Tobacco Documents Online

BAT BI~ITISI-I-AMEEICAN TOBACCO COMPANY LIMITED )~/ TO D. J. WOOD~ ESQ., R. & D. E. FROM DATE C. J. PENDRY 20th F ebr u_ar_Y.r~l~9 F¢_~--'--" ~,i "'~" . " ': 2 ?:F~5i~74 t t Dear Jeremy, PROJECT K~O" o° .... • .... ~ ..... °,° . "" I .... .°. .°... • ; .o. i~ : -" I attach a copy of Nigel Dix's summary of the outcome of Stage I of this project, for your information. Stage II will take place shortly but has been slightly delayed by the fact that we are looking into the possibility of using it as an opportunity to ex~plore ways of improving our product-testing methodology. I will keep you informed of progress. With kind regards Yours sincerely, C.J. Pendry. Enclosure: cc: N.W. Dix~ Esq. C:) CD O "-4 O~ BATCo document for Legal Services • Health Canada 20 May 1999

a" "4, • ° / o Q PROJECT K~J Main Points arising from the Explo_ratoryStudy • (London 4973) Introduction l' The aims of thisresearch were to explore, for the preparation of a quantified study, (i) the extent to which consumer appraisal of cigarettes of known differing characteristics aligns with in~ernal assessment by Taste & Flavour panels and (ii) the vocabulary used by consumers in describing smokin6 attributes and experiences. The research was carried out by two agencies, each conductinG a total of 18 semi-structured depth interviews among smokers of EMBASSY FILTER. This particular brand was selected for the research since it has averasetaste characterist'ics and with a market share of around 20%, its smokers could be easily located. Each respondent was asked to smoke a specially manufactured control cigarette having smoking characteristics similar to ~I~iSSY, and to compare this with either one or both of two test blends. One test blend had a similar level of irritation to the control, b~:~ t:ith an increased intensity of Impact; the other had a simi]ar intensity of Impact to the control, but with a greater level of irritations. In all other respects, the tesZcigarettes resembled the control as closely as possible. Summary of Findings The research undoubtedly confirms the opinion that consumer descriptions of their smoking experiences are vague and highly subjective. M~ny words were used to describe smoking experiencez and even where differen~ respondents used the same word, it camno~: be assumed that similar sensations were being described• However, the majority of respondents apparently found little difficulty in distinguishing between the cigarettes sz:cked in the test, which perhaps indicates thaZ differences between cisare~:t~s in terms of Impact and Ir~itaticns are of some relevance and hence that the inclusion of these characteristics within the Taste & ?iavou: ~ustifiab!e "* system i~ c A~hou~h by no means clear cut separutc patterns emerged in the descriptions of these differences, and it is ~.l~ that a quantified study could provide increased knowled<e of the ways in which consumers perceive and relate the Impact ~nd Irritation sensations. In addition to the Taste & Flavour aspects, the study has ~'eve~le- "a nt~.mb~r of other criteria upon ~.:hich consu-_,'..rs appar-~ntly b':::." -..hal;. asness.-_ent of c.~ar~ttes and it is intended that theme, be furt;.:r tested ir the quantified research, to determine their s-,:nifi..:.~.-.-.,-~ and whether they should be included in future consumer tests, LLLiO0001 BATCo document for Legal Services : Health Canada 20 May 1999

-" 2 -- P How differences bet:.,een cig,~.rettes are ~e.'.~c'ribed Before'discussing the reactions of resp6ndents to the test cigarettes individually, it m~st be stressed that in many instances there is an overlap of terms used to.describe the two separate sensations of Impact and Irritations, with common use of "strong" and its supposed opposite "mild". This appears to be the most natural and readily forthcominE ~.Jay for smokers to describe ci- garettes, and in most cases only persistent probing by the inter- viewer produced more detailed response. There was a noticeable tendency.for smokers to compare the test cigarettes with other brands, although still using the strong/mild context - "like PLAYLRS No.6, that Sort of strongness". .. a) The !moact V.~riant Although many different words and phrases are used to describe this blend, there is a recurring theme of effects'upon the throat and chest. In general, the cigarette was thought to be strong and by some respondents "strong but smooth". b) The Irritation Variant Again, many ~,:ords were used to describe this cigarette, most overlapping with d~scriptions of the Impact variant. However, whereas the latter was largely seen to affect the throat and chest, the Irritation variant• was more frequently considered to cause unpleasant effects in the mouth, particularly on the tongue. In general, the cigarette was considered to be strong, but str0n~ in this case being qualified by roughness and harshness. c) Flavour Sensations Although this research was primarily designed to study differences between cigarettes in terms of the irritant charac- teristics, it ~;as thought likely that flavour aspects would arise in the interviews and thus respondents were encouraged to discuss any flavour differences they noticed in cigarettes and how these are assessed. Firstly, it is apparent that the word flavour is not natural! used b v smokers of Vl.~:_._la cigarettes, respondents preferring • ; to discuss the "!:~ste" of a cigarette. The word taste is widely used and seems to refer to the overall smoking e~erience, rather than ~ust flavour, In facZ, it at.pears doubtful whether smokers of this ty:-e o£ product *~hink of "it ns hav~..cG "flavour" - to them a Virgin~:l cigarette ~- ,~_~es the natural tobacco taste: rather as one zi~ht ~" "-.'- of ~,nl .... ~he taste of .~.lain as opposed to flavoured yoghurt. Certainly, when describin[~ cigarettes of other than Virginia type. ~. :~-. American, Tur:<izh, respondents shot.:ed a greater a~.:areness of fi;,vour, de~cribing ~'foreign" cigarettes as havinG a ...... ~ed or ~,.~c~ flavour. c Q-/ C oD BATCo document for Legal Services • Health Canada 20 May 1999

° / ~o Tar ~nd Nicotine Aspects When rcs~ondents spontaneousiy mentioned tar and nicotine effects. they were related to the stainin~ attributes rather than to any aspects of taste, and only on probing by the interviewer did some zmokerm show a~Isreness of a relationship between tar/nicotine and "strength". Only a few respondents thought of nicotine as a type of drug and related this to the effects of stimulation or relaxation. All respondents hadheard of the tar and nicotine tables although few had actually bothered to study them. ~-I&%SSY was recollected to be "~ozewhere in the middle", but so lonE as their brand was not "at the top", smokers were not unduly ccncerned by the table. Cne or two fezale respondents had been influenced by the table to try SILK CUT but had since reverted to ~K~SSY. Overall, respondents seemed uncertain as to what tar and nicotine are and tended to =rouo then: together as "the dark, smelly stuff collecSed in the filter". Other criteria used to assess cigarettes The research also revealed a number of other apparently significa~ aspects upon which consuzers assess cigarettes and it is proposed th~ these be included in the quantified study to determine their importan~ They are: - firmness of filling - ease of dra~o - effectiveness of filter (perhaps related to ease of draw) - rate of burn - after-taste Prooosa!s for a o,:~ntitative stud.v Both Research Agencies involved in the exploratory work have been ozked to submit nro~o~als fo:" a quantitative study. 3~e objectives of this will be to determine the significance of the various aspects of consumer cigarette ap!~raiaal raised in the exploratory research, and to develop scales upon which such attributes may be rated in future product tests. 8th February, 197a. CID O C w BATCo document for Legal Services • Health Canada 20 May 1999

Q f y N.Y. nIX, ESQ., MARKET RESEARCH DEPT., HILLBANK. D lJP146E. D.J. ~OOD. .° 10th January, 1974. ., • o .o , ~. ° .° • . Project E:EW : , .,. I have now gone through in detail the reports from the two Agencies on the qualitative stage of Project EE~, and have some thoughts on how the quantitative stage might be handled. Presuming that the 18 smokers contacted by each Agency are fairly represents tire of U.K. smokers in general, it is clear that many of the adjectives commonly used by smokers when describing their impressions of a cigare@te, such as strong, harsh, rough, hot, cover a wide range of sensations. For instance the hlgh impact cigarette was described as strong, but so also was the high irritation cigarette; in fact many smokers thought the latter stronger than the former. There is also the comment in the Z.R.A. report ~hat consumers would consider French cigarettes "strong-flavoured". So the finding, in a consumer test, that one cigarette was considered too strong would really give very little information about the cigarette, and further probing would be necessary to find out in what sense it was too strong. At the same time there are clear indications, that consumers were able to notice differences between the three cigarettes, and with free choice of term8 were able to describe these differences. The high impact cigarette was seen by many as having an adverse effect on the throat, often described as a burnin~ sensation, whereas the high irritation cigarette had adverse effects on the tongue (and even the teeth with some respondents), and to a lesser extent on the throat. It is unlikely that these differences would have been revealed by the conventional type of questionnaire containing scales for strength, mildness, harshness, etc. What seems to be required is a questionnaire specially designed In order to bring out the more useful connnents that respondents were making. CD Years ago, In R. and D., we had a project called WATERL~ which attempted to relate the smoking characteristics of c~ representative flue-cured, Burley and Oriental grades with their leaf and smoke chemistry. A special questionnaire was -.~ designed for this project, and I enclose a copy because I think cD BATCo document for Legal Services • Health Canada 20 May 1999

.° / /. / the general approach, and even some of the specific questions, could be extremely useful in the quantitative stage of KEW~ For instance with minor modification question I could read: le EFFECT OH TONGUE: Consider only the sensation on the tip or front portion of the tongue. The • smoke caused a stlnglng or burnin~ sensation Very much Much etc. The most obviously relevant questions are No.l, relatinK to the tongue, and No.3, relating to the back of the throat, though some of the others might also be useful; or one might be able to construct addltlonal questions, alone the same lines, which more closely reflec~ the respondents' spontaneous comments. Anyway, I offer this as s suggestlon for you and Colin toconslder. ".- . .° / c.c. Dr. S,J. Green D.E. Conway, Esq. CD o O "-U C~D BATCo document for Legal Services • Health Canada 20 May 1999

/ .& RESEAIKCH PROPOSAL FOR PROJECT KEW ~/ September 1973 Prepared for: By BRITISH AMERICAN TOBACCO COMPANY LIMITED Marketing & Retail Analysis Ltd 178 Sha/tesbury Avenue, London WC2 Ol 836 8383/4 0 0 0 CD O0 BATCo document for Legal Services • Health Canada 20 May 1999

1, BACKGROUND In the past alternative methods of assessing and ev.aluating cigarettes by P. D. L. smoke analysis, Taste and Ylavour panels and consumer research have not always produced reconcilable results. A/though results from alternative methods have not necessarily been strictly comparable because of variations in test factors, the differences in the results have raised concern about the Taste and Flavour system as an evaluative technique and about the lack of understanding of the relationships existing between the three means of evaluation. B. A. T. therefore wish to conduct a strictly controlled project in which cigarettes with clearly noticeable differences between them in terms of Taste and Flavour evaluation are measured by both P. D. L. and consumer testing. The purpose of the consumer testing project will be: D To determine the extent to which consumer appraisal of cigarettes which have known, differing characteristics aligns with internal assessment by Taste and Flavour panels. 2. To study the vocabulary used by consumers in describing such known differences. The research will examine irritant and flavour characteristics separately and these proposals cover the first stage for irritant characteristics. O CD O O "-4 O BATCo document for Legal Services • Health Canada 20 May 1999

~Q METHOD Three blends will be manufactured to highlight impact and irritation differences. Embassy R. S.F.T. has been selected as a control brand with a wide franchise which has average characteristics lending themselves to systematic variation. The test cigarettes will however be in a King Size format because of manufacturing facilities. Three cigarette blends will then be tested: I. The 'control' cigarette based on Embassy. e Test blend I with similar irritations to the control but with a noticeably higher intensity of impact. 3. Test blend II with similar impact to control but with a noticeably higher level of irritation. Research would first use exploratory, qualitative research to prepare the ground for a second stage product placement, quantitative study. I. Exploratory Study Twenty intensive, semi-structured interviews would be conducted with Embassy smokers to provide: a detailed understanding of consumer recognition of various cigarette characteristics, especially impact and irritations. a knowledge of the language used by consumers. a guide for use in designing the questionnaire for the subsequent placement test. CZD O c~ cZD BATCo document for Legal Services • Health Canada 20 May 1999

3~ The interviews would follow a format which would be agreed "I with the client but which could be expected to be along the following lines: il establishment of respondents current and previous smoking habits. if. reasons for choice of regular brand. iii. dislikes of brands which would never be smoked. iV, Vi vi. further open probing of all characteristics which the respondent is ever aware of when evaluating a cigarette. specific further probing, where necessary, in the areas of satisfaction, irritation, strength/ mildness, hot/cool, harsh/smooth and a.rnount and type of flavour to establish comprehension and perception of these terms. further discussion with the respondent of awareness, comprehension and evaluation generally, of: = tar and nicotine content - differences in mouth, throat and nose irritations - flavour differences vii. open-ended description in the respondents own terms of the control cigarette and one of the test cigarettes. C C C ~C r'x. BATCo document for Legal Services • Health Canada 20 May 1999

1 viii. concentration o£ the r espondents attention on the perceived differences between the two cigarettes. discussion of the known differences between the two cigarettes as specified by Taste and Flavour evaluation. e C) uantitative Study Following the results o£ the qualitative study two double-rnonadic placement tests would be conducted o£: control cigarette versus test cigarette I control cigarette versus test cigarette II In each case respondents would be given 40 cigarettes of one blend at the first call and be asked to smoke only these cigarettes over the next 2/3 days. At the end of this period they would be interviewed on their opinions o£ these cigarettes before 40 cigarettes of the other blend was placed with them. At the third call Z/3 days later respondents would be interviewed on their opinions of the second cigarette and would also be asked to compare the two cigarettes. Strict control would be maintained to ensure that each cigarette was placed first an equal number o£ times. The detail o£ the interviews would be dependent on the results o£ the qualitative study but the questionnaire might be expected to cover: open-ended likes and dislikes; CD CD ,,,O BATCo document for Legal Services • Health Canada 20 May 1999

5. rating on descriptive scales of those T attributes established as being important in the qualitative study e.g. strength / rnildne s s satisfaction hot/cool flavour harsh/smoothness etc - like/dislike rating; specific differences noticed between the two cigarettes and the importance of these differences; - overall preference and reasons for preference. C~ C~ C~ ~0 BATCo document for Legal Services • Health Canada 20 May 1999

• 6. THE SAMPLE The sample would consist of male and female regular and occasional smokers of Embassy i%. S. F.T. They would be contacted by interviewers going door-to-door following a strict random route sampling procedure. On the placement tests matching of the two samples would be ensured through systematic rotation of the order of placing the cigarettes. Sample sizes are anticipated to be: 20 qualitative interviews. 200 control versus test I. 200 control versus test ~/. Alternative costs are given later for placements in: London conurbation only. London plus two other conurbations. Allowance is also made for either all or 50% of the qualitative interviews to be conducted by G W Flude. C C -,,j x,O BATCo document for Legal Services • Health Canada 20 May 1999

7, TIMING The qualitative work would take 4 weeks from the start of fleldwork to the report. The report on the placement tests would be available 6 weeks from the start of fieldwork. COST Costs, excludLug 10% VAT, would be: Z0 Qualitative interviews by G W Flude (10byGWF; 10by supervisor £400). £5OO 2x200 sample placement tests: London only London plus 2 conurbations £ I, 900 £ 2, 150 D C C C BATCo document for Legal Services • Health Canada 20 May 1999

J J 2331 PROJECT KEW October 1973 I OUTLINE OBJECTIVES AND SAMPLE The aim of this research is to investigate the vocabulary of consumer appraisal in respect of various cigarette blends with a view to relating it to data collected by a Taste and Flavour panel. This panel consists of a number of trained judges who evaluate cigarette blends in terms of 14 scientifically based characteristics centred on irritant/impact factors and flavour. The current research is restricted to the irritant/impact dimension. The successful fulfilment of the objective depends significantly on the extent to which consumers are able to verbalise and articulate the criteria by which they evaluate cigarette blends. This in turn depends on the relationship built up between interviewer and respondent. The attached interviewer guide should be regarded as a rough outline of the sort of information you might well get from respondents when they talk subjectively about c~garettes. In any one interview, all the concepts might not be mentioned and this in itself would tell us something about the way in which consumers regard particular smoking characteristics. Consumers should be encouraged wherever possible to put forward their own conceptual structures which they use (either consciously or unconsciously) to evaluate cigarette blends. Some respondents will inevitably be more able to express themselves than others. Greater patience and more effort will be required in interviews with those respondents who find it difficult to , verbalise. It is so easy for an interview with such a person to degenerate into a string of one sentence questions and answers. Under no circumstances should you prompt the respondents with words such as satisfaction or strength etc but you should find out --- O from them what they think the important characteristics of cigarettes ,~ are, how important they are relative to other characteristics, to wha~D --4 BATCo document for Legal Services • Health Canada 20 May 1999

o extent they are present in the control and test blends etc. The important question of comparison between the two types of cigarette in the test is in the description of each no__~t in which is preferred. Sample The procedure is that two packets of twenty cigarettes (one control and one test) will be left with the respondent two days before the interview. Half the sample will receive 20 Control cigarettes and 20 Test Blend 1 (high intensity of impact) and the other half will receive 20 Control cigarettes and 20 Test Blend 2 (high level of irritation). The packs will be unmarked except for a neutral code number. Half the sample should be asked to smoke the control cigarettes first and half the test blend and it should be made clear by the interviewer at the placement that the informant should try to smoke only the test cigarettes for the two days between placement and interview. The sample consists of a total of 18 regular or occasional smokers of Embassy Filter Tip Cigarettes who smoke no more than 20 cigarettes per day. The sample is subject to the following quota controls. CIGARETTES SOCIAL CLASS SEX AGE Total 18 Depth Interviews J to receive 20 control cigs. & 20 Test Blend 1 to be in social classes ABC 1 C2 DE This group should include : 5men and 4 women This group should include : 4 under 30 5 over 30 to receive 20 control cigs & 20 Test Blend 2 4 5 to be in social classes ABCI C2DE This group should include: 5 men and 4 women This group should include: 5under 30 4over 30 C C C CO BATCo document for Legal Services • Health Canada 20 May 1999

J 2331 October 1973 MIL RESEARCH LIMITED 1 & 2 Berners Street London W.I. PROJECT KEW DEPTH INTERVIEW GUIDE l. . Commence the interview with an openended question to all informants "Now what did you think of them?" In this way, it is hoped that those aspects of the cigarettes which are important.to each individual respondent will be mentioned at the very beginning. This should provide several leads which can be pursued in greater depth later in the interview. A more detailed description of the characteristics of the test cigarettes should be sought. In addition to describing the two test brands in isolation, there are several other possible approaches which the respondent may find easier. The method to be adopted should be the one which comes most naturally to the respondent. In the event of the respondent finding it difficult to describe the two test brands in isolation then the following alternative approaches may be suggested to him: i. Compare Test Brand 1 with Control 2. Compare Test Brand 1 and Control with Regular Brand 3. Compare Test Brand 1 and Control with known "strong" or "mild" brands. If the respondent has concluded his descriptions of the t test cigarettes without making any comparisons then . the interviewer should find out whether the respondent thinks they differ and if so in what respect. The respondent should be asked to describe any differences and the extent of these differences. <ZD c~ ~D BATCo document for Legal Services • Health Canada 20 May 1999

o It is important that respondents are encouraged to describe the characteristics of the cigarettes in their own words. The role of the interviewer in this instance should be to explore in detail the vocabulary that is used to express their responses on these and other criteria they themselves see as being relevant to the evaluation of a cigarette blend. . . Depending on the methods adopted by the respondent at 2 above, the interviewer should try to establish why that particular approach was adopted and whether or not it facilitated their evaluative process and if so in what way. For example consumer's responses may have a consistent basis - use of reference points such as regular brand or known strong or mild brands used - or they may be unrelated to one another. If not already mentioned, the subject of the tar and n~cotine content of cigarette smoke should be approached. Important points to cover are the extent to which respondents are aware of T & N content and the extent to which they can detect it during smoking - both in terms of their regular brand and the two test blends. Find out how important T & N content is to the evaluation of a cigarette and how the respondent would describe perceived differences. . Encourage the respondent to talk about the physical aspects of smoking - how cigarette smoke feels in the mouth, throat and nose and ascertain the importance of the part played by these •physiologicalfactors in the evaluation of the test blends. Particular attention should agai~ be paid to the means of expression used by the respondent. BATCo document for Legal Services • Health Canada 20 May 1999 O O O CO C~

% . Assess the extent to which differences in flavour are appreciated between the two test blends paying particular attention to how the respondent evaluates these differences and the words used to describe them. BATCo document for Legal Services • Health Canada 20 May 1999 C~ O O C3 C~ O

i ~ROJEC" " ' PR, O~-~SALS FOR CGI3PW.ER P.ESF-~C'~ .... . . o . ~ . ,.-....~ .... ~ ..... o.._ .... -. • , J ,.~.'. " :- "--:':e,:.n cicarettes which are kuc'.'.,~ to have ;i~--~,-:- ~ ' "" - " s...o.:~::..: c.leracteristie3 ~s dei'ir_cd %Y a subzidiery level, to e:.'.ami::.e the voc~.bul-"zZj ,.'.sod b7 consumers in dcscrihi.-.~ such differences, esDeciel!y • " ~'- dimen-~o~n "a~cd in Teste & Flavour it. 2e!~t?.o:~. to u~.e ........ nnne! assess:tents. e To undertPke P.D.L. Routine Brand ~nalysis of cigarettes differentiated in this w~v in order to ~rovide an ob~o~+~,~. ~.~,.---,. measure of ~.,~h~---~ ~iffere~ce~. B.~C;,, ~_,-0"[.., ,,-'D From time to time, instance~ have arisen in :.,hJch Ta--t~_ and F3 avour :e.zults h~ve ngt "-o~ in the di::-ectJ.or_ tkr:. • , . . i n~n5 ;_a,., ~.ee.~ ¢:~ected on ~-~',, b usi~ o~ c~;id~nn~ frov on :~hich t-:-, co,,..... ...... ~.., re;.c~-..o~n -.:hJc':: .-J.~--::t ":z-;:. " ...... ar.tie:,~ntec_..:- ".,.3end "~..,s~_!..-~ - on ~_, .~a-;-:.~ 0:" _,.. . .." ,hen,s _~.~',ve no% ; ~, borne oit. "= ..... ,,..~.~ ~.,-,ee ATpen: ~'" .). q RlO000[ BATCo document for Legal Services • Health Canada 20 May 1999

-2- attached in Appendix TT. Ho:.'e'..'er, a!the,_,,"h this e:-e"ci~: is of considerable-za!ue and "~'~ ~ - Z;,~ ~La_c,.,,~c ...... n~ture of t?e rel~.,,oDs:-_io :~--~-~ the .~..~rticular a_--.-.eszments r: • - 7".~F.':" : :" .-; " " "~ 4 ~._ -, • insi.-ht i~o the' relatfo---::~ ....... p.,~- e'-,istin~., between T/? eva].~'etions, the ~as,'~t- of P.D.L. ~-~-:~-~ ~ and tl-e result ~f (7,-: ..... :;:" "le'%~. tests, a careful.Iv controlled..~.z'o.~ec.~ should be set-u'~.~ in :~hich ci,.-..arettes that have c!e..-.~v not=c ...... differences bet-,.:een them in tez~..,s of Taste and Flavour evaluation ate measured by both P.D.L. and consumer ~ .~ ~_.....-;. =.,...':'.~ :,~ UsE,_,-':. T-ETJT,?S To improve our u.nder.ztondinq of the .na~:re of tl~e relaticn.--.-'/..p between ?~zte and F!avo::r azses,--...-..ents, P.D.L. anel~.'zis and consumer b.!e-d testinT and to provide guidance for the futuredevelopment and application of the TasSe and Flavour s::stom. LOCATIOTT It is reo,,cnS.zed that Ln :.'~zic'_~.e',;o~ r, orYet this rem-:-.-.b, t ~ re21: .... ..~ ~" ........ • takes "'~,,_..---e,. h~. ~ "- .-..c cc~:ci:'..~-ion.--- ~i].! h~ -~ ;."--.-Z ~-~'--:'~-'~__,,...._ "^.,~ ~lLnt~ r,~.-'-.e~ r.:-.:-:.. _'-,-,: ..... ",.": ~, h..'-.,:..-e limit-:.- 8~.D].J. "; : ~" . _ --~'."-¢re. " ................... arT.--.~u';..-.,~-,~.~'..:_ "o.c. ar'.,--.-cr.~ ........ for .~erf"9--T--i-'--- the re.---~?:'.*.-: e!te~.:here, r_.onethe].e-a, it is cons!~.ere.i th.~t the U.K. 4 s ~.-~'t:-"; -- " c.:-.tcble f~ t'-~ -" ~ ~ • - o~ . the foe: t::o~ the !~." .~. dc=.s not trade the:'e: ~o "1 -~'--~.--~ - ..... . Re.-~=~., rz;~ D~,.m-.rZ-.c-nz ~.-i!! b~ reac'~:v abl-~ ~;-. rp-z!r:_- cc.-.tr-.t, "-.i~-''~.. the ,~-or.c-:. 2~ C cx: c~ BATCo document for Legal Services • Health Canada 20 May 1999

-3- .. ..'- Agency exp. ertise is readily availab!e, enabling us to choose the aEency :faich eppears most ~itable. f • I It is ~roposed thnt this research be ~wlit into ~.~o entirely separate stakes, the first coverin~ the irritant characteristics of c!snrette ~moke i.e. Impact and irritations, and the second coverin~ flavour aspects. ~nese particular propcsa!s refer onlp to the irritant stage vfnich it is reconnended be completed before givin~ consideration to further research on flavour aspects. We believe the irritant characteristics should be examineC first s~nce these are ccmmonly re~arded as the most importont aspects of Taste and Flavour work ~nd are probably less complex than the flavour sensations. It is proposed that three blends be specifically manufact- ured by G.R.D.C. for the research to highliEht differeuce~ in Impact rmd irritations as follows: 1. A 'control' cicarette of widely accepted 'taste' characteristics. 2~ Q Test blend I - a ciuarette with similar irritations to the control but ~th s z:oticeably hisher intensit7 of Impact. Test blend il - a ci-arett~. ~:ith siui!nr Impact to control butw_~.~4~ a n0ticeebly hisher level of ir=-itation. it is considered necesss~: that th= control blend sl.oul& be bused upon a brand of the following specificaticns: (i) cener~llF i--no:m and "~dely ~:ccept~bie to ti'e consu-~ (!i) of 'av~rP~e tazte' cher~cteristics i.e. no extremes of irn~ct o~" i'-'ritations (iJi) :,:ill !on,~, itz?!f to s[:-te-ntic ver'.?.tien of Tmp~ct -~,: i_-ritations (iv) "-'a,--- a :.:ide-franc!'ise to faci!itnt~. !ocntion of ~'t_- c o.~_c-/.-.gr ~ If it is ncccDted thnt t'~4s research should be carried o~t ~n thc U.;[., then -_I ['- ASS'_" ~.S.F.T. m~ts thozc rcruir~nents~ and it ~_n ~:'.'= reco..-.zended thnt a ci:_-h-~c'~t~ with similar smo~:i.-.g ch-_racteristics to this brnnd he i 00001 BATCo document for Legal Services • Health Canada 20 May 1999

° -4- f . manufactured by G.R.D.C. ~s control. It ~.~ll, however, be necessary to ~roduee this in s Kin~ Size format since current ~a.-~!e nanufac~rin9 f~cilities at Group R & D. Cent'~e do not cover p_~ck~in~ of Regular Size brands. ~,en so, ~.;e do not ~egard this ss !i!:~!y to 2ffect the validity of the research or conzuuers oerce~tlon of the contwol and a!ternmtive v~riations, ~ince preliminary examination of the suokin~ characteristics of the K.S. version ~_ndicate them to be broadly the same as those of the reualar size ci.~arette. It is proposed that the research be carried out in the form of t:~o paired-comparison blind ~.ests, with Test blend I beinE cor.?~red ,:,~_th the control in one pair and Test blend iI bein~ coupe zeal ~.~ith control in the othe Within each p~ir, re.~p~nd.ents will be ~ive~ 'aO ciGsrett.es of each version (i.e. control and te~t) and as:-e& to E.no~: one for the nex~ 2-3 days end then to smoke the other for the fol!o~,,Snz 2-5 deys. After they have had an opDortun~ to smoke b.oth verzions in the pair, respondents vil! be interviewed for their op!rions of the cicarettes. Regardinz the t2~e of ' ~--" ,.H~ it is _ ~nte_ .re ..... ~ c mp!eye.~, recommended that a t~:o-stage approach be used: 20 intensive, "semi-struct~red inte~zie~:'s to be carried out in order to provide: a detailed un~ers:,an~in--, of the extent" to -..:hich consumers reco..-nize ~nd describe the vnrj.ous char~ct~.ri~ties of ct,~',_'.~Lte smoke, particuia_,-iy I~.pact and irritntions b) 8 r~,~e in dezi<nln~ the ~" " ._. c ues~orms~_re for the cfa~ntitative s t-a d..-. ~ese 20~'-~o.~ ~_rv,~o-. ,;~- should cover t~-_. following _ooints: a) BATCo document for Legal Services • Health Canada 20 May 1999 O o oo o<3

-5- b) :.,heth-,r rem'.~ondemt~ ~ "'- of .".nd ,.a,_ h.~mrd the for nicotine content of cigarette smoke and~,.."~.~:~-..~,,: .... the'; con detect then r1::r4~-_.. ¢mo~in,,': _~ ~o) ',:h~'..~..= .... they feel the t,-'o cigare'6tes th=y hove smoked rafT.- in this waD" end how they. de,oct ~:,:~..~s, c) throat and nose irri~ations and how impor.Zen~ the[: feel these ,to be. d). the extent to which respondents appreciste f!.=.vo-l_~ diffem'ences in ci~a-~ettes end how they evaluate the"r 2. q,asntit~t!v~ S~;.v To obtain sound ststistic~l ~guidmnce for ara'.,inm f_~rm conclusions from this rcsear,'.h, it is considered, ncc.es~= the results ~nd ideas rai~e~ ~t the e:.~!orntor-j s:~.-fe. ....... =~ -t Dcc.sible to outline the i:~fo~lat5o:: th'-t i- it is ~ ~.~d ths~ t:'-~ ' ~ -'~ ~-~ '.'il~ once the results of the cumlitative inter~.'ie'.:s are ]:no'-. ~D.~.e cuestio-~ire will probably cover ,,~,-~ . . s) differences noticed b~t:-'e~Jm the two products smo'..,e:~ .'=r..-i ..--.it!rude3 to t:':e:'e b) ~..~,di~ional =triP, dins! com.~:ents on t.,c two ~.-,od',,cta individu all.: c) ench F..~.-c'?,uct to "-.- rated u'~,em m ser.~es o" ~c[:ler: these to he finmlisg~ ,~fter the e:.,~....icfato.-~.- renra-..'.- "cut :;ill be of the t77.e: S:',; IF.;- E zmtis f~.ction hot/coo! t oh..-., c c o .--: --- ~..i iT: tTq~c of fi~.vou',z har~h/zToo ~h ~. ~ "~"~"~-.~; "'- ' ~ C f." ' . • ~ ..... ...... . . "." --'". .': " " 7 ,- ,'~ "' 2. "'., "" .... Y.'~ ","~""~ ~':'-~ .... " ........ " .... .~,'5 :,"* :"- ...... " ' w 0 ..,.=,.. oo BATCo document for Legal Services • Health Canada 20 May 1999

are used to end so h.~.';~ the r, aximum cl:~nce of discrinin~:t- inc between the contrcl and test version.~. For tbc. cu:.l .......... rtvc:$', "~t :~s envis~,-ed tnaz 20 re .... o._& ....... - (flO ,~cr pelF) should ~,o.. . ~de~.unte. For ~u_~.:... eu.nntJ ,t.-:t~:,:. sin:c-:: it 4 s _-~o.._~ ......... zn~t ar.. .......... ,.n of ~,. ~ -_' (i.e. a tot~l of aOO) sl-o'~Id he used: the s~m~!e You- ee.ch ~a4- ,.4 ~l "~. co~,~--o~. ~ -.,' ~,: gse ~.-:d c.lr.~- ":o as _ .......... • ........ . ..... - CO2:~ .. ..-., , to be representative of ~.~ASSY smokers and the two sanples will be .c!o.~e!:: _-:arched. AG~TCY In view of the develoT:-'.ent~! nature of th':.s rez~erch ~t " ~-~,:~-c~ .~.~--t~"-~--l:- imnort~nt that the a-:er.c:." 1S CO.,~.~ .... "- :. - .............. . used be select,-~d with caz-e. It is prdposed to obte.in quotes from .'3.]i.~.B., P.H.R, and H.O.~., since the.-,- hc::e all had eonsidez'ab!e e.,..-ferience of rezearchinF, ci':arc'Gt~.~- C¢~'~f~ L_22~ Eati..-.ated costz (m:bSect to "'~': re~.~ .... cn e:ter dizcussion wit:- egency) are as .fo!~owz: r E,...;~ ._ _ ~ ,_ Pie::earch giO0 Quentit::tiv-e..'~°---- ~-~"..:~ ~.,_ "£2900 .... ,.,.~ =.~ ~'9 Xf'.'% Cost ~f a~ ~..-:_-'. :.'t ......... ~'ot~l_ =... ~-- '-= ~- '--~~ co~: £a'_39 9I::_-' ::& C.u--!it-...~i:-e r:,----:!ts .'-' -:--.-::s P.fter fief: ,~tcT't. ...... : ~:"~ '" .... ~ "-'O after .i-:-.-! st'-.t:t. Cot.ire! :-i,:n:'.: ~,0").:') c; -nr~:t,,s 9.--.c':..-':i in ~- r.~:.: ,%,% T~ct _=,le':.: I : 9,'.D~.'~ ci'-Pr.'-ttc.~ :-ncT-e.~ in 3 "-0:.: <D ~m co ~m BATCo document for Legal Services • Health Canada 20 May 1999

-7- To~'t Blend 7I: o CO0 ci-~rettes nac~:ed in ~ ~o,,~ 20 :.'[.L. r,~r-,::- ~,',; COd"P, ""='" Wit':'..~n ~.~r.b h!e,'d, hP!f should be !~b~.l!ed "z-.o::e f~-'st" an.4. h::~.:" ".":T.:~ke ~" " S.~C~:Ic . '~t?.;.~ o....--;~rg~t~..T..~.}%o,.I.'LE. "T..c. availoh] e et l~a-"t 2 wee!-Ts x~-., ,. o.-~, start of J'5 eld-':',:'::. In o:'/.::r to horn r~,,e~". ~.:nects of ,,.~,,4 ~-~414.-Z, in t:-:.2 test ci-arettes., the-,- will be subjected to a weir-hi selection prccess after ua.~:afe.cture. ~.'D and !A~.] 5th Jonuarv, ~,973- 0 <:D O O OO 4z~ BATCo document for Legal Services • Health Canada 20 May 1999

APPE}~I X i As ~ ~~ ~ " --~-~-~"~ ~ inst~_nces have arisen in which Taste & Ylavour re.~.a!ts have not gone in the direction that micht h,~';e been expected on the basis of evidence from P.D.L, Smoke Analysis, consttmer testing and on occasions, subjective s~zoking by members of Leaf ~-~-+ .... Some exs~mples of these are as follows: I • 2o 3o • o In the development of Flair (Switzerl~_nd), a T/F panel assessment indicated the test blend to meet the brief. However, this conflicted with the leaf blender's judgement which was subsequently supported by consumer testinG. In a Project Cosmos for Vunezuela, two different T/F nanels (Millbank and G.R.D.C.) returned different results for a test cigarette, neither of which agreed with assessments by Leaf people in both Venezuela and London• In the development of a Virginia br~_nd for the Middle East, a Millba~c panel indicated 'no difference' between two cigarettes, despite a difference of 43% in ~xtractable Nicotine between them, Blends developed for t'he rationa!izszion of pAT,L :.t~LL K.S. FILTU7 in _h/ro~~. <-.'..-re tested b:,. T/P :a.'.-els. l]~'" ["{]l'u~']", ........ z..,{ ............... ;---.-~l ~-,-~ and DeP.-.ar3" all of "...,hich _e_~:..d a _~so ..... o_~ -- .... ~ ..... - Howev.~r, their results were dimectly contradicted by cons~e:" testing. A Miilban1: .~anel indicated that of t'.:o blends developedz ~c- COI'-~?~A K.S.F.T. (Aniola), on=_ had a higher intensity of Imcac~ than the otl:er. This result directly conflicted with the ooinions of Leaf Depar~:=ent (Y.,illbank), P.D.L. Smok~ Anal-,-~i of Ex%rac=ab!e Nicotine and info:-mal re~earch c_-.:ong a s.-_zi! grc: of eonstun=-rs. These discr---z_nci,~ hay.~ led Leaf Dep--_rtz-:nt (::r.. R.D. -s':: t:1~', ce~ o:'~ !ach of i-~.%ai] e~ C~.-~ ...... 7.r.~t " " " .':" - - u.n:=crs~&na~n/ of _,.'_: ~--: ..... :.~'- .... ~ hi '- .... n ~ ..... ,...,,.;. ...... S 0".7.% "$I!q0.'-';:.9~-." ~ ........... . Z.qP.. su32ec+~--~ On'~{n'~ and those of cons'~::3rs: ar.d to Sll:][Oi't "" ..- d.er.~ is o.P--~'-,~ ..... to in':is'.i-"_t2 ".".:':iv re!~_:ic:zshio ~:.:r':.-::='. -- ; sa-'s .... cru': o:? this i$ t~:nt "..;~_ ..._:'~-z~_ very ~...._ ~ ~--; .~ OJLi"'''~~-~ .~" empirical r~la~.is::-hip over a ion.] period of ti:n_~ b-z...'zen <D j-..,.' .... ~%t bii -'.:~in- b:'-.ui.z ÷=~'" to thic I~ so-..-.~. ~.: .... ~-~ " there is a ..... " " _ .... " .... a ~o .stabl~-:~ th.': cosition oz T/? pan-i regui.~.:-- .,~+" 4 the ~-,~'-i ,n~ ~,:.~-:.~,, BATCo document for Legal Services • Health Canada 20 May 1999

J~, ,'.r'"o , 'Ji .'. PROJT~T ].q~..I %.~en the questS'on wn~ first raised of studyi-:C the possible re'_.ationzhi':~s e'-:izuin-q bet~-:oen '-'aste and 3'la,'our. Con.~t'.,-.:~.'_- ?='o'iuct =estin.s ~.n'~ P.D.L. zuoPe- anal[.'[i3, it :;-~.'; felt thPt con~ide~'ab!e infoz-:.tion mi~ht be obt-~inc.d hit examinSn~ y.cr..= reneerch pro,-jcct.'-: -.:i-.:;re the test ciz~:revtes had been subjected to a!l th'.'oo means of assessment. However, the data turned out to be much more limited than anticipated: the results of only. five U.K. initiate.=. projects ::~re av~i!sb!e~ ~.'here st least t':o of t':e ~e.=nu of testin~ had been applied, and recuests to eve~_'sess co:.::-ni~ for simile2 ~nfo'~z~ion ::ez'e u:~suce~sf'al. It %-:as therefo:'? decided to ~xazlne the five U.}[. projects, of ~.Sich th_~c[: concerned Virginia cigarette blends and t%.'o U.S.~, The3~ :'ore: i) Project Osprey the development of P!'~Y~S h~. z. },~i:G SIZE. ~'~e consumer reseae:h ::~::. car~'ied out in the C::~:~_nel Island :- 1938. ii) l~coject Chsrlton the ~leve!op,-.~nt of ---,.iBASSY EJ--_::'XI'I-- for the ;-;idd!e East. Research co::!': in ~9~5. iii) P~;oject Alpha the de,re!c-..~ent of 3LT's version -: JOH;: -- "~ .... ' r, ~_~:. SPECI.-'-L J.esearc.: . srr± : out in ~97~. iv) .=;~L ;'L:_LL KSF v) PALL IIALL ~ ?Llf. - ~ . :...~,~.-".'-'- ,--~ -,~..-. .... ,, ~--_~ Testiu-. of a!ternntive b!en.~-" _-'-.r t'.-.- retionalisotion of ~,4_~o~e£..~ .,.r.::.--.-c- of this brand. Testing too'-: nit.co i- ~97o. a nsr~lle! test to the ~,:-T.T. ;L'/_L .'.'S- project. O O< • ..-.,4=:, -.[:is r...':ic'.: r,..-'."l., in.':.: "-'eL--:'--' el '. ''.'.::::.'.'. "--':- - . "%~ sati.-:fne-e.~il.-." t::er.e n-~.--r:~...- r-.-~ r, tic-~:'~'n=_ .... : ~, a z:eci~i "-::: ..... pt'o~ect c:.ou'-(; be I:~LCIGZ£C- ...... ~: .!) di:.:."e:':-::ces h..t:.'-'e~t ci~::-.:.-t, tes a:'c, c!oc,:l:: dofin=d (:': ":' & 2!:.'.'n--" t~'.-'...-.) BATCo document for Legal Services • Health Canada 20 May 1999

-2- ii) strict control is maintained over the testin~ procedurc~ F!ED!![GS i) l=ro3ect Oswrev The results indicate en encoura~inc corresnondence bet~.:een the product test and Taste and Flavour panel evaluation as follows: a) panel ratinm of Izoact reflects consumer 'strenNth' and 'satisfactionS' results b) Irritation rawl~in~s by the p~nel ere reflected in consumer measure= of ',=,oo~,,..=~/_.o_pne~-' 'hotness/coolness' c) the nanel ra.n/-:Jz~._= of Amplitude coincides with 'amount of flavcur' assessment by consumers. Results of routine brand (P.D.L.) analysis were unobtsinab!e. ii) Pro~ect Croy~.on Impact and Ar.n!ih~de ratings by the Taste r, na Flgvour panel corr=sT.c.nd :~er~tc!.-..- :.ell with the ~roduct test rez-u!ts -':~ ~.l_,~;_-....c .- ,;v.~.;e.-.: the c, ii-~.r~ ~-¢: ,, ....... in terns of ir:,~.te.tions sre l~.ot reflected ~-- ti-o cor.au'_-.~r: a) ~) t,. .... :-~ Of ':" .... ~" P-...--x.-/s::ooth:.,~.sc.'~, "-"~.._. no~ ,~,,-'~e~.;'..~ ~,__ b:.- -:-~.. iii) ?ro/~.~t A!~ha C The _~er.ults of t!:e thwee :;eans of ~.s~eszz~enb shaw a clor.e co m. BATCo document for Legal Services • Health Canada 20 May 1999

-5- ~) b) Taste end Flavour p~nel ratinT for Izpnct is similcr to conct:-.:er '."~. ....... ,~.,_= ~..-" ' m,d.. 'nildr.czz ' readin-.z. . little or.-'9 "/iffcr~nce -- tc~:.~ ~-'.. conzu-.er ~zse~c::.::z of 'irritatiOn' ~nd 'bitte.-n. esz' is reflected : ...... :,'-'~ scor'r~:z for !='rit~.tions. c) panel ..-m~:in- of Anslitude is similar to 'amount of of f!~vour' ratir-, b-," conzuuers. 6) ana!vs;s of extractable nicotine supports panel Iup~ct and consumer '.=zrencth' ~nd 'mildness' ratings. iv) P.t,I:L I:ALL K.S.F. Results did not tie in es night be expected: a) differences deternined by the consumer in terzs of 'stren~-th' and 'uil~mess' snd sho~m up in an~ljzia of ex~crac~ahl~ nicotinc ~:ere not apparent in r,^.~ Flavour penal scores for Impact. b) panel ra,~7:i~': of irritations is contraaicbed by consu.-_,~r aase.%ar, ent o~ 'irritation' and 'bitterness'. v) ?~,,L ![:j,_-r, "...".S._-'. i-'7.!. A more eneoure~in- !~vel of correzpon~:nc~ is :~-4~.:; pro~ ec; : a) COnCUr"-.``` .... rC::kin.-.z of '=~-~-~'.~._...~_. ~.._~-~" '-~"~-='____ ... COinsii~ v:izh .n~nel r.ntin ~ ___ b) little or .-.9 difference_..4-~4^~+~-;.._.,. ~ ~, .~o.,..~,~ ~..~ .....ci--nrette.- ~-. consumer "_'atin-~ o-" 'i:'vittzicns ', but :-:<orc__.'4 ;,-~..~__~ .... .... ~ Fla':our ,-.'=n,;! :corot. c) e"-~$"-~-. .._. of _..._..~,_.e~-'~--~^~',~ nicotine ccntradicts con-::-..-'." 0 "~"" ..... $72; 0 0 Con-u-.>-:- -. 'e',:'c/- hl,-u.l te:-t :','-':.".!:;s c.:':[ .-.D.L. r:..-.]:'.'i.', :." BATCo document for Legal Services • Health Canada 20 May 1999

-4- sho~..n in the table below: ~/~ Impact Connu-er Reseorch neasuremen~ Irritations 7 'strength' 'satisfaction' 'smootb~=ess/sharpness' 'irritation' 'bitterness' Amplitude 'amount of flavour' P.D.L. E~ractable Nicotine I "~,,,.'~,-'#'," there ~ s not sufficient data avail~bl~ oor ~ ......o*_ c~nci_~,-.n. :.,ith'- "':'-~ "- " ..... .......... ~ ~ ", ^7 ........... ava=zazl÷ t~ s].lo:; zz~-n concl~si¢,nz t~ be drz',m., in addizion, :r is i-r;or-.~n; to bear in mind that t:-.~se pro.jects were no~ set-up with the . . t .~ these v-a~J.ous ~,,~:,.-'~:- n.~ .... o~-= of inte='-re~at:~..G hu re'sults of u .... is ann -~-:n~ zz-~'_'c us-.' noz n~ve "-,.e,. as ri-oro::s 8 control of technacu~.s_ an._" ci.~'o:'eZto. _~..,nl¢-u_ :~ as ~.;ould _ ...... ,,c~.,_~-'~ • o ::.D o oo cr- BATCo document for Legal Services • Health Canada 20 May 1999