Sun Media Corporation Sdn Bhd v The Nielsen Company (Malaysia) Sdn Bhd (23NCVC-83-2011)

The Plaintiff is Sun Media Corporation Sdn Bhd (“Sun Media”). It publishes a national newspaper called “theSun”. It started out as a paid newspaper but in 2002, the newspaper was circulated without any charge. Its circulation is mainly in the Kuala Lumpur/Petaling Jaya (KL/PJ) area with pockets of concentration also in the other big cities in Peninsular Malaysia like Georgetown and Johor Bahru. [2] The Defendant, The Nielsen Company (Malaysia) Sdn Bhd (“Nielsen”), is part of a global group of companies operating under the “Nielsen” brand. They represent themselves as being a global leader in media, consumer and market research. [3] Among the media players and stakeholders, the Defendant is known for its annual syndicated survey called the Nielsen Media Index or “NMI”. It is a wide-ranging survey that seeks to measure the consumption of different print and electronic media, viewership of satellite TV channels, listenership of radio channels and product and service consumption habits of the general population as well as a host of consumer products and 3 services such as telecommunication, credit cards, insurances and banking services, shopping, fast food and beverage consumption, automotive, holidays and travels and the like. The national survey since 1968 is conducted quarterly with the results being published on a biannual basis and available only to paid subscribers with qualifications on the results available, restrictions on usage and disclaimer of liabilities. Problem [4] For quite a while the Plaintiff has been nursing a nagging suspicion that the Defendant has been negligently, or worse still, maliciously under-representing the readership of theSun by its flawed design, methodology and conduct of its NMI Survey. [5] What seems to irk the Plaintiff more than anything else is that going by the results extractable (“Data”) from the NMI Survey and based on the circulation figures of the Audit Bureau of Circulation (“ABC”), the Plaintiff’s newspaper theSun would end up with a readership of less than 1 i.e. 0.8 for the year 2008 and that cannot but be ridiculously false! [6] There were meetings between the parties held to address some of the concerns of the Plaintiff but to no abatement of the Plaintiff brewing discontent. Despite Sun Media’s disapproval of Nielsen’s NMI Survey 4 Methodology, it decided to commission Nielsen to do a syndicated survey in 2008 called the Prime Media Index Survey (“Prime Survey”) of its readership in the Klang Valley from Nielsen’s existing database of online respondents with input and questions contributed by Sun Media. [7] According to Sun Media, the results of the Prime Survey showed that the NMI Survey results cannot be relied on, and is false, as in persistently under-representing the readership of theSun especially amongst the “affluent”. Pleadings and Prayers [8] Perhaps emboldened by the findings and conclusions of this Prime Survey, Sun Media then launched this suit against Nielsen in 2011. It framed 3 causes of action against Nielsen for Negligence, Defamation and Malicious Falsehood. The period of complaint of Sun Media is for the years 2006-2010 of the NMI Survey. [9] Sun Media contended that Nielsen had stubbornly and strenuously refused to accept that there were flaws in the design of the NMI, its methodology and the manner in which the NMI was conducted such that the results of the NMI Survey yielded an under-represented readership 5 (“Offending Publication”) to the detriment of its advertising revenue and consequential loss of profits. [10] According to learned counsel for Sun Media, his client was constrained to publish two articles dated 2.3.2011 and 3.3.2011 (“Offending Articles”) in theSun newspaper in which Sun Media took issue with the accuracy of the NMI. [11] Reacting to the said two articles published in theSun as contended by Sun Media though Nielsen would see it as a case of responding to the said two articles, Nielsen published their letter of 8.3.2011 (“Offending Letter”) to its subscribers and customers by which Nielsen sought to provide its clarification on the 2 articles published in theSun and to allay fears that its subscribers may have with respect to the reliability and validity of the readership figures in the NMI. [12] It prayed inter alia for general damages to be assessed for the Offending Publication in the NMI Survey and in the said Offending Letter of Nielsen to their subscribers. It also prayed for special damages of RM303.52 million for loss of profit from advertisements. There was also prayers for a Declaration that the readership data published in the NMI Survey Report from 2006-2010 in respect of theSun newspaper is false, 6 inaccurate and misleading and some related injunctive reliefs besides interest and costs. [13] Nielsen on the other hand counterclaimed inter alia for damages for defamation based on the 2 Offending Articles that Sun Media carries in theSun. [14] By consent this trial was bifurcated with the trial being confined to findings on liability first. Principles Whether the Defendant owes the Plaintiff a Duty or Care [15] First I shall deal with Sun Media’s claim for negligence. [16] It must be stated at the outset that we are not dealing with personal injury arising out of the negligent act of another person who stands in a proximate relationship with the person who has suffered a personal injury arising out of the action or inaction of the tortfeasor which damage in the personal injury is reasonably foreseeable as in a running down case. [17] Those categories of cases are more settled by the law of negligence in the application of the neighbour principle propounded since the seminal case of Donoghue v Stevenson [1932] UKHL 100. 7

Neither are we dealing with a relationship of a professional with a client where there is reliance on the advice given by the professional whether for a fee or gratuitously, and some economic loss ensued as a result of the advice given. There is a legion of cases governing negligence of professionals like accountants, lawyers, architects, fund managers, bankers and the like. [19] Very importantly we are also not dealing with a case where advice, findings and conclusions are given in the context of a contractual relationship where there are clear terms expressed as to the limitations of the survey methodology, restrictions on use of the survey results. That would apply to the subscribers of NMI Survey of which Sun Media is not one during period in question i.e. 2006-2010. It is important to mention this, at this juncture both for analysis and analogous comparison and extension of new categories of a duty of care. [20] Where there is sought to be recognized a new category of a duty of care where contractual relationship is missing and the reliance is on pure statements made or which can be inferred and a so-called economic loss arising out of reliance on such a statement, then the Courts must tread 8 more cautiously and more so when there would be some remedies available under some other causes of action. [21] When there is no precedent to fall back on and where there is only the guiding light from principles previously stated in the law of negligence, one can appreciate the dilemma expressed by the Law Lords in Hedley Byrne & Co Ltd v Heller & Partners Ltd [1964] AC 465, where the neighbour principle was extended to cover negligent misstatement. Lord Hodson said, at p. 514: "I do not think it is possible to catalogue the special features which must be found to exist before the duty of care will arise in a given case." [22] Lord Devlin in sounding the depths said, at pp. 529-530: "I do not think it possible to formulate with exactitude all the conditions under which the law will in a specific case imply a voluntary undertaking any more than it is possible to formulate those in which the law will imply a contract." [23] Ultimately it is for the Court to decide on whether or not there is a duty of care as in a tortious duty in a given factual matrix that does not quite fit into the traditional pigeon holes of personal injuries, professional 9 relationships, and contractual relationship where the act complained of is in the nature of statements made and the loss is purely economic loss. [24] From decided cases in our Courts, the position is that the Court should be slow in extending the categories of duty of care in cases involving pure economic loss and statements or advice given when or where there is no privity of contract or proximity of relationship between the parties. [25] The House of Lords case of Caparo Industries Plc v Dickman [1990] 2 AC 605 makes reference to the “three-fold test” of 1. harm must be reasonably foreseeable as a result of the defendant’s conduct, 2. the parties must be in a relationship of proximity and 3. it must be fair, just and reasonable to impose liability. This test was popularized by Bingham LJ in the Court of Appeal in Caparo Industries case (supra) where he confessed as follows: "It is not easy, or perhaps possible, to find a single proposition encapsulating a comprehensive rule to determine when persons are brought into a relationship which creates a duty of care upon those who make statements towards those who may act upon them and when persons are not brought into such a relationship.” 10

In Caparo Industries case (supra) the House of Lord held that the auditor who prepared the accounts of a target company, Fidelity plc, does not owe a duty of care to a shareholder like Caparo who had some shares in Fidelity plc and subsequently increased it in reliance on the audited accounts of Dickman, the auditor. It turned out that the accounts showing profits was false and that Fidelity had in fact made a loss. That was not all as after Caparo Industries had taken control of Fidelity plc, the company was in a worse state than had been revealed by the directors or the auditors. It sued Dickman for negligence in preparing the accounts and sought to recover its losses. This was the difference in value between the company as it had and what it would have had if the accounts had been accurate. [27] Human enterprise as we know it, and with that business and commerce, encompasses a myriad of human relationships permitting of various permutations such that it would be easier to ask the question of whether the parties are in a proximate relationship than it is to give an answer. 11

Be that as it may, one must endeavor to apply the “three fold test” with a view to discern if there is a duty of care that the law should impose or that liability should be assumed by the party said to be in breach. [29] With respect to harm that must be reasonably foreseeable, it may be said that a duty of care is owed to those who are subscribers of the NMI Survey who are mainly advertising agencies and advertisers. If the results are false or misleading, perhaps unwise and ineffective decisions may be made by advertising with certain media but then again, advertising agencies and advertisers deal with a multitude of factors with respect to placement of advertisements such as target audience, reach of the media, language of the media, costs, preference of clients themselves and feedback from the respondents to the advertisement as to how they had come to know of a certain product or services being advertised and the like. It would be quite difficult to say what are the determinative factors in matters of this nature with respect to where to place client’s advertisement. If at all there is, then it would be a case where there is perhaps a cause of action in negligence by the advertising agencies or the advertisers who had spent unnecessary costs on advertising in what may be regarded as ineffective print media. 12

However with respect to advertising agencies and advertisers who made decisions on where to advertise it would be difficult to identify the loss or damage that is foreseeable for in such matters it is difficult if not impossible to attribute loss to having recommended for clients to advertise in an ineffective medium. [31] More than that those who subscribe are subject to the contractually agreed disclaimer and exemption from liability where the use of the information derived from the NMI is concerned. [32] As for non-subscribers they do not have access to such information in the NMI and any access must be by means of unauthorized sharing or disclosure of the NMI Survey results. That aside it is difficult to foresee how it can be said that there would be a clear loss of advertising revenue for the simple reason that advertising decisions are the results of a whole bunch of variables. Whilst reliance of the NMI Survey is one factor to be considered, it certainly cannot be, the be all and end all of all advertising decision-making. In dealing with an increasingly sophisticated audience, media agencies are as astute as they are anxious of getting the maximum mileage for their clients, stretching every advertising ringgit to its limit. 13

It is a fact that Sun Media was not a subscriber to Nielsen’s NMI Survey for the period in which they leveled their attack of under-representation of their readership, i.e. 2006-2010. It had subscribed to the NMI Survey from 1995-2000. [34] Where proximity of relationship is concerned it is certainly not a contractual one and the only proximity is that the print media itself is the subject matter of the NMI Survey. The dicta of Lord Devlin in Hedley Byrne (supra) at pages 530-531 is worthy of reference and reproduction as follows: “I shall ... content myself with the proposition that wherever there is a relationship equivalent to contract, there is a duty of care. Such a relationship may be either general or particular. Examples of a general relationship are those of solicitor and client and of banker and customer ... Where there is a general relationship of this sort, it is unnecessary to do more than prove its existence and the duty follows. Where, as in the present case, what is relied on is a particular relationship created ad hoc, it will be necessary to examine the particular facts to see whether there is an express or implied undertaking of responsibility. I regard this proposition 14 as an application of the general conception of proximity. Cases may arise in the future in which a new and wider proposition, quite independent of any notion of contract, will be needed.” (emphasis added) [35] If at all there is “proximity”, it is that theSun was one of the 23 newspapers being surveyed under the NMI Survey. The fact that media companies in general are interested in how they fare or are perceived would not by itself be sufficient for the Court to say that there is an assumption of liability where the results of the NMI Survey is concerned. [36] There is merit in Nielsen’s submission that Sun Media, not been a subscriber of the NMI for the period it claimed to have suffered pecuniary loss, is not a user of the data and further has not shown how it had relied on the NMI Data to its detriment. On the contrary, the documents show quite clearly that from 2001 onwards Sun Media has been challenging the reliability of the NMI data (See Sun Media’s letter dated 22.2.2001 CBD1 at pg.1). The General Manager - Marketing wrote as follows: “Re: The ACNielsen Media Index I write to confirm that we will not be subscribing for year 2001. 15 You will recall our meeting just before the Chinese New Year where we expressed our disappointment and disagreement with the survey results. This particularly in connection with the total number of readers of the Sun. We understand that surveys are like “streetlights” and certainly not microscopes. We also understand that it is not necessary to have direct correlations between circulation and readership trends. But, we are now in a situation where your survey results depict trends that is in direct contradiction to our actual circulation returns.” [37] There is no room for one to say that Nielsen has by its conduct assumed responsibility or that there is justification to impose liability on Nielsen. Nielsen specifically restricts its subscriber/customer from publication or unauthorized use of the data under Clauses 1.3 to 1.6. These restrictions are applicable to Advertising Agency, Advertiser and Media Owner so that there would not be further reproduction or sharing of those data with others and non-subscriber except under limited cases in 1.3(a) to 1.3(e). [38] More importantly, learned counsel for Nielsen highlighted that Nielsen has expressed the following limits to their liability under Clause 3.1 to 3.3: 16 “3. ACCURACY

3.1 The Customer shall recognize that the Data are estimates derived from sample surveys carried out in accordance with market research methodology and as such are subject to limits of statistical error.

3.2 ACNielsen shall use its best endeavor to ensure the accuracy of the Data but does not warrant the accuracy of any Data provided, nor does it accept responsibility for any error contained in or any omission from the Data or in any event any loss damage or injury suffered directly or otherwise arising therefrom.

3.3 ACNielsen is not responsible for any use to which the Data are put or for any interpretation placed upon them and the Customer undertakes to indemnify ACNielsen against any legal proceeding or law suits claims for damages or costs which may be taken against or incurred by ACNielsen pertaining or relating in any way or form to the Data” (emphasis added). [39] It would not be unreasonable for Nielsen to ask the rhetorical question that bearing in mind that Sun Media was not a subscriber of the 17 NMI Survey at the material time, and thus could not even rely on the Data to its detriment, can Nielsen be said to owe such a duty in a vague legal situation, when in contract it would not? Is this then not indicative of Nielsen adopting a ‘non-assumption of responsibility position’? [40] The fear that should one’s study published concerning a certain company’s product or services arising out of a survey undertaken with all its limitations and the disclaimers that come with it would expose one to suits of negligence would very likely dampen the spirit of research and study and the statistical method is a method of quantitative research which is open to criticism and even vilification as finally there is the court of informed users of these data and not forgetting the court of public opinion. [41] If there are defects and deficiencies that affects reputation and character then there is the existing law of defamation designed to address and redress the ensuing damage to character, if any, arising from the publication of the NMI Survey. If there is a misrepresentation of the print media that lowers its standing in the eyes of the fair-reader then perhaps there is a case for defamation or even malicious falsehood. [42] Here is a case where if the Court were to extend the category of negligence, it would create an incongruent and anomalous situation where 18 a non-subscriber would have better rights in tort compared to a subscriber whose cause of action is in contract with restrictions on the ability of the subscriber to sue under contract. This would be a fit and proper case where it would not be fair, just and reasonable to impose liability. [43] Realistically, in this case, those affected directly would be advertising agencies and advertisers who rely on perhaps the findings in the NMI Survey to help them make decisions resulting in foreseeable losses that they are able to prove and where one can easily and readily anticipate. There is no evidence that those who had advertised have shunned Sun Media or had advertised less with them based on the advice of the advertising agencies whose advice, Sun Media said, is predominantly drawn from the NMI Survey. Costs and circulation and preference of customers and target groups that the media/customer wants to reach out to would be relevant considerations. [44] On the other hand, media owners, who also would be interested in the NMI Survey results would be interested more for the purposes of planning and strategizing and marketing to their target audience [45] I can understand where Sun Media is coming from in their argument that since 2002, and with the shift to the free-newspaper model, Sun 19 Media’s revenue from theSun has been almost entirely dependent on the advertisement rates and the volume of advertising purchased in theSun. It was forcefully put forward that Sun Media’s case is that the NMI Survey has a direct impact on its advertising revenue. It is more a case where unsuspecting advertising agencies and advertisers had relied on the NMI Survey resulting in loss of advertising revenue to Sun Media. [46] However the evidence adduced showed that advertising revenue has, on the contrary, increased during the years complained of in 2006-2010. Are we then to say that this is a case of “in spite of” rather than “because of” under-representation of theSun readership? There is simply no evidence of a direct correlation between the results of the NMI Survey and the allegation of loss of advertising revenue. Being a relatively late comer to the print media industry compared to the other 2 major English national newspapers there is always the costs differentiating factor that can be harnessed to good use and the fact that it is distributed free would perhaps put the newspaper within reach of those young executives who perhaps would not be keen to spend on newsprint but would rather go for free online or hardcopies version of newspapers made available in theSun’s receptacles placed at strategic places such as commuters’ stations. 20 [47] It would not be out of place to say that the “Y” and millennial generations gather their news from various sources streamed to their hand phone, which appliance one would not leave home without. The handphone from which the younger generation store and capture everything and read the news is likened to an extension of the human anatomy. In an unanimous opinion of the United States Supreme Court on the need for a warrant before a search may be done on a cellphone of an arrested person, Chief Justice John Roberts of the United States Supreme Court in Riley v California 573 U.S. (2014) at page 9 observed rather wittily of a cellphone “which are now such a pervasive and insistent part of daily life that the proverbial visitor from Mars might conclude they were an important feature of human anatomy.” [48] The Court should be slow to extend categories of negligence to cover economic loss from a non-subscriber of the NMI Survey where the loss is not foreseeable as in showing the correlation between loss in advertising revenue and the alleged under-representation of the readership of theSun where if there is any damage to reputation, the law of defamation, which in our jurisdiction has been fully developed considering the fertile field of litigation in this area, would be more than able to attend to it. 21 [49] To impose liability would not only stifle research via the statistical method but would also expose a person like the Defendant "to a liability in an indeterminate amount for an indeterminate time to an indeterminate class,” to borrow the enduring words from the pen of Cardozo C.J. of the United States Supreme Court in Ultramares Corporation v Touche, 174 N.E. 441, 444. The prospect of that problem would be enough to weigh heavily, probably conclusively, against the imposition of a duty of care where there is no personal injury, no contractual or other relationship of proximity and where there is an available remedy in the tort of defamation or malicious falsehood. [50] Policy-wise in cases where contractually the use of the results in restricted to only the subscriber and even then with disclaimer of liability, this Court would also be slow to extend such new categories of negligence. [51] Learned counsel for the Nielsen referred to the Federal Court’s Decision in CIMB Bank Bhd v Maybank Trustees Bhd and Other appeals [2014] 3 MLJ 169 at pg. 187 & 190-192. There the Federal Court had to deal with a case of negligence pleaded against an arranger in a bond placement exercise. In the Information Memorandum issued there was a disclaimer on the part of the arranger. The disclaimer in the form of 22 an ‘important notice’ is at para 35 of the judgment at pg. 187 of the report. The Federal Court in deciding on whether the arranger owed a duty of care to bondholders who purchased the bonds on the strength of what was stated in the Information Memorandum came to the conclusion that the arranger did not owe a duty of care. The Federal Court was not convinced that there was that degree of proximity or a ‘special relationship’ to indicate a duty of care and also because of the disclaimer. [52] His Lordship Ariffin Zakaria Chief Justice (as His Lordship then was) at para 48 of the judgment took a cautious approach and observed as follows: “It is also worth noting that both the High Court and the Court of Appeal, without considering the special facts and circumstances of the case, simply ruled that there existed a duty of care on the principles of ‘foreseeability’, ‘proximity’, ‘neighborhood’ and ‘fairness’. In applying those general phrases, it is important to bear in mind the warning given by Lord Roskill in Caparo Industries plc v Dickman [1990] 2 AC 605 where he said: ‘But ... such phrases are not precise definitions. At best they are but labels or phrases descriptive of the very different factual 23 situations which can exist in particular cases and which must be carefully examined in each case before it can be pragmatically determined whether a duty of care exists and, if so, what is the scope and extent of that duty’”. See also the case of Dato’ Ariff Wan Hamzah & Ors v Hwang DBS Investment Bank Bhd & Anor [2013] 1 MLJ 526 at pg.548. [53] Looking at the whole circumstances of this case where the Plaintiff suing is not a subscriber of the NMI Survey and where the party most affected economically are perhaps the advertising agencies and advertisers and seeing the restrictions placed even on subscribers, this Court would hold that there is no duty of care owed by the Defendant Nielsen to the Plaintiff Sun Media in the circumstances of this case. Whether there has been a breach of that Duty of Care in the Design, Methodology and Conduct of the NMI Survey [54] Assuming for a moment that there is a Duty of Care, I shall now consider whether there has been a breach of that Duty. [55] It is said that there are flaws in the NMI Survey methodology such that there is an under-representation of the upper Social Economic Segment (SES) group of income earners in the KL/PJ area and the 24 inevitable error was that theSun’s readership in the years 2006-2010 had been under-represented. Sun Media postulated that this is caused by the barriers in interviewing respondents from this category. [56] Sun Media stopped subscribing to the NMI Survey starting 2001. It had disagreed with the NMI Survey results. Nielsen had in response to Sun Media’s complaint, offered to take Sun Media through its NMI Survey Methodology. Several meetings were held between the parties in 2007 to clarify and discuss Sun Media’s queries. [57] It must not be forgotten that a survey is not a census. No two surveys conducted at different times will yield same results. There are statistical errors within a range that is acceptable in the industry. The NMI Survey Methodology [58] I am satisfied that steps have been taken to ensure that the methodology employed meet the acceptable standards of statistical research method in a survey of this nature where one is concerned with the issue of reliability and validity. As stated the NMI Survey seeks to measure the readership of newspapers and other print and electronic media including viewership of satellite TV and listenership of radio and a whole host of consumer products and services. Where print media is concerned it 25 is close to about 30 of them including weekly publications. It covers the general population of the whole of Peninsular Malaysia. [59] It will of course suffer from deficiencies from the point of view of those who are keen on the same consumer habits or perhaps even readership of a particular segment of the population in a particular area like the KL/PJ area. For that the party concerned would have to commission a customized survey with the input of the parameters the party wants such as the Prime Survey 2008 that Sun Media did commission Nielsen to do. [60] All Survey methodologies suffer from some inherent weaknesses and limitations and Nielsen has informed all their subscribers of these through the NMI Agreement, NMI brochure and the NMI & IMS client training materials (See CBD 1 at pg.83-86 and pg.87-90, CBD 7 at pg.4233 - 4291 and pg.4394-4400). [61] To begin with the sample size would be quite important. The annual sample size is about 10,000 with surveys done every quarter and the results published twice a year on a 12-month basis for the periods of a July-June and January-December. The subscribers may access the results of the NMI Survey through the use of the Interactive Market Systems (“IMS”) software and/or its predecessor, Periscope. The IMS software is a 26 data analysis program that allows the NMI subscribers to produce readable data relating to their target segments by demographics, media consumption and product usage. It contains data analysis functions such as Cross Tabulation, Average Issue Readership, Reach Accumulation, Gross Reach, Net Reach, Average Frequency, and Gross Opportunity to See (See para 10 of the Statement of Agreed Facts). [62] The contents of the IMS database for each subscriber varies depending on the subscription package. [63] In other words one can extract as much information from the Data by way of input of parameters for each particular print medium or a combination by way of comparison coupled with other permutations that a subscriber might be keen in for example young working adults in the KL/PJ area who reads an English daily, is a graduate and own a house and a car who go for holidays overseas once a year. Its versatility lies in the fact that the data is malleable enough to be crunched to yield different permutations that are permissible within the parameters surveyed! [64] Sun Media was a subscriber to the NMI Survey for 6 consecutive years from 1995-2000 and was aware of the terms of the NMI Agreement and the limitations and disclaimers on the usage of the data. 27 [65] The survey is done in a face-to-face interview. There is no suggestion that this sample size is inadequate for the population of Peninsular Malaysia though of course, the bigger the sample size the more accurate perhaps the findings but then that has to be balanced by the costs factor of the survey. The sampling population consists of Malaysian citizens aged 15 years and above and projected to the population estimates of Peninsular Malaysia according to the numbers obtained from the Malaysian Department of Statistics (“DOS”). There is sufficient randomness in the sampling method of stratified and systematic random sampling by use of the Primary Sampling Units or PSU. [66] Miss Wong Mae Suen, DW2 had explained this stratified random sampling process in her witness statement (DW2-WS) as follows: “19. Q: Please explain the NMI sampling plan. A: The NMI sampling plan contains information on the number of sampling points (also known as sampling areas or PSUs or maps) for each survey quarter, the target interviews to be achieved by ethnic group and each of the 12 locations and the sampling procedure. ...

Q: What happens after the preparation of the NMI sampling plan? 28 A: The next step is the selection of PSUs/maps for interviews. This is Stage 1 of Nielsen’s 3 stage stratified random sampling procedure for the selection of the NMI respondents. For 2006-2010, the maps were selected through Nielsen’s master sampling frame.”

Q: What is a master sampling frame? A: A master sampling frame is a database of maps or areas used for the NMI Survey Nielsen first purchased these hand-drawn maps from the Department of Statistics sometime during the 1970s and they have been maintained and updated by Nielsen until today. The number of maps may vary from year to year, depending on whether there are any new areas of development or demolished areas within Peninsular Malaysia. In 2006, there were approximately 15,500 maps and by 2009, there were about 18,000 maps in the frame. Each map contains approximately 220 DUs [Dwelling Units] (CBD 1 Tab 52, p. 94-107). ...

Q: How is a PSU/map selected for the NMI Survey? A: From the master sampling frame, each PSU is classified as low/medium/high density in respect of the ethnic/racial composition of 29 the DUs within that PSU. The PSU classification is set out in the sampling pans, see for example CBD 5(1) p.723, as follows: PSU Classification Low Between 20 to 60 DUs of the targeted ethnic group Medium Between 61 to 120 DUs of the targeted ethnic group. High Above 120 DUs of the targeted ethnic group. a) First Step – a PSU frame by Ethnic Group is created. For example, a Malay Frame would consist of all PSUs within a minimum of 20 Malay DUs. The PSUs are classified as Low Malay, Medium Malay and High Malay. This is to determine the DU skipping interval to be used by fieldworkers later during fieldwork. b) Second Step – PSU selection. The PSU selection is done by Ethnic Group, i.e. based on the number of PSUs required for each Ethnic Group and Location in the sampling plan. Taking the earlier Q2 of 2008 sampling plan, a total of 5 PSUs are to be selected from the Penang Market Centre (PG MC) Malay frame. This 30 process then repeated for the Chinese and Indian frames.” (emphasis added) [67] With respect to the use of geographical maps as the basis of sample selection she explained in Q&A 24 of her witness statement (DW2-WS) as follows: “24. Q: Why does the NMI use geographical maps as the basis for its sample selection? A: The main reason for using PSU maps is that we do not have a complete and accurate list of all members of the population. Dividing the country into small geographical units serves as a sampling frame because the entire population can be mapped to one and only one of these geographical areas.” [68] The estimation of ethnic composition of each PSU map. As she stated in Q&A 25 of her witness statement (DW2-WS): “25. Q: How does Nielsen estimate the ethnic composition of each PSU map? A: When a PSU map is created, the number of DUs by ethnic group in each PSU is estimated through physical observation and counting. 31 The ethnic group for each DU is determined through visual observation such as Quranic verses hung outside a Malay home, Chinese altar outside a Chinese home or mango leaves hung at the doorstep of an Indian home. For unidentifiable homes, the neighbours are asked.” [69] It is representative and random enough of the various categories of consumers surveyed in a nationwide survey where regard must be had to the projection to the population based on the latest data available from the DOS and then weighting to be done according to age, race, gender, location and income bracket. It is acknowledged that there are difficulties with respect to conducting face-to-face interviews with people in the affluent top income category and sufficient use of boosters have been put in place to create any sedimentation as it were. [70] Each side has called their own Expert Witness; Mr Suresh Nimbalkar (PW7) from India for the Plaintiff, Sun Media, and Mr Jonathan Jephcott (DW6) from Australia for the Defendant, Nielsen. Much has been submitted on by learned counsel for both sides against the other Expert Witness where competencies and credibility is concerned. 32 [71] Mr Jephcott’s experience seems to be broader and his expertise deeper where statistical survey methods and analysis is concerned compared to Mr Nimbalkar whose experience is more confined to that of India alone. Mr Jephcott has a BA (Econs) Hons. in Economics, Maths & Statistics from the University of Sheffield and has a diploma from the Institute of Statisticians (Market Research) (See Mr. Jephcott’s CV at Annexure 1 D-RER). Mr Jephcott has more than 50 years of experience in market and media research and has worked in no less than 25 countries globally in a leadership role. His various publications and professional achievements are set out in his CV. [72] Mr. Jephcott also possesses various professional qualifications and memberships such as Fellow of the Royal Statistical Society, Chartered Statistician and member of ESOMAR. More recently from about 2008, he was the founder of ARF Online Research Council (ORQC), Chairman, ARF ORQC Online Quality Metrics committee, member of ESOMAR’s On-line best practice committee, member of the Professional Development Committee of the AMSRS and Examiner for AMSRS PQMR qualification. [73] Mr Nimbalkar’s tertiary education is a Bachelor’s degree in Chemical Engineering and a Master’s degree in Management Studies. I agree that 33 these are unrelated qualifications to market research or applied statistics. He appeared rather defensive when questioned on his educational background and conceded that where relevant professional studies is concerned he had only studied managerial statistics as a subject and taken some related lectures but did not see it as necessary to set out in his CV as he did not like talking about himself. [74] Mr Nimbalkar conceded that he is not a member of any international research body outside India. His relevant experience seems to have been derived from his involvement in the Indian Readership Survey (“IRS”) for about 8 years when he was with the Hansa Research in Mumbai from 2006 to April 2016. He confirmed that Hansa lost the IRS Contract in 2013 to Nielsen India. He had tried to give the impression that Hansa did not lose the contract but had declined the contract. However, upon further questioning, he then disclosed that Hansa together with another research company IPSOS actually bid for the contract but subsequently “opted out” due to certain disagreement on the pre-conditions. [75] Much was said about Mr Jephcott’s previous association with Nielsen Group. It was Mr. Jephcott himself who volunteered information on his previous employment with Nielsen. During 1991-1998, he had a regional 34 role and was involved in TV audience measurement in Malaysia. He testified that he worked out of Australia and apart from an audit of all the systems and media in Asia which would have included the NMI, he had no prior direct involvement in the NMI in Malaysia (See Notes of Proceedings 9.11.2016 pg.26 line 23 - pg. 29 line 16). More importantly, he had spent more years working for other major competitor companies such as AGB, TNS and Synovate before and after his tenure with Nielsen. He was consulted by many others in this business including IPSOS, the survey company that now Sun Media has used in recent years after the filing of this suit to give an alternative presentation of the circulation and readership of theSun. [76] Learned counsel for Nielsen had referred to Zulhasnimar bt Hassan Basri & Ors v Dr Kuppu Velumani P & Ors [2016] 3 MLJ 625 at pg. 645- 646, on the proposition that there is no issue even when expert is an ex-colleague. Even non-disclosure of conflict of interest does not automatically disqualify an expert as was held in Toth v Jarman [2006] EWCA Civ 1028. The key question is whether opinion is independent. [77] I am satisfied that his previous association with Nielsen has not impaired his independence in any way in the way in which he was not 35 reticent to express his disagreement with the other witnesses of Nielsen. I am prepared to accept both the expert witnesses as experts in their own right and where their opinions are at variance I shall state the reason for my preference. [78] Whilst Mr Nimbalkar had attempted to poke holes and pour water on the credibility of the NMI Survey such as the sample size, the rejection rates, the frequency of updates of the PSUs, the mechanism to ensure that personal biases are minimized, the problem with sedimentation and the perennial problem of challenges in reaching respondents in condominium and gated and guarded communities and the like, they are not of such seriousness as to undermine the NMI Survey from what is general and approved practice in a national survey statistical method. [79] Much criticism was also directed at the fact that Mr Jephcott, the media research expert for Nielsen, had expressed that the “NMI has changed little in design and implementation since its establishment...”. I do not think that fundamentals in terms of size of sampling units and randomness of sampling have changed much through the years and this is a field of research where like all things, if it is not broken then there is no good reason to fix it. Where there is change of demography there has been 36 an update of the population sample and Nielsen has put in place mechanisms to avoid sampling biases and to address sedimentation issues in sampling. [80] Sun Media also took umbrage with Nielsen that the raw data as in the questionnaires were destroyed 2 years after they were filled up and a year after the release of the NMI Survey results. Mr Liao Peng Guan DW4, the IT head of Nielsen, had referred to the ICC/ESOMAR International Code (On Market and Social Research) and to Article 7(c) thereof where it was recommended as follows: “Personal information collected and held in accordance with this Code shall be: ... Preserved no longer than is required for the purpose for which the information was collected or further processed” [81] There is no evidence of any suppression of evidence on the part of Nielsen where the keying in of the data into the software is concerned. Neither had Sun Media put Nielsen on Notice that these questionnaires should not be destroyed as they would be proceeding with a discovery application on it. Therefore there is no room to suggest that there should be 37 any adverse inference drawn merely because the raw data as in the questionnaires could not be available for examination or audit. The Prime Survey 2008 [82] It was argued for Sun Media that the falsity of the NMI Survey with respect to readership of theSun when compared to the other English dailies like the Star and NST can be proved by way of the findings in the Prime Survey 2008 which Sun Media commissioned Nielsen in November 2007 to undertake with respect to parameters supplied by Sun Media and agreed to by Nielsen. The main objective of the Prime Survey was to provide Sun Media with an accurate and up-to-date picture (understanding) of the readership of affluent Malaysians and determine their attitude towards theSun newspaper and its major competitors. The results of the Prime Survey were intended to be used by Sun Media as its marketing tool (See CBD 5(8) at pg.3788). [83] This Prime Survey was conducted between January to March 2008 by inviting certain online respondents of the following characteristics to participate and complete the online questionnaires –

Residing in KL/ PJ;

Aged 20 and above; 38

Personal income of RM2500 or more per month;

Engaged in a white-collar occupation; and

Used the internet. [84] The Prime survey was conducted online and in English only with a total sample of 2521 potential respondents selected from Nielsen’s online panel. The total responses received for this Prime survey was 888. It cannot be overemphasized that unlike the NMI Survey, the Prime Survey was not a national readership survey of the random population in Peninsular Malaysia. Instead, the Prime Survey was based on a selected sub-population of online respondents selected from Nielsen’s online panel. Comparison of NMI Survey to Prime Survey [85] Sun Media contended that the market share of theSun is appreciably higher than the one surveyed under the NMI Survey and as such the NMI Survey must be unreliable and indeed false and wrong. [86] The table on comparative readership between the NMI Survey and Prime Survey is as set out in para 10B.5 of the Amended Statement of Claim (“ASOC”) is set out below: 39 [87] Taking the readership figure from the 2 Surveys and dividing that with the circulation figure from the Audit Bureau of Circulation (“ABC”), one would get a readership per copy or RPC score of 0.9 under the NMI Survey for 2008. It is admitted that this has been derived from an additional information from another source i.e. the ABC figures which Nielsen does not include in its Data. The RPC of theSun then is a conclusion that is arrived at outside of the available Data that is extractable by a subscriber. This is called “Inferred RPC” as submitted by learned counsel for SunMedia. 40 [88] The ABC is an independent non-profit bureau with a membership that consists of various members from the media industry such as representatives from newspaper publishers (including theSun, NST and The Star), the Malaysian Advertisers Association, the Association of Accredited Advertising Agents Malaysia and the Media Specialists Association of Selangor and Wilayah Persekutuan. [89] I accept that the role of the ABC is to report the audited circulation data of newspapers and magazines of its members for the benefit of the Marketing Communications Industry of Malaysia. [90] There was a question introduced by Sun Media into the Prime Survey in helping Sun Media understand the perception of a respondent as to how many persons read the respondent’s copy of a daily (English) newspaper. The question was as follows: “Approximately how many other people read YOUR copy of the DAILY NEWSPAPER which you read?” [91] This is called the perceived pass-on rate and from that figure one then proceeds in the same manner to calculate the “Perceived RPC” of 2.9 from the Prime Survey 2008. The RPC number is a useful estimate of the popularity, or reach, of a particular newspaper. 41 [92] The readership figures of theSun published by Nielsen, in the period between 2006 and 2010, are set out in paragraphs 11A to 11D of the ASOC (the “NMI Figures”). As explained by Nielsen, “Readership” is not synonymous with “Circulation”. They are different concepts and distinct measurements. They are not comparable especially in the case of theSun being a free newspaper. “Readership” looks at the number of people who claim to have read the newspaper yesterday. [93] There are clearly material differences between the Prime Survey 2008 and the NMI annual survey such that it would not be reasonable to compare one to the other with respect to yielding the same results where readership is concerned. There are differences in the Population Profile, Size and Projection to the national population. Here are some of the differences between NMI and Prime such that a comparison with the conclusion that NMI is wrong and would not be logical nor reasonable. [94] First, NMI is a syndicated survey nationwide whereas Prime is customized and focused on the KL/PJ area. NMI does not target the readership of theSun or any other media or publications. The NMI Survey covers more than 30 different types of media, product and services and 42 readership of newspapers and magazines is only but one part of the entire Survey. [95] As explained by Mr. Richard Hall, Nielsen’s Managing Director, in his evidence (See Examination-in-chief of Mr. Richard Hall DW1-WS Q&A 32): “I say so because the NMI is a national survey, which methodology is not skewed in favour of any one particular client or product, and which is not intended to provide a deep analysis of any particular product or market segment. The targeting of a specified market or product would be done through customized research such as the Prime survey and not a syndicated survey of national level readership (and other consumption) like the NMI. A feature of a national level readership and consumption survey is that the same research methodology and sample design is used across the board for all the titles (and other products and services) that form part of the survey, and that the survey occurs on an ongoing basis. This ensures that the consumption of all products within the survey is measured against the same parameters and that they are data trends.” (emphasis added) 43 [96] In NMI there is a stratified random sampling whereas in Prime it is preselected from a sub-population of the Defendant’s online profile; more of a dipstick survey for a more particular point in time rather a survey throughout the 4 quarters of a year. [97] Profile is also different of their population surveyed in that for NMI it is consisting of citizens only, 15 years old and above, face-to-face interview, against all income levels and the closest to Prime Survey where income is concerned is the RM3,000 per month salary. [98] The Profile in the Prime Survey included non-citizens like expatriates, hotel guests and would cover foreign workers as well; all earning RM2,500.00 and above per month, 20 years old and above, white collar workers who are CEOs, directors, professionals, executives, managers, businessmen (“PEMBs”), internet users and English speaking. [99] Sample size wise that of NMI is 10,000 whereas that of Prime is about 888 that responded. [100] NMI is projected to the whole population of Peninsular Malaysia based on age, gender, race and region where weighting is concerned whereas for the Prime Survey the sample is projected to a universe of 403,000, which is Nielsen’s estimate of the population of Top Management 44 and Affluent Consumers in the KL/PJ area. As there were no available Government statistics for this target group, the universe estimate was obtained by using the 2003 Nielsen’s Asian Target Market Survey (“ATMS”) base. ATMS is a survey across 7 Asian cities, namely Hong Kong, Singapore, Kuala Lumpur, Taipei, Manila, Bangkok and Jakarta using both telephone as well as face-to-face interview. [101] As explained by DW3 Miss Andrea Douglas who was the Executive Director who did Market Research in Nielsen Malaysia in 2007 when the Prime Survey was commissioned: “The weighting and projection of the Prime results was also different from the NMI. The Prime sample was rim weighted and projected to the estimated population of top management and affluent consumers (as defined in the Prime survey proposal at p. 4004 of the Common Bundle of Documents (“CBD”) 5(8)) living in Kuala Lumpur/Petaling Jaya whereas the NMI sample was cell weighted and projected to the general population of Peninsular Malaysia aged 15 and above.” [102] The Plaintiff has no problem accepting the results of the NMI Survey that are in its favour in its presentation in 2007 to its stakeholders. Some of these results are for instance steady growth of readership in the KL/PJ 45 area, higher income profile, urban concentration, university education, decision makers and laptop owners. [103] I think one cannot cherry pick the favourable from the less favourable for the conclusions are drawn from the same common pool of Data that are extracted from NMI by use of the software given to subscribers which allows for various permutations to be generated based on the parameters that one may input. The brunt of Sun Media’s complaint [104] Perhaps the sore point in Sun Media’s complaint is that if one were to compare the Readership Per Copy (RPC) of theSun vis-a-vis that of the other 2 English dailies in the Star and the NST, one sees that the RPC is less than one. It goes without saying that when one is not comparing like to like with the same parameters then one should not be surprised that the results may differ though there are other findings that are the same and Sun Media is not complaining. Sun Media maintained that their suspicion of under-representation of its readership in the KL/PJ area among high income earners or RM3,000 per month salary and above has been proved right in the Prime Survey for the readership figure increases substantially and even surpasses that of NST. 46 [105] However Nielsen does not calculate or rely on the RPC where its results is concerned. It does not claim to and does not measure RPC for the newspapers. It makes no reference and does not take into consideration the circulation figures available from the ABC. It only measures readership by the number of readers as projected to the total population. [106] The RPC is derived from dividing the readership figure with the number copies in circulation. Here is where the difference and divergence lie in the audit method of paid newspapers compared to free newspapers. Whilst one can expect a paid readership to have at least one reader per newspaper subscribed, if not more as would be the likely if the newspaper is subscribed for office or home reading. It is a reasonable inference drawn from why people would pay for a newspaper if not for the purpose of reading, both in keeping abreast with the news and for general information and the sheer joy of reading things of interest. [107] However, the same cannot be said of a free newspaper. Unless all of it is handed out by a staff of the media company normally at strategic places where there is high human traffic like at commuter stations where waiting time for the commuter and commuting can be usefully occupied 47 flipping through the pages of the newspaper or even glancing at the caption to pictures therein and that would have been considered under the Survey method as having read, there is bound to be leakages and copies not picked up and read and picked up but for other purposes other than for reading. [108] The circulation figures are supplied by Audit Bureau of Circulation (“ABC”) and from the evidence, we see that there is very restricted audit where circulation is concerned for a free newspaper like theSun in this instance. As gathered from the evidence that emerged from the cross-examination of the Plaintiff’s witnesses, the ABC does not audit what is not picked up at the end of the day from the various pick-up points. What they audit, and even then confining only to copies made available for collection at the pick-up points were covering only hardly 1% or so of the total number of pick-up points. [109] Sun Media further relied on its own stock count report. That report also does not refer to copies left behind at the end of the day. There are times when copies are left untouched or taken for other purposes other than reading as testified by the Defendant’s witnesses. It would not be unfair to say that there are times when more than a copy is picked up 48 because it is relatively thinner when compared to the other 2 main English dailies. [110] Surely one can take cognizance of the fact that there is an industry of second-hand or recycled newspapers in Malaysia. It would be different if Sun Media actually engages workers to physically hand out all copies published every day. It can then be said that the RPC would be close to 1 on the assumption that the one who stretches out his hand to collect from the staff of Sun Media would read it for a minute or so at least based on the agreed definition of recently read copy. The word for it is “commuter newspaper” presumably read more by commuters when they are waiting for public transport and during their journey. [111] To have an RPC of less than 1, which information is not provided by Nielsen but rather derived from the circulation figures furnished by ABC, is not necessarily alarming or indicative of the fact that the NMI Survey results as complained by Sun Media is inaccurate. The fact of leakages is real as borne out by the stock count report and the propensity of human nature to take sometimes more than what is required for purposes other than for reading. 49 [112] The traditional model of paid subscription on the other hand is audited by the ABC based on receipts of payments and it is generally accepted that a house or office that subscribes it would have at least 1 person reading it with the likelihood that others at home or in the office would be reading it as well. [113] To say that the NMI Data complained of is inaccurate and an under-representation, one would have to believe that almost every copy that the Plaintiff publishes of theSun that is being distributed as being effectively picked up and read by at least one person. The Plaintiff has not shown any evidence that that is the case and it is merely speculation that every copy left at the pick-up points is being taken for the purpose of reading. [114] This is what Sun Media expects which is that the readership of theSun to be at least 1 in that every copy is being picked up for reading. That has not been borne out by the evidence. No doubt the Prime Survey serves a limited though useful purpose; to test whether one’s strategy in reaching out to one’s target audience has been successful and to see a discernible trend in the reading habits and preference of the High-income earners in the KL/PJ area. It was designed to provide a picture of the readership of the Sun vis-à-vis its major competitors by testing against a 50 pool, not necessarily representative, of readers in the KL/PJ area drawn from the respondents with the criteria specified. [115] To compare NMI Survey to Prime Survey where the ratio of readership and RPC are concerned would be to compare an apple to an orange. I agree with the Defendant’s learned counsel that they differ in many ways in terms of design and universe, survey methodology and objectives. They are not comparable and certainly would not be any indicator of the correctness or falsity of one or the other. [116] It must also be noted that because Prime Survey is a customized survey as opposed to a syndicated nationwide survey, it has in its questionnaires asked question online of its respondents with respect to perceived pass-on rates. The question of those who responded as having read theSun yesterday is the question of who they think would also have read the same copy of theSun newspaper. It is a pure guesstimate and not even an estimate at that. Surely guesstimate figures cannot be relied on too much as compared to the figures in the NMI Survey which does not take into account this perceived pass-on rate. [117] The Prime Survey is not a random survey and indeed all the complaints that Sun Media’s Expert Witness Mr Nimbalkar has on the NMI 51 for its so-called lack of randomness can most appropriately be leveled at the Prime Survey 2008. It is precisely what a survey is not, though it is good to test one’s theory or hunch with respect to what one has in mind on effectiveness of one’s targeting strategy. [118] There is the related complaint that the middle to upper income segment of the population has not been sufficiently surveyed such that when projected to the population in Peninsular the end result is skewed against these target audience of Sun Media. This group do people have come to be referred to in literature as the PMEBs (Professionals, Managers, Executives and Businessmen). However the NMI as a national survey measures readership across all income levels and it does not discriminate the respondents by profession or designation. As stated, the only qualifying criteria is that a Respondent is 15 years and is a Malaysian citizen. [119] As pointed out by Nielsen, there is no official statistics or government records of any defined “middle to upper income Group” or “PMEBs” which would be necessary as a benchmark for measurement. About the only one page document produced by the Plaintiff and not a complete one at that during Submission is the income level of this category of people closest to 52 the PMEBs but then their criteria for monthly income includes more than just their salaries and includes the value of rent for their homes as well. It certainly cannot be relied on for comparison purpose. [120] Nielsen does recognize that there would always be the “harder to reach” respondents given their social economic status (“SES”) or demographics. That is a perennial problem for all Survey work undertaken at a national level for all income groups and across different dwellings and here the problem arises from the strict protocols where security is concerned for those who stay in condominiums and gated and guarded communities. [121] Mr Nimbalkar, Sun Media’s Expert, took umbrage with Nielsen’s Survey Methodology and Design here as he concluded that there was an over sampling of the lower income group and retirees and an under sampling of the higher income group and the PEMBs. Nielsen acknowledged these challenges in a national survey compounded further by the fact that listing in the White Pages phone directory has dropped by more than 50% in 2007 and if I may say, with greater concerns now for issues of privacy and contact numbers and even emails. 53 [122] No viable alternative methods have been proffered by Mr Nimbalkar and even if some hybrid method of pure internet Survey is conducted, again one cannot run away from the fact that it may then well lack the element of randomness as those who would to respond are already drawn from a pool of those internet users who consents to be interviewed. As I said before no method is perfect and free from deficiencies and like all survey results, one would have to interpret it with some measure of discernment and discretion. [123] Mr Nimbalkar then said that a pure online Survey is acceptable though it might lack the randomness that he complained of with respect to the use of the PSUs. At paragraph 12.18 of Mr Nimbalkar’s third expert report (P-RER 2), he reproduced relevant excerpts from an article titled “Report of the AAPOR Task Force on Non-Probability Sampling” released in June 2013. The article acknowledged that “Sampling methods used with opt-in panels have evolved significantly over time, and, as a result, research aimed at evaluating the validity of survey estimates from these sample sources should focus on sampling methods rather than the panel themselves”. 54 [124] Mr Nimbalkar noted that “On page 13 of the AAPOR report, the task force has noted quite a few instances of use of non-probability sampling (panel) that have yielded results that are as good or better than probability based surveys when measured against an external criterion”. [125] Mr Nimbalkar was thus emboldened to conclude that “In summary, the AAPOR report says that due to difficulties faced (entry barrier, refusals etc.), costs and time, pure probability sampling is becoming more and more difficult. As a result, panels have become very popular. A panel study could be used for population inferences provided it is corrected for biases either prior to the survey (e.g.by using quotas) or after the survey (e.g.by using weighting)”. [126] I must say that in countries like Hong Kong and New Zealand where internet penetration is high and where privacy issues are of prime consideration, an online interview might be the next best alternative though there would still be the problem of insufficient randomness and with that the problem of how representative it would be with the perennial problem of biases that have to be addressed from a pre-selected list of internet addresses. There is also no known directory of internet users and where one breaks down in dwelling units and location, as well as ethnicity and 55 income groupings, there will be many other media owners that would have a bone to pick with Nielsen! [127] Mr Andrew Whitney, DW5, Nielsen’s regional measurement science leader for Media, confirmed that internet survey has been used by Nielsen in Hong Kong for about 10 years already and that Nielsen will be doing that in 2017. However he was quite satisfied that the booster method has tackled this problem of sedimentation and that the problem is offset or adjusted by the booster. For the period 2006 to 2010 Mr Whitney had overall responsibility for the sample design and weighting strategy applied to both the NMI survey and Prime Survey. [128] I am satisfied that this challenge has been sufficiently addressed with the use of booster samples to be achieved on a quarterly basis (See CBD 5(1) at page 839) as follows:  Main sample – 2200  Male booster – 75  Upper SES booster – 125  Location booster – 100  Total sample – 2500 quarterly (10,000 yearly) 56  The Main sample is split 65% urban and 35% rural. All booster samples are from urban areas with higher SES and are set with target achievements of this by location/ market centers and dwelling types such as bungalow/ SemiD, Condominium/ luxury apartments. [129] Miss Wong Mae Suen DW2 in Q&A 20-21 of her Witness Statement has explained the booster sample as follows:

Q:What is a booster sample? A: A booster sample is a supplementary/additional sample that is used to complement or increase the response for a particular demographic or geographic part of the population. In the case of the NMI, interviews from locations, and males from upper socio-economic status locations were likely to have lower response rates and therefore target samples for these segments were set and were referred to as the “booster sample”. The total booster sample for each NMI quarter was 300.

Q: Why did the NMI have a booster sample size of only 300 out of the total sample of 2,500 per quarter? 57 A: 300 out of the total sample of 2500 per quarter is 12%, which is roughly the incidence of upper social economic status segment of the population. The 300 booster samples were meant to off-set low response rates from the demographic segments identified above. The main and booster samples per quarter are combined before the total sample of 2,500 is weighted by gender, ethnic group, age and the 12 locations.” (emphasis added) [130] Learned counsel for Nielsen had summed up the conclusions from its expert Mr. Jonathan Jephcott (See para 10-13 D-ER) as follows: “10. In the period of 2006 to 2010, the NMI survey conformed with global best practice for a national multi-media survey. There is no evidence to suggest that NMI results for theSun for this period were deficient as a result of the survey design. ▪ TheSun’s circulation figures not comparable to NMI readership figures

TheSun’s circulation figures are not comparable to the NMI readership results reported for the paper. As detailed below, the circulation figures reported for theSun measure the number of copies 58 of the paper that delivered to fixed distribution points as determined by Sun Media, from which they can be picked up by passers-by. The readership of theSun is estimated from the results derived from the NMI, a national randomly selected sample who, when interviewed claimed to have read the paper the previous day, which is then projected to the population estimates from the Malaysian Department of Statistics.

The sampling design for the NMI conforms to statistical best practice as regards a general national survey. The disparity between the NMI readership results and theSun’s circulation figures does not mean that one of the two measures must be wrong or deficient within the context of their collection and compilation. TheSun is the only free newspaper in Peninsular Malaysia and unlike paid newspapers the circulation for theSun is the claimed average number of copies delivered (distributed) to various unmanned collection points. This number is not an indication of its readership nor is it necessarily, comparable with circulation information based on sale receipts.

TheSun’s allegation that the NMI should take into consideration theSun’s special position as a free newspaper is irrelevant, misplaced 59 and runs contrary to the concept of a national media survey. There is no recognised print readership methodology that explicitly takes into account circulation figures. Such figures are sometimes used as an informal sense check on survey based estimates but its value in this role is controversial. Even if a special method existed that would equate readership to circulation, NMI, has a duty to its subscribers to be even-handed in catering to a broad range of subscribers titles.” [131] I would agree with Mr Jephcott’s assessment and analysis of the data from NMI when compared with that of the Prime Survey where theSun’s readership is concerned. There is no evidence that Nielsen has arrived at the figures for theSun readership contrary to globally acceptable National Survey research method or that the design is faulty or that the conduct of the NMI Survey is flawed. [132] It is of course easy to criticize but no alternative constructive suggestions have been put forward which can help this Court in assessing whether there has been a breach of a duty of care on Nielsen’s methodology, design and execution of the NMI Survey, assuming for a moment that there is a duty of care to begin with. 60 [133] I find Mr Jephcott’s final conclusion as both measured and mature, mellowed by years of experience in market survey research across different jurisdictions in para 74 of D-RER: “74. TheSun’s Report attempts to challenge the quality of the NMI survey at a detailed level, not by reference to published research standards but rather by appeal to the author’s experience whilst operating the IRS on behalf of Hansa Research. Where technical opinions are offered, they are largely based on unsubstantiated assertion.

As stated in the body of this Rebuttal Report, the premise of theSun’s Report is compromised by significant misunderstandings of the NMI methodology and in particular by its treatment of the ethnic stratification of PSUs.

The Report speculates in several places that the NMI results are tarnished by corrupt influence from competitive publishers who have manipulated the survey’s fieldworkers. There is no evidence to support this speculation. Indeed, in my opinion the probability of this happening and not being caught is negligible given the QC procedures employed by Nielsen across the survey stages. Whilst 61 such behaviour as suggested by theSun’s expert may take place in some markets, I have never experienced this and, if there are genuine concerns on this front I would have to believe that they should be investigated and substantiated. I do not believe that claims of likely corruption without some serious evidence are something that should be included in an expert report.

TheSun has made much of the size of the NMI readership estimates compared to the Malaysian ABC estimates of distribution/circulation. In theSun’s Report, this argument has focussed on the RPC statistic which because of its apparent range across titles, is offered as supposedly clear evidence of errors in the NMI survey.

As I have stated above and in my Report dated 24.5.2016, the greater part of published research in this area shows claimed circulation to be an unreliable correlate of readership and therefore the RPC is not a robust concept. No attempt is made in theSun’s Report to either justify the quality of neither the claimed circulation estimates nor more relevantly to justify the 62 claimed circulation figures for theSun involving an untested method for this particular title.

The final issue of note relates to theSun’s expert’s contention that the Prime survey both validates theSun’s claim of under-reporting and highlights a superior method of measuring readership in Malaysia. I believe both points have been demonstrated in this Rebuttal to be false and misleading.

For my part, I would stand by my contention that NMI complies with what was recognised as best practice in the period 2006 to 2010 and that there is no evidence to suggest that the NMI results for theSun for the relevant period were deficient as a result of the survey design.” (emphasis added) [134] Perhaps the best evidence of the fact that the NMI is generally reliable is provided by the evidence of Mr Prashun Dutt, a marketing specialist with Sun Media, who admitted rather sheepishly that he had used the many favorable aspects of the NMI Survey results in a promotional blitz undertaken by him during the period of 2007-2008 in his PowerPoint presentation to their stakeholders. 63 [135] As a marketing consultant then, he compiled and prepared a power-point presentation by analysing the NMI data (a document disclosed at trial), among others, and the following was used by Sun Media in its advertising presentation in December 2007 (See EIC of Mr. Prashun Dutt PW6-WS Q&A 4-5): “Pg.6 (NMI Q3 ’06 – Q2 ’07 vs. Q1- Q4 ’06) Readership Growth (AIR/ Read Yesterday) ▪ Sun is the ONLY English Newspaper which has registered GROWTH in Total Readership (24%), especially in the critical KL/ PJ area (28%) and in the White Collar/ Affluent Sector (20%) Pg. 7 (NMI Q3 ’06 – Q2 ’07 vs. Q1- Q4 ’06) Readership Growth (Read Past Week) ▪ Sun has also registered significantly Higher Growth in Past Week Readership (38%) -> from 371,000 to 512,000 Pg.9 (NMI Q3 ’06 – Q2 ’07) Readership Profile (Income) 64 ▪ NMI reconfirms that readers of the Sun have a Higher Income Profile compared to Star, NST and the Average of Any English Daily. ▪ 76.2% of the Sun’s readers reflect RM3,000+ monthly income,

54.6% reflect RM5,000 and 38.5% RM8,000+ monthly income. Obviously Sun is not attracting “lower income” free-loaders ▪ Sun Readers reflect much HIGHER AVERAGE INCOME and would thus have HIGHER PURCHASING POWER. Pg.10 (NMI Q3 ’06 – Q2 ’07) Readership Profile (University Education) ▪ NMI also reflects that readers of the Sun have a HIGHER EDUCATION PROFILE compared to the STAR, NST and the Average for Any English Daily. (NST and M Mail are even lower than the English Daily Average) ▪ 57% of the Sun’s readers have University education, compared to 50% of Star readers. ▪ SUN Readers reflect much HIGHER SOCIO-ECONOMIC PROFILE (Higher Average Income x Higher Education Profile). 65 Obviously SUN is Not attracting “lower income”/ lower SEC free-loaders. Pg. 11 (NMI Q3 ’06 – Q2 ’07) Readership Profile (Urban Bias) ▪ NMI reflects that >87% of THE SUN Readership is concentrated in the Large Urban Centres. The Sun there has a significantly higher URBAN BIASED readership compared to STAR, NST and the Average for Any English Daily. ▪ 81% of the Sun’s readership is CONCENTRATED in the KL/ PJ area – Malaysia’s largest urban agglomeration which also generates >75% of the country’s consumer expenditure. Pg. 21 (NMI Q3 ’06 – Q2 ’07) Readership Profile (Shopping Habits) ▪ The Sun’s Urban Bias Readership and Married ‘family-profile’ also influences Shopping Habits. ▪ >85.3% of SUN readers shop in Hypermarkets and nearly 95% visit Shopping Malls. 66 ▪ This is a significantly higher percentage compared to STAR, NST, Malay Mail and the Average for Any English Daily. ▪ The Sun’s readers are thus most likely to be purchasers of premium brands, lifestyle products, clothes and fashion items. Pg.22 (NMI Q3’ 06 – Q2 ’07) Readership Profile (Decision Makers RM5K+) ▪ THE SUN Reflects the highest percentage of readers who are Decision-Makers earning RM5,000+. ▪ This is a significantly higher percentage compared to STAR, NST, Malay Mail and the Average for Any English Daily. ▪ SUN readers are thus most likely to be purchasers of premium brands, lifestyle products, clothes and fashion items. Pg.23 (NMI Q3’ 06 – Q2 ’07) Product Ownership (PC/ Laptop/ Notebook) ▪ The Sun’s readers reflect the highest percentage ownership of Computers (PC/Laptop/Notebook). 67 ▪ This is a higher percentage compared to STAR, NST, Malay mail and the Average for Any English Daily. ▪ The Sun’s readers are most like to be purchasers of premium brands, lifestyle products, clothes and fashion items.” (emphasis added) [136] Mr. Prashun Dutt admitted that this promotional exercise, aided and augmented by the NMI was beneficial to the revenues of the Sun newspaper (See EIC of Mr. Prashun Dutt PW6-WS Q&A 5A). He was candid in confessing that under cross-examination: “5A. Q: What was the effect of these presentations? A: With those presentations, Sun Media was able to provide its clients/ advertisers and advertising and media specialist agencies a better understanding of the true profile of theSun newspaper’s reach and relevance in the market. Ultimately, this resulted in Sun Media’s advertisement revenue from theSun.” [137] From 2011 onwards Nielsen stopped reporting on theSun newspaper. There was no need to for a media company that had taken the uncompromising stand of Nielsen had persistently under represented its readership to its detriment and loss. 68 [138] On or about 2014 or earlier Sun Media then engaged another survey company known as IPSOS to carry out a customised survey for their own marketing purposes. They went on a huge marketing blitz with the IPSOS Survey 2014 by full page advertisements in the Sun since year end of 2014 with their own claimed readership and circulation information (See DBD 9 at pg.99-100). Sun Media is perfectly entitled to do so and leverage every advantage to the fullest as it owns theSun and so is in a position to mould public perception or misperception of its readership and the efficacy of its reach to the affluent and high-income earners. [139] The point that Nielsen wanted to drive home was that Sun Media had not relied on the NMI Survey during the period in contention, to their detriment or loss. [140] It is said that a spring cannot produce both fresh water and salt water at the same time; having being derived from the same source! Neither can a salt spring produce fresh water. It is unlike the human mouth out of which may pour forth both praises and curses. The fact that Sun Media was able to extract valuable and favorable survey data from the NMI and used them in their marketing in 2007-2008 when it suited them, speaks volume on the credibility of the NMI Survey. The fact that the pool of NMI Data when used 69 with the IMS software allows such a detailed drilling that distills useful information to help Sun Media discern trends and discover correlation between their product and the people who consume them, would underscore the integrity of the NMI Data. [141] Assuming for a moment that there is a duty of care, I am constrained to conclude and indeed convinced that there is no evidence before the Court to show that there has been a breach of that duty. Whether the Damage here is foreseeable from the point of liability [142] I should also consider foreseeability with respect to liability only and not from the point of view of assessment of damages as parties had agreed that this trial is first to determine liability and only after liability has been determined then should a further date be fixed for assessment of damages. Even assuming that there is a breach, the damage caused here is too remote. [143] From the evidence adduced the Plaintiff’s revenue has been consistent throughout the years in question. PW1 Mr Chan Kian Seng the Managing Director of Sun Media offered not an iota of evidence to show that Sun Media’s advertising revenue has been impaired by the readership data of the NMI. His evidence taken as a whole is that theSun’s advertising 70 has been driven by factors such as circulation, rate increase or marketing strategy. As pointed out by Nielsen when it stopped publishing theSun’s readership information since 2011, Sun Media’s still maintained its revenue trend of about RM50million from 2011 to 2015. [144] As stated it is the subscribers who would suffer if at all it can be shown that they had spent unnecessarily on advertising with Sun Media’s competitors instead of with theSun. Advices are also drawn, as we hear, from the evidence from the survey results of other sources. There are a myriad of factors influencing one’s decision to advertise and one must of course take into consideration clients’ own preferences and the advertising rates and charges and the target audience of one’s products and services and not just circulation and readership. [145] No evidence has been led as to how the advertisers have lost in terms of revenue or profit which is attributable to NMI Survey or that the Plaintiff has suffered which is linked to the NMI Data or results in question. [146] Sun Media’s argument is that based on the evidence of the 2 marketing witnesses called by them, there is a correlation between where the media agencies would recommend their clients to advertise and the results of the NMI Survey. To put it more bluntly Sun Media’s complaint is 71 that because Nielsen had under-represented the readership of theSun, the natural consequence is that advertising agencies and advertisers would prefer to advertise in the Star and the NST. [147] However PW4 Mr Mirza Mohamad whilst wanting the Court to believe that the NMI was have an impact on theSun’s business potential, finally admitted under cross-examination that it is the circulation figures that have a role to play in advertising agencies deciding the strategy for their client. [148] He confessed under cross-examination that he had once written in a publication when he was the Chairman of the Audit Committee of the ABC as follows: “ABC Media has championed industry issues concerning media data and remains the sole supervisory body that upholds standardisation, accountability and veracity of circulation figures. In short, they provide the ultimate yardstick which forms the basis of planning and buying media for advertisers and the starting point of publishers’ sales pictures.” (emphasis added) [149] No advertising agencies or advertisers have come forward to say on behalf of the Plaintiff, Sun Media that they had actually made a wrong decision to advertise in the Star or NST instead of theSun and that had 72 there been no under-representation they would have recommended more advertisements to be placed in theSun. I do not think advertising agencies are so powerful to decide on where to advertise without the concurrence of their clients. All clients have their advertising budget and depending on the target audience with respect to consumption of products and services they may well spread their advertising ringgit amongst the media that they perceive would help them reach their target audience more effectively bearing in mind there is also a host of competing print and digital media where advertising eyeball is concerned. The clients themselves would have their way of tracking their prospects with respect to from which source they had heard of the products or services marketed. [150] I would say that in a competitive industry like the advertising industry, the advertising agencies are people with their ears on the ground and that they are discerning enough, relying on both statistical method and their own perception, with respect to where they should put their advertising money. [151] In fact another media marketing specialist called by the Plaintiff, PW5 Faraz Khan, admitted that every advertising agency would have their own proprietary research to rely on in advising clients including his own. When 73 asked what materials he would refer to and rely on after Nielsen stopped publishing data on theSun since 2011, he confirmed that he had used his own research and the circulation information. [152] If where they had put their money is not yielding results, they would be prepared to try other media. One must not forget to mention that different newspapers would differentiate themselves differently where comparative and competitive advantage is concerned; where rates, coverage, circulation and target audience are concerned, not to mention that their own marketing specialists and staff too would play a key role in grossing in the advertising revenue. [153] Whilst the Star and NST are published every day theSun has no publication on Saturday and Sunday and depending on perception of the advertisers and advertising agencies, some might think that being non-working days the average reader would have more time to linger longer with the Saturday and Sunday edition of the dailies. [154] There are just so many factors and variables that it would not be reasonable to say as a media company that one has suffered a loss of advertising revenue merely because of under-representation of the affluent 74 and with that the relative high-spender who would have a deeper pocket for all new-fangled products and services that may be rolled out. [155] It would not be unfair to extrapolate that high income earners would not be that affected where spending on print media is concerned and either their company and their home would perhaps have a paid subscription of a national newspaper. While they may read theSun and I have no doubt that they would, they will also be the likelihood of them reading one or both of the 2 other newspapers in the Star and the NST if they are English educated. [156] It is true that perception is reality until that reality is changed and no greater power and control can be given to Survey companies and their Survey results than the one controlling the advertising budget would want to give them. Here we have to admit that the one controlling the purse would have the final say and they are often surrounded by many specialists and experts who are not easily beguiled by Survey Reports. [157] I would agree with Nielsen that the factors affecting Sun Media’s advertising revenue were mainly due to its own activities and business decisions. It has little or nothing to do with the NMI or NMI data. This is proved by Sun Media’s own evidence and revenue trend. It follows that its 75 claim for loss of profit attributable to the NMI Data to be far too remote and wholly unsustainable. [158] The Plaintiff’s claim for negligence is thus dismissed for the reasons given above. Whether there is publication of the NMI Survey readership result (the Offending Publication) [159] I now move to consider Sun Media’s claim for defamation. [160] The question is whether there is publication in the format for which the claim is made. I understand that these results are extractable from the pool of NMI Data made electronically available to the subscribers depending on the package subscribed. The software provided by Nielsen to its subscribers is called the IMS software. [161] By inputting the necessary parameters one would be able to generate the permutations that one is interested in. However it must be pointed out that these are cold data with no commentaries. The circulation figures derived from ABC of the various newspapers are not available in the NMI Data. 76 [162] In para 11 of the ASOC, Sun Media set out the Offending Publication which they had extracted from the NMI Survey and NMI Data, albeit without the proper authorization of Nielsen because Sun Media was not a subscriber for the period of 2006-2010 in question, as follows: 77 78 79 [163] Nielsen’s argument is that the NMI Data complained of are found in the IMS system. It is a data analysis program that allows the NMI subscribers to produce readable data relating to their target segments by demographics, media consumption and product usage. There is no evidence whatsoever to suggest that the NMI Data complained of was extracted or published in the pleaded manner. Nielsen’s argument is that Sun Media’s claim is doomed to fail. [164] I agree with Nielsen that this is not an ordinary case of reading of an article published in a newspaper or magazine. This is rather a set of figures found in the IMS system to only its subscribers which requires among others, cross-tabulation to derive to the data which Sun Media now alleged to be defamatory of it. [165] The issue of publication is fundamental in a claim for defamation. Nielsen argued that Sun Media has failed to prove publication of its pleaded NMI Data complained of. As the NMI Data and its extraction and cross-tabulation are only available to subscribers who have contractually agreed to the terms of access, use and the non-sharing of the Data, there is also strictly speaking no evidence before this Court on which subscribers 80 of the NMI or third party that has knowledge or read the NMI Data complained of and understood that they were defamatory of Sun Media. [166] However the law must not be interpreted anachronistically in total apathy to the advancement in technology. I am prepared to say, that by way of inference, there is publication of this kind when data are made available and extractable. Whilst there could be numerous permutations as one may have interest in, it is not too fanciful and indeed quite reasonable that a subscriber would be generally interested in the readership of those who read the 3 major English daily newspapers in Malaysia namely the Sun, Star and NST. At any rate the circulation, readership and revenue generated and reach of these media would be where the income is. [167] Learned counsel for Sun Media had referred this Court to the case of Soh Chun Seng v CTOS-emr Sdn Bhd [2012] 5 MLJ 208. I would thus agree that it is more than a fair probability that subscribers would be keen on the data as depicted in Sun Media Amended Statement of Claim and indeed a natural inference to be made though in the CTOS case, the information printed out was more direct by inputting the particulars of the person whose credit-worthiness is being searched. 81 [168] The difficulty involved in the present case is not so much that the information extracted is convoluted but that one can generate as much as one would like to with respect to the 30 over media and other product and services that have been surveyed with respect to its consumption habits and patterns and the profile of people using those products, service or reading those dailies. [169] So long as the Offending Data is extractable to all subscribers who are minded to obtain such information by way of comparing the circulation figures with respect to the Star, NST and theSun, that would be sufficient publication for otherwise one would have an artificial situation in which no information is considered published when one can extract the information by way of specific focus for purposes of comparison in a cross-tabulation. [170] That cannot be right nor reasonable just because the versatility of the NMI Data is such that it can be “manipulated” and “crunched” to yield as many permutations as one is interested in depending on one’s interest. That would be to allow new-fangled technological advancement and innovation to escape the long arm of the law at least where the claim for defamation is concerned. 82 [171] The tremendous reach of the software must go in tandem with the responsibility that comes with it not unlike the aphorism that with great power comes great responsibility, attributed often to Spider-Man in the August 1962 issue of the comic book “Amazing Fantasy”. Whether the words published are defamatory of the Plaintiff [172] I move to next consider whether the words are defamatory. I agree with learned counsel for Nielsen that to determine whether the words complained of are defamatory of Sun Media one should follow a 2-pronged test in that - a) Question of law – It is the task of the Court to determine upon proper construction of the words complained of, whether they are capable of conveying defamatory meaning as pleaded by the Plaintiff;

Only upon the satisfaction of the Court that the words complained of are capable of bearing a defamatory meaning, the Court is to make further enquiry i.e. question of fact - whether the words complained of are in fact defamatory. This is a question of fact, dependent upon the circumstances of each particular case. See the case of Chok Foo Choo @ Chok Kee Lian v The China Press Bhd [1999] 1 MLJ 371 at pg.374-377. 83 [173] In determining whether the words complained of are defamatory of the Plaintiff, this Court needs to determine the natural and ordinary meaning of the Article. What amounts to a ‘natural and ordinary’ meaning, can be seen in the case of Jones v Skelton [1963] 3 All ER 952 at pg.958: “The ordinary and natural meaning of words may be either the literal meaning or it may be an implied or inferred or an indirect meaning: any meaning that does not require the support of extrinsic facts passing beyond general knowledge but is a meaning which is capable of being detected in the language used can be a part of the ordinary and natural meaning of words...” (emphasis added) [174] I agree with Nielsen that the Offending Material does not in any way convey the meaning of theSun being inferior in quality or coverage of news or that it is no longer popular, reliable and cannot be relied upon for advertising purposes. The NMI figures do not refer, indicate or mention poor readership or deterioration of the readership of theSun newspaper as pleaded by Sun Media in para 13 a), h) and i) of the Amended Statement of Claim (“ASOC”). [175] Neither does it imply or suggest that theSun has low pass-on rate or popularity/reliability of theSun for advertising purpose as alleged in para 13 84 c) of the ASOC. The NMI Data does not measure circulation figures of theSun and so this defeats the allegation of Sun Media in para 13e) and l) of the ASOC. Much less does it refer or imply the incompetence of theSun as a newspaper business as alleged in para 13f), g) and k) of the ASOC. In fact the Managing Director of Sun Media PW1 Mr Chan Kien Seng coyly conceded this though not without some reluctance during cross-examination. [176] I would agree with Nielsen that the NMI Data complained of does not disparage theSun as a newspaper or lower its reputation and a reading of the data does not suggest the defamatory meanings as alleged by Sun Media. It is my finding as set out below that the words which are purely figures here are not capable of conveying a defamatory meaning. [177] As I said before, these are cold data shorn of any emotive elements. It bears no description and no adjectives or adverbs are employed. It is not being unlike one who is listed as last in class. Somebody has to be last and the class might well consists of geniuses in the making. It is just a tabulation of marks. So too in this case, it is a comparison that one may draw on the number of readers or readership amongst the 3 main English newspapers and nothing of any comments, whether it be with respect to 85 why some respondents had preferred one newspaper to the other or certainly nothing to do with one’s ability or the lack of it in managing a media company or business. [178] I do appreciate that under defamatory material is given a wide definition under Section 2 of the Defamation Act 1957 to encompass any material capable of conveying a meaning. Section 2 defines “words” to include “pictures, visual images, gestures and other methods of signifying meaning”. [179] Section 3 of the Interpretation Acts 1948 and 1967 defines the meaning of “words” to include “figures and symbols”. Speaking of the section in the context of admissibility of short messaging system text messages in one handphone, Suriyadi FCJ said in Yam Kong Seng & Anor v Yee Weng Kai [2014] 4 MLJ 478 at 489 as follows: “[25] ... Parliament is not unmindful of the march of time as s. 3 of the Interpretation Acts 1948 and 1967 even legislated that 'writing' or 'written' includes electronic storage or transmission or any other method of recording information or fixing information in a form capable of being preserved... ” 86 [180] Whilst figures bundled together with words may be defamatory, they seldom are standing alone. Whilst they allow for comparison to be made they are not caustic nor condemnatory. Whilst they allow conclusion to be drawn they leave that to the subscriber to form their own opinion for they are not in the business of promoting or demoting anyone. [181] Whenever any comparison is done, a subscriber extracting the relevant Data would of course be able to conclude with respect to readership number how one is compared to the other two. One may say that readership is highest in the Star, followed by the NST and then theSun. That is not to say that theSun comes in last because there is no ranking that comes with the Data extracted and what one has is the raw data. The Data that is extractable may potentially be used for purposes of comparing one media with another. The conclusion is left to the subscriber to form and none has been called bearing in mind that the results of the NMI Survey are only available to subscribers only. [182] I would say that the results generated consisting purely of figures are in the context of this case not defamatory of Sun Media or that any innuendo might be arrived at that is of a defamatory nature. In fact with respect to this curiosity of the Sun Media on the people’s perception of the 87 quality of its contents and other reasons why some readers would prefer other publications are precisely what a Prime Survey would do and that too have been captured in the Prime Survey 2008 results and findings. [183] Nielsen’s submission is that contrary to the purported defamatory meanings ascribed by Sun Media, a reading of the NMI Data complained of instead bear the alternative meanings pleaded by Nielsen at para 12A to 12D of the Amended Defence, that is the NMI Data complained of provides the –

Estimated readership estimates for Sun Media, the Star and NST for period of January to December 2006 and July 2006 to June 2007;

Estimated readership estimates for Sun Media, the Star and NST in KL/PJ for the period of January to December 2006 and June 2006 to July 2007; and

Cross tabulation of the NMI results for the period of July 2007 to June 2008, July 2008 to June 2009 and July 2009 to June 2010. [184] I agree with Nielsen’s submission that they have justified the alternative meanings as set out above, as required to establish the Lucas-Box plea of justification. 88 [185] In the Federal Court case of Syarikat Bekalan Air Selangor Sdn Bhd v Tony Pua Kiam Wee [2015] 6 MLJ 187 at pg.209 in adopting the case of Lucas-Box observed as follows: “[51] In substance, the Lucas-Box plea of justification as decided by the English Court of Appeal is as follows:

If a plaintiff, in its defamation pleadings, gives a natural and ordinary meaning to the impugned words, the defendant may then rely on stating in his defence what he alleged was the natural and ordinary meaning of the words complained of; and

A defendant in defamation proceedings who wishes to rely on a plea of justification must make clear in the particulars of justification the case which he is seeking to set up and must accordingly state clearly and explicitly the meaning which he seeks to justify” [186] Sun Media had also pleaded innuendo. However none of the particulars of innuendo pleaded by Sun Media has been proved in Court. [187] In Lewis v Daily Telegraph Ltd [1964] A.C. 234 at pg.264 it was held that a plaintiff may plead an innuendo in establishing that because 89 there were extrinsic facts which were known to readers of the words, such readers would be reasonably induced to understand the words in a defamatory sense which went beyond or which altered their natural and ordinary meaning [188] I agree with Nielsen that without any evidence from Sun Media to prove that the subscribers or any third party has knowledge of any extrinsic facts pleaded by Sun Media in para 13A of the ASOC, there can be no innuendo meanings ascribed to the NMI Data complained of. [189] I state that in the absence of such a critical evidence before this Court, there is no reason that this Court should allow or even entertain the possibility of an innuendo meaning to the NMI Data complained of. Whether the Defence of Qualified Privilege is available to Nielsen [190] Assuming that the words were defamatory, I hold that the defence of qualified privilege would apply in this case. [191] I agree with learned counsel for Nielsen that the following facts as established by Nielsen form the basis for the defence of qualified privilege– 90

Nielsen is an established information and measurement company, part of a global group of companies headquartered in the United States with a presence in approximately 100 countries;

Nielsen delivers media and marketing information and analytics about what consumers watch and read and what consumers buy in

The NMI Survey is a long-standing continuous Survey conducted by Nielsen since 1968. The results of the NMI Survey are currently reported twice a year on a 12 month basis for the periods of

The NMI Survey is a wide-ranging Survey and estimates among others, the readership of newspapers, viewership of free-to-air and satellite TV channels, listenership of radio channels and the product/service consumption habits of the general population in

The NMI Survey applied the best practice methodologies in data collection, sampling procedures, quality control and weighting to estimate necessary information; 91

The NMI data are derived from sample-based Surveys. The NMI Survey uses an annual sample of 10,000 random individuals aged 15 years and above in Peninsular Malaysia to collate the relevant information for this Survey;

Nielsen only makes the NMI Survey findings available to its clients who are subscribers of the Survey;

The NMI Data complained of was part of the NMI Survey findings which contained information that was of interest to the subscribers of the NMI Survey at that material time.

Nielsen was therefore under a legal duty to furnish and make available the NMI data to its subscribers in accordance with the Nielsen Agreement entered into between Nielsen and its subscribers;

The subscribers to the NMI Survey had a similar and corresponding interest to receive and be informed of the findings of the NMI Survey; and

The NMI Survey was not targeted at any particular newspaper publication and reported only estimates of among others, the readership estimates for the newspapers in Peninsular Malaysia for 92 the material period. In fact, the NMI Surveys no less than 23 newspapers! [192] The circumstance under which the defence of qualified privilege may be invoked has been considered in Adam v Ward [1917] AC 309 at pg.318 & 334, where it was held that the publication of a letter in the British and Colonial Press was made on an occasion of qualified privilege when it was sent by the Army Council to protect an army officer who had been falsely attacked in Parliament. The defence of qualified privilege is available under the following circumstance: “...a privileged occasion is, in reference to qualified privilege, an occasion where the person who makes a communication has an interest or a duty, legal, social, or moral, to make it to the person to whom it is made, and the person to whom it is so made has a corresponding interest or duty to receive it. This reciprocity is essential.” [193] For a case more directly relevant in the context of publication to subscribers, learned counsel for Nielsen cited the case of Rex Bashford v Information Australia (Newsletters) Pty Ltd [2004] HCA 5 at pg.9-11 93 where in holding that the defence of qualified privilege is available the Court observed as follows: “...What set the respondent’s Bulletin apart from some other paid publications was the narrow focus of both its subject matter and its readership. Because its subscribers were only those responsible for occupational health and safety matters, and because it dealt only with those matters, there was that reciprocity of duty or interest between maker and recipient which attracted qualified privilege.” (emphasis added) [194] In order to defeat this defence of qualified privilege, Sun Media must establish malice on the part of Nielsen, which malice must be expressed or actual. As support for this proposition the case of Hoe Thean Sun & Anor v Lim Tee Keng [1999] 3 MLJ 138 at pg.142 was referred to where it was held that the onus lies on the plaintiff to prove malice and not the defendant to his bona fides or absence of malice. When the occasion is privileged, the bona fides of the defendant is always presumed. [195] I agree with Nielsen that in this case, whether in pleadings or throughout the course of the trial, Sun Media has failed to discharge the 94 onus to prove malice in para 16C of the ASOC. There has been absolutely no intrinsic or extraneous evidence of actual malice against Nielsen. [196] In fact learned counsel for Sun Media accepts the fact that there are results extractable from the NMI Survey that are in favour of theSun in its presentation in 2007 to its stakeholders including advertising agencies and advertisers. Some of these results are for instance steady growth of readership in the KL/PJ area, higher income profile, urban concentration, university education, decision makers and laptop owners as part of its readership profile. The NMI Data were used by Sun Media’s own witness, Mr. Prashun Dutt (PW-6) in his power point presentation to Sun Media’s clients and advertisers in December 2007 (See PBD 7 pg.1–74). [197] This fact was also conceded by Sun Media’s witness, Mr. Prashun Dutt during his cross-examination by Mr Robert Lazar (“RBL”) (See Notes of Proceedings 27.7.2016 pg. 38, line 22-36 & pg. 39 line 1-31) as follows: “RBL Now, Question 5 in your witness statement. You were asked this question. ‘Can you share some of your critical findings and summarise the outcome of your efforts during October, 2007 to June, 2008?’ 95 So this would be the period before the findings of the Prime Survey. PRASHUN Before the Prime was activated for commission. Ok. And then you say in your answer that you then collated information from several available sources and databases including Nielsen’s NMI. Yes? Right. And then, in the third paragraph, you prepared a PowerPoint presentation and this was presented to clients and advertisers. Clients, advertisers, advertising and media specialist agencies in December, 2007. PRASHUN That is true Sir. Yes ok. So this is the marketing part of your - PRASHUN theSun’s marketing. Yes. theSun’s marketing. Ok, fine. So you prepared the PowerPoint presentation. PRASHUN I also did the presentation. 96 You did the presentation. PRASHUN At the Hilton. Ok fine. And then over the page, you then say that the findings from these sources including the NMI. Including the NMI, you summarised it in five sub-segments. Which you then set out at page 11. PRASHUN Yes Sir.” (emphasis added) [198] As a result of the 2007 power point presentation, it was conceded by Mr Prashun Dutt, who exuded no exhilaration that would normally accompany such a positive disclosure, that Sun Media’s advertising revenue had increased. See Notes of Proceedings on 27.7.2016 pg.47 line 20-36 & pg.48, line 1-17 as follows: “RBL And then you summarise this with your answer in Question 5A. As a result of this presentation, you say this resulted in an increase in theSun Media’s advertising revenue. That is your evidence, Mr Prashun. 97 PRASHUN Yes there was an increase. There was an increase. PRASHUN A little bit, not much but there was an increase. Ok, right. Now, we haven't got the final details on the advertising income. But that is something that - PRASHUN You can use the ADEX and use other factors. We did try and do that. PRASHUN There was some increase, yes. We will get the evidence from the right people. PRASHUN No, but your statement is right. There was some increase. There was, yes. No, it’s your evidence. PRASHUN No, there was some increase. Yes, right. Ok. PRASHUN Let me confirm that. Yes. 98 PRASHUN As the consequence of using all this.” (emphasis added) [199] I accept the submission of learned counsel for Sun Media that it does not quite matter that there may be 10 laudatory validly drawn conclusions that are extractable from the NMI Survey but so long as there is 1 defamatory statement, that would be sufficient to found a cause of action in defamation. Whilst that may be so, I would say that in the overall context of this case, the laudatory conclusions and the positive results extractable from the NMI Survey would lend assurance to and indeed underscore the fact that there is no malice on the part of Nielsen in the publication of the NMI Survey. [200] In fact the way the software is designed is such that after the input of the raw data there is hardly any human interference that can be effected on the pool of Data and depending on the parameters that one would input there could well be as many permutations as one may have a proclivity and penchant for. The end result may well be that some may not be as favorable as one would like to have it but that is no justification for a defamation action. 99 [201] In Seray-Wurie v Charity Commission of England and Wales [2008] All ER (D) 311 at para 35 it was held that a mere assertion of malice is insufficient. [202] Halsbury’s Laws of Malaysia, 2014 Reissue at pg.155 has this helpful passage: “What is malice? Express or actual malice is ill will or spite towards the plaintiff or any indirect or improper motive in the defendant’s mind at the time of the publication, which is his sole or dominant motive for publishing the words complained of.” (emphasis added) [203] I would therefore hold that the qualified privilege is available and no malice has been shown by Sun Media on the part of Nielsen where the publication of the Data in the form extracted by Sun Media is concerned, derived as it is from a national research carried out with no particular emphasis and much less targeting any particular newspaper like theSun here. Whether Sun Media has proved, that the Offending Letter sent out by Nielsen to their clients to clarify Sun Media’s assertion that the NMI Survey is unreliable, is defamatory 100 [204] With respect to the Sun Media’s claim for defamation for the Offending Letter sent out by Nielsen to their subscribers arising out of 2 publications in theSun, learned counsel for Nielsen raised a preliminary point in that the claim by Sun Media ought to be dismissed by reason of its failure to plead with sufficient particularity the words complained of. [205] I agree with Nielsen that it is trite law that in a libel claim, it requires special treatment in so far as the drafting of the pleadings are concerned. The seminal works of Bullen & Leake & Jacobs's Precedents of Pleadings (17th Edition) at pg. 636 explained as follows: “Pleading the statement of claim Libel. The words must be set out verbatim in the statement of claim. It is not enough to set out their substance or effect (Harris v Warre (1879) 4 CPD 125 at p 127; Coffins v Jones (1955) 1 QB 564). Where the defamatory words form only part of a longer article or programme, the claimant must set out in his statement of claim only the particular passages of which he complains as being defamatory of him (DDSA Pharmaceuticals Ltd v Times Newspapers Ltd

1 QB 21, CA). Question and answer must be set out if 101 the libel is contained in both together (Bromage v Prosser

4 B & C 247).” (emphasis added) Gatley on Libel and Slander, Twelfth Edition, Sweet & Maxwell at para 26.11 pg. 989-992, deals with this strict requirement as follows: “Setting out words complained of: libel. In a libel claim the words used are material facts and they must therefore be set out verbatim in the particulars of claim, preferably in the form of a quotation: it is not enough to describe their substance, purport or effect. (emphasis added) [206] Decided cases exemplify this principle as in the case of DDSA Pharmaceuticals Ltd v Times Newspaper Ltd and another [1972] 3 All ER 417 (Court of Appeal) at pg.419 where Lord Denning MR observed as follows: “In the second place, the pleading is defective because it throws—and I use that word deliberately—on to the defendant a long article without picking out the parts said to be defamatory. Some of the article is not defamatory of anyone at all. It describes only the method of importing drugs. Other parts of the article are defamatory of some unnamed chemists, but not of the plaintiffs at all. 102 Yet other parts may be defamatory of the plaintiffs. To throw an article of that kind at the defendants and indeed at the court— without picking out the particular passages, is highly embarrassing. Master Bickford Smith put it very sensibly: 'It is tremendously embarrassing to claim the whole of the article as a libel. There is a tremendous amount of the article which is not defamatory of [the plaintiffs]. You must pick out the particular bits and rely on the rest as extrinsic or surrounding facts giving a defamatory meaning to the words.' That ruling is in accord with the practice as it has been known for many years. The plaintiffs must specify the particular parts defamatory of them. For instance in this particular case there is a reference to a 'London-based operation'. If the plaintiffs say that it means the plaintiffs, they should say so. They should insert 'meaning thereby the plaintiffs'.” (emphasis added) [207] There is merit in this preliminary point raised by Nielsen as Sun Media in its ASOC merely reproduced the entire Nielsen letter complained of and failed to sufficiently identify and plead the words purportedly 103 defamatory of Sun Media. That preliminary point alone is sufficient for this Court to dismiss Sun Media’s claim for defamation based on the 2 clarification letters sent out by Nielsen to their clients. [208] It can be seen particularly from paragraph 3 to 5 of Nielsen’s Offending Letter complained of that these merely concerned Nielsen’s explanation concerning the NMI including its methodologies as well as the ABC circulation. Para 5 merely state that the subscribers may contact the Nielsen account manager should they need additional details of the NMI methodologies. I agree that these cannot be defamatory of the Sun Media. [209] I hold that the words used were not defamatory for it merely seeks to provide its side of the story as to why there is no cause for concern on the issues raised by theSun. [210] At any rate the defence of qualified privilege would apply. Nielsen was under a duty and interest to convey the contents of the Nielsen letter complained of to its clients at that material time and the clients of Nielsen had a corresponding duty and interest to receive and be informed of the contents of the Nielsen letter complained of. [211] The Nielsen letters complained of was prompted by and in response to Sun Media’s “public attack” through its newspaper front page coverage 104 of its articles entitled “Absurd figure” in theSun newspaper dated 2.3.2011 and “Need to evolve way audiences are measured: MSA” dated 3.3.2011. [212] I would thus dismiss the Plaintiff’s claim for defamation. Whether the Plaintiff has proved malicious falsehood on the part of the Defendant in the publication of the NMI Survey [213] In Ratus Mesra Sdn Bhd v Shaik Osman Majid & Ors [1999] 3 MLJ 529 at pg.542 it was held that to establish the tort of malicious falsehood, the plaintiff has to prove that –

the defendant has published about the plaintiff words which are false;

they were published maliciously; and

special damage has followed as the direct and natural result of the publication save when section 6 of the Defamation Act applies. See Ratcliffe v Evans [1892] 2 QB 524 at pg.527-528 and Mak Khuin Weng v Melawangi Sdn Bhd [2016] 5 MLJ 314 at pg.320-

[214] Likewise Gatley in para 21.1 at pg. 811 expressed the elements of the claim as follows: 105 “At common law the claimant may maintain an action for malicious falsehood if he can show that: (1) the defendant published to third parties words which are false; (2) that they refer to the claimant or his property or his business; (3) that they were published maliciously; and (4) that special damage has followed as a direct and natural results of their publication” (emphasis is ours) [215] Looking at the evidence adduced in totality, it has not been proved the NMI Survey is false and that the same were published with malicious intent. In fact Nielsen has ceased to include theSun as part of its NMI Survey for subsequent years from 2011 onwards, thus avoiding any future possible conflict with Sun Media, who according to Nielsen, had chosen not to believe its NMI Survey when it comes to readership. [216] I do not see the decision to stop including theSun in the NMI Survey as an admission of guilt and liability but rather a decision taken to avoid being embroiled in unnecessary litigation where no amount of convincing and persuasion is possible for someone who does not believe the NMI Survey for its reliability and validity. 106 [217] In Spring v Guardian Assurance plc and others [1993] 2 All ER 273 at pg. 288, it was held that the test of malice in the tort of malicious falsehood is the same as the test in relation to the torts of libel and slander. Thus, it will not be sufficient for Sun Media to establish Nielsen’s malicious intent simply because there were inaccuracies in the NMI Data complained of. [218] Sun Media’s allegation of malice on the part of Nielsen is also untenable as it has been proved by its Media Presentation in 2007 and Mr. Prashun Dutt PW6 that there is value in the data favourable to the Sun. Sun Media has utilised this information for its marketing and to increase its advertising revenue. Mr Prashun admitted under cross-examination, though not without some reluctance, that the advertising revenue had increased. (See Notes of Proceedings on 27.7.2016 pg.47 line 20-36 & pg.48, line 1- 17) [219] As pointed out there were laudatory and favorable conclusions that could be drawn from the NMI Survey which Sun Media had sought to harness in its Media Presentation in 2007 to its stakeholders and such Survey results is hardly the handiwork of one who is bent on spreading malicious falsehood. It is when one chooses to cherry-pick the results that 107 one becomes guilty of skewing the results with the danger of distorting the big picture. [220] What is more is that Sun Media has failed to lead any evidence as to why Nielsen’s NMI Survey methodology is skewed in favour of the Star or NST and to the specific detriment of Sun Media. There is simply no reason for any Survey error to coalesce and as it were concentrate on theSun newspaper as one would expect any deficiencies in the Survey Methodology to cut right across the board. Even if there are errors leading to inaccuracies in the NMI Survey that would not be sufficient to establish malicious intent. [221] The expert witness of Nielsen, Mr. Jonathan Jephcott had given evidence that all survey measurements are by their very nature “estimates”, and are subject to errors and variations of several types including selection bias. The objective of survey-based research is not one of pretending to deliver absolute accuracy but of maximizing relative accuracy. Surveys are not meant to be absolute truths (See para 38-39 D-ER). [222] Learned counsel for Nielsen was careful to highlight that it was never Nielsen’s case that the NMI Survey results is absolutely accurate. Even from the inception through the NMI Agreement, Nielsen has qualified the 108 NMI Survey results to be estimates, and subject to statistical errors. This very fact is known to Sun Media and it was expressly acknowledged by its own letter dated 22.02.2001 stating that “We understand surveys are like ‘streetlights’ and certainly not microscope.” (See Sun Media’s letter dated 22.2.2001 CBD1 at pg.1) [223] As I had earlier found that there is no malice proved that would defeat the defence of qualified privilege, that same absence of malice would apply equally to defeat Sun Media’s claim for malicious falsehood. See Ratus Mesra Sdn Bhd v Shaik Osman Majid & Ors [1999] 3 MLJ 529 at pg. where it was observed as follows: “Since I have found against the plaintiff in his claim in libel, logically therefore I ought to dismiss this claim for malicious falsehood”. See also Raub Australian Gold Mining Sdn Bhd v Mkini Dotcom Sdn Bhd & Ors [2016] 12 MLJ 470 at para [31]. Whether the Plaintiff’s publication of the 2 Offending Articles on the Defendant’s NMI Survey is defamatory [224] With respect to Nielsen’s counterclaim for the 2 Offending Articles published in theSun by Sun Media, I hold that the words used, though expressed in strong language and even perhaps a condemnatory tone, is 109 not so caustic as to be defamatory of Nielsen. Nielsen as a survey company must surely be made of tougher material and that criticisms should be welcomed as being good and positive for the industry as a whole. [225] It would not be unfair to say that this is a case where even if “you are right it does not mean that I am wrong.” Where research from survey results are concerned there must be a free flow of arguments for and against that should be allowed both to promote intellectual rigor and resilience as well as to expose weakness and highlight areas of improvement. [226] No survey result especially a national survey published to subscribers only, with clear statements of limitations and disclaimer, is ever perfect and at the end of the day the consumer of the survey results would be the best judge of the reliability and validity of it. [227] I would also dismiss Nielsen’s Counterclaim for defamation. [228] As there is no appeal by Nielsen on the dismissal of their Counterclaim, I shall not labour the matter any further. 110 Pronouncement [229] To recapitulate, the whole of the Plaintiff’s Claims in Negligence, Defamation and Malicious Falsehood are dismissed with costs for the reasons given above. The Defendant’s Counterclaim for defamation is also dismissed. [230] It remains for me to decide on costs. Based on the submission prepared by both sides on the kind of quantum of out-of-pocket expenses spent by the Defendant and also the length of trial, the novelty of the issues and the various interlocutory applications and the number of witnesses and having in mind the seniority of counsel and the fact that this is not a case that can be considered a walk in the park but rather complicated, I would order costs to follow the event and award a single costs covering both getting-up and out-of-pocket expenses of RM600,000.00 to be paid by the Plaintiff to the Defendant and allocatur as provided for in the Rules of Court 2012 to be paid to Court before the party who prepares the order for costs may extract it. 111 [231] I would like to put on record my gratitude to all the counsel who have most ably conducted this trial and also the scholarship that has gone into the various of submissions and further submissions and clarification. Dated: 26 October 2017. Sgd Y.A. LEE SWEE SENG Judge High Court Malaya Kuala Lumpur For the Plaintiff : Malik Imtiaz together with Clinton Tan (Messrs Thomas Philip) For the Defendant : Robert Lazar together with Yee Mei Ken and Teh Soo Jin (Messrs Shearn Delamore & Co) Date of decision: 27 February 2017