5.5 measures to evaluate the effectiveness of tobacco...

287

BBaacckkggrroouunndd

The cigarette package serves as thecornerstone of tobacco marketingand advertising campaigns (Slade,1997; Pollay, 2001). Package designhelps to reinforce brand imagerycommunicated through other mediaand plays a central role in retailmarketing. The importance ofcigarette packaging only increasesas other forms of marketing arerestricted, as indicated in the fol-lowing quote from a Philip Morrisexecutive: "Our final communicationvehicle with our smoker is the packitself. In the absence of any othermarketing messages, our pac-kaging...is the sole communicator ofour brand essence. Put anotherway—when you don’t have anythingelse—our packaging is our mar-keting.” (Alechnowicz & Chapman,2004).Governments in many juris-

dictions have begun to apply greaterrestrictions on tobacco labelling. Asmuch as half of the package is nowused by regulators to communicatethe health effects of smoking.Governments have also begun to

prohibit packaging elements thatare deemed to be misleading tosmokers. As a consequence,labelling policies have begun to alterthe traditional appearance of thecigarette package. The importance of tobacco label-

ling policies is highlighted in Article11 of the WHO FCTC (WHO, 2003).Article 11 sets international stan-dards for packaging and labelling oftobacco products in three broadcategories: 1) mandatory healthwarnings; 2) restrictions on branddescriptors, such as the use of “light”and “mild”; and 3) information oncigarette contents and emissions1(Figure 5.26).

HHeeaalltthh wwaarrnniinngg llaabbeelllliinngg

Cigarette packages in the vastmajority of countries carry a healthwarning (Aftab et al., 1999). How-ever, the position, size, and generalstrength of these warnings varyconsiderably across jurisdictions.FCTC Article 11 requires thatpackage health warnings mustcover at least 30% of the packagesurface and be “large, clear, visible,

and legible” (WHO, 2003). Beyondthese minimum requirements, Arti-cle 11 also states that warnings“should” cover 50% or more of apackage’s principle surfaces, and“may” be in the form of pictures.To date, at least eight countries

have implemented picture-basedhealth warnings that meet theFCTC’s “recommended” standard(see Figure 5.27). A number of otherjurisdictions, including the EuropeanUnion, have recently implementedprominent text warnings which meetthe minimum FCTC standard. Moreobscure text warnings remain inmany other markets, including theUSA, China, and Russia.

CCoonnssttiittuueennttss && eemmiissssiioonnssllaabbeelllliinngg

There is general agreement thattobacco packaging should includeat least minimal information aboutsome of the hazardous and addi-ctive constituents in tobacco andtobacco smoke. FCTC Article 11requires that packages contain“information on relevant consti-tuents and emissions of tobacco

5.5 Measures to evaluate the effectiveness of tobacco product labelling policies

1Tobacco labelling policies apply to a broad range of tobacco products, including a range of combustible products, such ascigars, and the packaging of loose or “fine cut” tobacco, as well as non-combustible tobacco products. However, much of thissection will focus on labelling policies for factory-made, “pre-packaged” cigarettes given that they are the primary target oflabelling policies, and the area in which most research has been conducted. Labelling policies for other types of products willbe described briefly in a separate section to follow.

section5.5janvier13:Layout 1 13/01/2009 10:07 Page 287

IARC Handbooks of Cancer Prevention

288

1. Each Party shall, within a period of three years after entry into force of this Convention for that Party, adopt andimplement, in accordance with its national law, effective measures to ensure that:

(a) Tobacco product packaging and labelling do not promote a tobacco product by any means that are false, misleading,deceptive or likely to create an erroneous impression about its characteristics, health effects, hazards or emissions,including any term, descriptor, trademark, figurative or any other sign that directly or indirectly creates the falseimpression that a particular tobacco product is less harmful than other tobacco products. These may include termssuch as “low tar”, “light”, “ultra-light”, or “mild”; and

(b) Each unit packet and package of tobacco products and any outside packaging and labelling of such products alsocarry health warnings describing the harmful effects of tobacco use, and may include other appropriate messages.These warnings and messages:

(i) shall be approved by the competent national authority,(ii) shall be rotating,(iii) shall be large, clear, visible and legible,(iv) should be 50% or more of the principal display areas but shall be no less than 30% of the principal display areas,(v) may be in the form of or include pictures or pictograms.

2. Each unit packet and package of tobacco products and any outside packaging and labelling of such products shall,in addition to the warnings specified in paragraph 1(b) of this Article, contain information on relevant constituentsand emissions of tobacco products as defined by national authorities.

3. Each Party shall require that the warnings and other textual information specified in paragraphs 1(b) and paragraph2 of this Article will appear on each unit packet and package of tobacco products and any outside packaging andlabelling of such products in its principal language or languages.

4. For the purposes of this Article, the term “outside packaging and labelling” in relation to tobacco products appliesto any packaging and labelling used in the retail sale of the product.

WHO (2003)

FFiigguurree 55..2266 WWHHOO FFCCTTCC AArrttiiccllee 1111:: PPaacckkaaggiinngg aanndd llaabbeelllliinngg ooff ttoobbaaccccoo pprroodduuccttss

products as defined by nationalauthorities.” At present, however,national authorities have takenmuch different approaches tolabelling constituents and emis-sions, and there remains con-siderable disagreement regardingwhat should be considered “rele-vant” information. The current regulatory practice

in many jurisdictions is to requiremanufacturers to print levels for

three emissions in the mainstreamsmoke: tar, nicotine, and carbonmonoxide (CO). Emission levelsare generated by machine-smoking cigarettes according to astandard set of puffing conditions;typically the International Stan-dards Organization (ISO) method,which serves as the currentinternational standard. However,in light of research indicating thatthe tar and nicotine levels

generated under the ISO testingmethod are unrelated to individuallevels of exposure or risk (Burns etal., 2001), there are growing callsfrom within the tobacco controlcommunity for the ISO numbers tobe removed from packages. Somejurisdictions have supplementedthe ISO numbers with additionalemission information. For exam-ple, Canada increased the list ofemissions that must be reported

section5.5janvier13:Layout 1 13/01/2009 10:07 88

Measures to evaluate the effectiveness of tobacco product labelling policies

289

FFiigguurree 55..2277 PPiiccttuurree--bbaasseedd wwaarrnniinnggss



290

(by adding benzene, formal-dehyde, and hydrogen cyanide totar, nicotine, and CO), andrequires a second set of numbersfrom a more intensive machinesmoking method for each emis-sion (Figure 5.28). Otherjurisdictions have replaced quan-titative emission values withdescriptive, non-numerical infor-mation on hazardous emissionsand toxicants. A consensus hasyet to emerge on “best practices”for this area of tobacco labellingpolicy.

BBrraanndd ddeessccrriippttoorr llaabbeelllliinngg

Tobacco manufacturers incor-porate a variety of common termsinto the names of their cigarettebrands. Words such as “light” and“mild” are ostensibly used todenote flavour and taste;however, “light” and “mild” brandsare often promoted as “healthier”products and are typically appliedto brands that generate lowermachine levels of tar (Pollay,2001; Pollay & Dewhirst, 2002).Not surprisingly, “light” and “mild”brands are perceived by manyconsumers to deliver less tar andlower risk than “regular” or “fullflavour” varieties despite evidenceto the contrary (Ashley et al.,2001; Shiffman et al., 2001).A growing number of juris-

dictions, including Brazil and theEuropean Union, have prohibitedthe use of “light” and “mild” onpackages. Similar prohibitions areproposed in FCTC Article 11:"tobacco product packaging andlabelling do not promote a tobaccoproduct by any means that are

false, misleading, deceptive orlikely to create an erroneousimpression about itscharacteristics, health effects,hazards or emissions, includingany term, descriptor, trademark,figurative or any other sign thatdirectly or indirectly creates thefalse impression that a particulartobacco product is less harmfulthan other tobacco products.These may include terms such as“low tar,” “light,” “ultra-light,” or“mild”. Although there is evidenceto suggest that other packagingelements, such as the use ofcolour, may also create mis-leading perceptions of risk(Wakefield et al., 2002), “light” and“mild” descriptors are the onlypackaging elements to berestricted to date.

MMeetthhooddoollooggiiccaall iissssuueess iinn eevvaalluu--aattiinngg ttoobbaaccccoo llaabbeelllliinngg ppoolliicciieess

Evaluating tobacco labellingpolicies presents several uniquechallenges; this section reviewssome of the principal methodo-logical and analytical consi-derations.

“Alternative” tobacco products:

Labelling policies have generallybeen designed with factory-made,pre-packaged cigarettes in mind.However, a substantial proportionof tobacco users throughout theworld use tobacco products thatare either packaged in a differentway, or have no manufacturedpackaging at all. This hasimportant implications for patternsof exposure to health warnings.

For example, consumers who buyloose or fine-cut tobacco, withoutany manufactured packaging,may not be exposed to producthealth warnings. Even consumerswho buy fine-cut tobacco, sold ingovernment-mandated packaging,will have different patterns ofexposure than those who smokemanufactured cigarettes, and whoare likely to be exposed to thewarnings each time they reach forthe package. As a result, studiesconducted in markets with aconsiderable proportion of fine-cuttobacco sales, such as the UnitedKingdom, New Zealand, andThailand, may need to stratify forfine-cut versus manufactured ormixed use. Smuggled or con-traband cigarettes may also alterpatterns of exposure in caseswhen the contraband product isnot manufactured to the samelabelling specifications.

Issues in attribution: dealing withmultiple sources of health infor-mation:

Health behaviours with multipledeterminants present a challengeto policy evaluation. The problemof attribution is particularly acutefor health warning labels. First,labelling policies are oftenimplemented simultaneously withother tobacco control measures,including increases in taxation andsmoke-free policies. As a result, itis difficult to isolate the effect of anindividual policy on overall pre-valence. Second, many of thespecific themes and messages inlabelling policies are commu-nicated through other sources.



291

EEuurrooppeeaann UUnniioonn ((UUnniitteedd KKiinnggddoomm)):: TThhrreeee IISSOO eemmiissssiioonnss

CCaannaaddaa:: SSiixx IISSOO eemmiissssiioonnss aanndd ‘‘HHeeaalltthh CCaannddaaddaa IInntteennssee’’ eemmiissssiioonn

AAuussttrraalliiaa:: DDeessccrriippttiivvee iinnffoorrmmaattiioonn

FFiigguurree 55..2288 CCoonnssttiittuueenntt llaabbeelllliinngg ppoolliicciieess iinn tthhee EEuurrooppeeaann UUnniioonn,, CCaannaaddaa aanndd AAuussttrraalliiaa



292

Mass media campaigns andhealth professionals often targetthe same health effects,particularly with regards to com-mon diseases such as cancer andcardiovascular disease. Theimpact of package-based labellingpolicies may also be confoundedwith health warnings in othersettings. Various jurisdictionsrequire health warnings in retailoutlets and warnings on printadvertisements for tobacco pro-ducts. Third, perceptions of riskand health knowledge areinfluenced by an inter-related setof factors at the individual, social,and environmental level. Fewstudies are able to measure morethan a small number of thesefactors within a single study andnone can fully isolate thecontributions of each. These reali-ties underscore the importance ofthe methodological featuresdescribed in Section 2.1. Inaddition, environmental scans ofother mass media campaigns andpolicy interventions can provideimportant context.

“Wear-out” and impact over time:

It is widely accepted that thesalience of advertising and healthcommunications is typically grea-test upon first exposure (Bornstein,1989; Henderson, 2000). The initialimpact of comprehensive labellingpolicies, such as the introduction oflarge graphic warnings onpackages, is often magnified bymedia coverage. As a result,measures of effectiveness are likelyto be strongly associated with theimplementation date. This has

implications for regulators in termsof ensuring periodic changes to thewarnings, as well as studies thatcompare labelling policies acrossjurisdictions. For example, a recentstudy found that new text-basedwarnings, introduced in the UnitedKingdom in 2003, were con-siderably more likely to be noticedthan Australian text-based war-nings, which were only slightlysmaller, but had been in place formore than eight years at the time ofthe survey (Bornstein, 1989).Ideally, labelling policies should beevaluated at similar post-imple-mentation dates; at the least,differences in follow-up periodsshould be clearly noted and takeninto account when interpretingfindings. There is preliminary evidence

to suggest that not all measures ofeffectiveness decline at the samerate over time. “Proximal” mea-sures of salience, such as noticingwarnings, may erode more quicklythan “distal” measures, such asreporting that health warningsmotivate quitting and increasethoughts about the health risks ofsmoking (Hammond et al.,2007a). It is even plausible that forsome smokers the impact ofhealth warnings could increaseover time. For example, thecessation and telephone quitlineinformation included in manyhealth warnings may only becomerelevant to smokers as theycontemplate quitting. In a popu-lation-based survey, however, theebb and flow among individualswill balance out, and one wouldstill anticipate decreases in mea-sures of effectiveness over time.

Youth:

One policy-relevant questionconcerns the impact of warninglabels in reducing youth uptake.Evaluating the impact of healthwarnings among youth duringperiods of smoking initiationrequires a different conceptualapproach. Given that the cigarettepackage serves as the medium forlabelling policies, consumptionlevels may be positively asso-ciated with knowledge of thewarning labels. In other words,individuals who smoke 20cigarettes a day will be exposed tothe warnings more frequently thanindividuals who smoke less thandaily. Furthermore, “occasional”youth smokers are less likely tobuy their own package, reducingthe likelihood of exposure towarning labels, compared to moreregular smokers who are morelikely to buy their own package(Leatherdale, 2005). As a result,individuals who smoke morefrequently are more likely to recallthe content, location, and otheraspects of labelling policies, acounter-intuitive association at firstglance (Robinson & Killen, 1997). A second issue concerns

longitudinal studies that usemeasures of exposure orknowledge as predictors of futuresmoking behaviour among youth.During youth and youngadulthood, the rate of smokingundergoes significant increases.As youth smoking behaviourincreases, so too will theirexposure to the package and theirknowledge of the warnings. Thus,whereas a negative association



293

between exposure and futuresmoking behaviour may beexpected for anti-smoking cam-paigns in other media, this is notthe case for warning labels. Failure to account for the

somewhat counter-intuitive asso-ciation between smoking andexposure to health warnings canresult in misleading interpretationsof data. For example, one studycharacterized an associationbetween increased smoking andincreased knowledge of healthwarnings as “paradoxical,” andalso found evidence that UShealth warnings were ineffective(Robinson & Killen, 1997). Thismay have been the case; how-ever, without a comparison group,the authors had no way ofknowing whether the increases insmoking behaviour were greater,less, or no different than theywould have been if no warnings ormore comprehensive warningshad been implemented. It may be,for example, that fewer youthinitiated smoking than would haveotherwise occurred without thehealth warnings. In fact, this wasthe pattern reported in a longi-tudinal study comparing changesin youth smoking in Canada andthe USA following the introductionof graphic warning labels onCanadian packages. Smokingrates and knowledge of thewarnings rose among Canadianyouth as they aged; however, theincrease in smoking was sig-nificantly less than among USadolescents and the increase inknowledge of the warnings con-siderably greater (Fong et al.,2002). Overall, this study under-

scores the importance of suitableresearch designs and appropriateinterpretations of the data whenevaluating warning labels amongyouth.

Evaluation of individual messages& content:

Beyond the question of whetherhealth warnings are generallyeffective, there is a growing body ofresearch on the individual elementsof a warning. These elements canbroadly be categorized in terms ofdesign and content components.To date, much of the research hasfocused on important design ele-ments including the size, position,and use of pictures on the package(Strahan et al., 2002). In contrast,relatively few studies have exa-mined the content of individualmessages. Population-based surveys that

compare labelling policies acrosstime or jurisdictions are somewhatill-suited to the task of evaluatingindividual warnings. Policies typi-cally differ on more than one di-mension, and policy changes ty-pically involve increases in the size,number, position, and type of in-formation presented in eachwarning. Evaluating individual com-ponents or messages becomesmore complicated as the number ofwarnings and complexity of in-formation increases; it is far easierto evaluate the effectiveness of asingle warning through survey-based research than to evaluate thecontent of 16 individual warnings. When assessing the impact of

individual warnings, it is alsoimportant to consider that many

health warnings are tailored toparticular sub-groups of smokers.Warnings on the risks of smokingwhile pregnant, for example, havelittle relevance for older males.Thus, it is conceivable that somewarnings may perform very wellamong sub-groups who comprisethe target audience, but relativelypoorly among the population as awhole. As a consequence, surveymeasures may need to beadapted and the findings mayneed to be stratified amongrelevant sub-groups. One mightexpect the tailoring of warnings toincrease, as the use of picture-based warnings increase, alongwith the typical number of rotatingwarnings in a given jurisdiction.In general, population-based

surveys may be most appropriatefor identifying the overall ef-fectiveness of a set of healthwarnings. However, the task ofevaluating the content of individualwarnings is best suited toexperimental or qualitative designs,in which the content and designcan be systematically varied.

Geographic & cultural differences:

Very little research has examinedpotential geographic and culturaldifferences in the effectiveness ofhealth warnings. Although thefundamental principles underlyingthe effectiveness of warnings areunlikely to vary across culturesand regions, the effectiveness ofindividual messages may indeedperform differently. First, smokersin different parts of the world havedifferent levels of existing healthknowledge. This has implications



294

for the type of messages to beincluded in warnings. For exam-ple, Australian smokers may havea relatively higher level of healthliteracy than smokers in otherregions, which may account forthe decision to include a warningfor “peripheral vascular disease”on packages. Picture-basedwarnings may be particularlyimportant in populations with lowerliteracy rates (CRÉATEC, 2003).In addition, the images used inone jurisdiction may not be equallyeffective in another. For example,several of the picture-basedwarnings that appear onVenezuelan and Uruguayanpackages, and elsewhere, usesymbols that may be culturallyspecific. Finally, similar sets ofwarnings may be more effective inareas where smokers haverelatively little access to anti-smoking information from massmedia or health professionals.Few of these issues have beenaddressed to date; however, theyare likely to gain prominence as agrowing number of jurisdictions inAsia, Africa, and the Middle Eastenhance their labelling policies tomeet Article 11, and must rely onan evidence base that derivesfrom relatively few Western andLatin American countries.

Evaluating the removal of infor-mation:

Unlike other labelling policies,restrictions on brand descriptorsresult in the removal, rather thanthe provision of, information. Thispresents a challenge to evalu-ation, particularly when the

information being removed is usedas a brand identifier. In the case ofbans on the use of “light” and“mild,” the terminology that waspreviously used to identify a classof products no longer exists.Smokers may retain the samemisleading perceptions of theseproducts after the terms havebeen prohibited, but surveymeasures can no longer refer to“light” or “mild” cigarettes in thesame way as in the past.Therefore, survey measures mustbe designed so that the wordingand meaning of questions remainsconstant before and after theremoval of these terms. Thiscreative challenge is only nowbeing confronted by researcherswith the recent advent of “light”and “mild” prohibitions. Oneapproach, discussed later in thissection, is to make the res-pondents’ “own brand” thereferent for questions. Another implication of the

“removal” of brand information isthat the beliefs associated with“light” and “mild” cigarettes arelikely to persist for some time afterthe descriptors disappear frompackages. This situation is similarto advertising, promotion, andsponsorship bans; one should notexpect beliefs to change imme-diately upon the implementation ofthe policy, but more gradually overtime. Indeed, anecdotal evidencesuggests that many retailers andconsumers continue to use theterms “light” and “mild” well aftertheir removal. Other packagingelements and aspect of cigarettedesign may also reinforce thesame beliefs and perceptions as

the “light” and “mild” descriptors.These considerations are impor-tant in terms of how the data areinterpreted and how the “effective-ness” of light and mild policies isconceptualized.

Defining misleading descriptors:

There is widespread confusionamong both consumers and manywithin the tobacco controlcommunity regarding several keyterms relevant to labelling policy.Many fail to make the distinctionbetween “light” and “mild” and “lowtar.” Whereas “light” and “mild” areterms used in the name of abrand, strictly speaking “low tar”refers to the emission levels undermachine testing. Although there isa very strong correlation betweenthe two (manufactures oftenattach “light” and “mild” des-criptors to brands that generatelower tar levels under the ISOsmoking machine), one can havea “light” cigarette that does notgenerate “low tar” levels and viceversa. Strictly speaking, injurisdictions with bans, “light” and“mild” cigarettes do not exist,whereas “low tar” cigarettes do.To complicate matters further, theterms “light” and “mild” can also beused to refer to sensory propertiesof a cigarette. Thus, smokers maystill retain the concept of acigarette as “light” or “mild” evenin the absence of a branddescriptor. Given the potential forconfusion, survey measuresshould be explicit about theintended meaning of these termsand should avoid using theminterchangeably. This becomes



295

apparent when measuring theseconcepts in jurisdictions where the“light” and “mild” brand descriptorshave been removed.

MMeeaassuurreess

This section provides an overviewof the key constructs andindividual measures that havebeen used to assess labellingpolicies. The constructs rangefrom the extent to which labellingpolicies are noticed and pro-cessed, the extent to which theyalter key beliefs (such as levels ofhealth knowledge), to their impactupon downstream behaviouraloutcomes. These measures canbe organised within a conceptual

model, as illustrated in Figure5.29. Other psychosocial varia-bles, such as social norms andbeliefs about the tobacco industry,could also be added to this model,but have been excluded in theinterest of brevity. The followingsub-section begins with a reviewof quantitative measures, followedby qualitative measures, and abrief discussion of the role ofindustry documents (Tables 5.27-5.39; see also Appendices 9 and10).

Measures of labelling salienceand processing:

Health warnings must be cog-nitively processed to be effective.

The extent to which information isprocessed or elaborated upon hasbeen shown to be the mostimportant determinant of memoryand attitude change in response tonew information (Anderson,1990). A number of measureshave been developed to assesscognitive processing of healthwarnings as a general indicator oftheir salience. These measuresrange from more “shallow”measures of processing, such asa general awareness of warnings,to “deeper” measures of pro-cessing, including reading thewarnings and thinking about themwhen they are not in sight(Borland & Hill, 1997a; CanadianCancer Society, 2001; Hammond

CCoonnssttrruucctt NNoottiicciinngg HHeeaalltthh WWaarrnniinnggss

MMeeaassuurree “In the last month, how often, if at all, have you noticed the warning labels on cigarettepacks?” (Never, Rarely, Sometimes, Often, Very Often)

SSoouurrcceess Hammond et al., 2006a; Hammond et al., 2007a

VVaalliiddiittyy The time reference varies across different versions: some questions include no timereference (“How often do you notice…”), whereas others refer to the “last month” or “lastthree months.” The response categories also vary and are often collapsed into a smaller number of categories in analysis. The basic question can also be asked within the contextof noticing other forms of anti-tobacco media (e.g. “In the last 6 months, have you noticed advertising or information that talks about the dangers of smoking, or encourages quittingin any of the following places? (Yes, No to a list of 9 media channels, including on cigarettepackages)).

CCoommmmeennttss Overall, a straightforward measure of the salience and processing of warnings that shouldbe considered within the core set of variables to assess health warnings. As close to a “goldstandard” in this domain as exists. Using the same wording to ask about salience of othermedia channels provides a useful comparative index for the salience of various healthinformation channels. A recommended and essential measure for evaluating healthwarnings.

Table 5.27 Essential Measure of Labelling Salience and Processing



296

PPOOLLIICCYY

PPOOLLIICCIIYY--SSPPEECCIIFFIICCMMEEDDIIAATTOORRSS

GGEENNEERRAALLMMEEDDIIAATTOORRSS

CCeessssaattiioonnkknnoowwlleeddggee

Self-efficacyQuit-lineinformationWebsiteinformation

BBrraanndd aappppeeaall

Attractiveness

QQuuiittiinntteennttiioonnss

SSmmookkiinngg bbeehhaavviioouurr

Changes in consumptionQuit attemptAbstinence

AAvvooiiddaannccee

Covering warningsRequesting packs

AAffffeeccttiivveerreeaaccttiioonnss

FearDisgust

HHeeaalltthh kknnoowwlleeddggeePPeerrcceeiivveedd rriisskk

Beliefs abouthealth effectsThinking/worryabout health risks

OOUUTTCCOOMMEESS

HHeeaalltthh WWaarrnniinnggss

SSaalliieennccee && pprroocceessssiinngg

Awareness of warningsKnowledge of warningsNoticing of warningsDepth of processing

FFiigguurree 55..2299 CCoonncceeppttuuaall ffrraammeewwoorrkk ffoorr tthhee eevvaalluuaattiioonn ooff hheeaalltthh wwaarrnniinngg ppoolliicciieess



297

CCoonnssttrruucctt ((aa)) GGeenneerraall AAwwaarreenneessss

MMeeaassuurree “Have you seen health warnings on cigarette packages?” (Yes, No)

SSoouurrcceess Borland & Hill, 1997a; Health Canada, 2005

VVaalliiddiittyy Good face validity; associated with policy strength.

VVaarriiaattiioonnss Response categories are consistent across measures. Alternative wordings include: “Are you aware of health warnings on cigarette packages?” and “Are there health warnings on packages?” Some questions refer specifically to the release of new warnings. For instance, “Are you aware of any recent changes to health warnings on cigarette packs?” and“Have you noticed any changes to the health warnings on cigarette packages since[date]?”

CCoommmmeennttss Provides an overall measure of general awareness. Limited value in examining changes and comparing across jurisdictions, given almost universal awareness among smokers.Most useful for examining policy implementation and rollout when the question makes reference to “new” warnings, or in jurisdictions with very weak health warnings and noprevious research. Overall, an informative measure, but only recommended under thesecircumstances.

CCoonnssttrruucctt ((bb)) RReeaaddiinngg//LLooookkiinngg CClloosseellyy aatt HHeeaalltthh WWaarrnniinnggss

MMeeaassuurree “In the last month, how often, if at all, have you read or looked closely at the warning labelson cigarette packs?” (Never, Rarely, Sometimes, Often, Very Often)

SSoouurrcceess Hammond et al., 2006a; Hammond et al., 2007a

VVaalliiddiittyy Face validity; good convergent validity with other measures; good predictive validity for strength of policy and motivation to quit smoking.

VVaarriiaattiioonnss The time reference varies across different versions. Also, some versions refer to reading, other versions use broader language, such as looking closely, and some versions includeboth terms. Looking closely may be more appropriate for pictorial warnings.

CCoommmmeennttss Strong correlation with noticing, but conceptualized as a “deeper” measure of processing.Overall, a recommended and important, but not essential, measure of salience andprocessing that may be particularly relevant for textual aspects of warnings.

CCoonnssttrruucctt ((cc)) DDiissccuussssiinngg tthhee HHeeaalltthh WWaarrnniinnggss WWiitthh OOtthheerrss

MMeeaassuurree “In the last month, how often, if at all, have you talked about the health warning with others?” (Never, Rarely, Sometimes, Often, Very Often)

SSoouurrccee Hammond et al., 2003

Table 5.28 Additional Measures of Labelling Salience and Processing



298

et al., 2003; Christie & Etter, 2004;Hammond et al., 2004a; HealthCanada, 2005; Koval et al., 2005;Hammond et al., 2006a; Ham-mond et al., 2007a). Measures of general “aware-

ness” are typically endorsed by avast majority of respondents,including non-smokers, regardlessof the type of warning level. Thesequestions are often used toexamine the implementation, or“roll-out,” of new package war-nings following a change in policy.This information is critical for anypopulation-based survey conduc-ted shortly after the imple-mentation of a new policy, given

the uncertainty regarding whenhealth warnings begin appearingon packages. In contrast to general mea-

sures of awareness, the extent towhich smokers notice, read, andthink about the warnings appearsto be highly dependent on thesize, type, and location of thewarning (Borland & Hill, 1997a;Health Canada, 2005; Hammondet al., 2007a). These measures ofprocessing are also subject to theimplementation date. Severalstudies have used measures ofprocessing collected from thesame population over time andcan be used to measure the

“wear-out” (i.e. decrease in thesalience of the warning labels) ofhealth warnings (Health Canada,2005; Hammond et al., 2007a).Additional data of this type mayhelp to answer the question as towhether the rate of decline amongmeasures of salience is asso-ciated with design features, suchas the size of warnings and theuse of pictures. In most cases,these measures have beenanalyzed as individual items,although in one case a depth ofprocessing scale was developedand tested (Hammond et al.,2003). In that instance, nine itemswere used to create a scale to

VVaalliiddiittyy Good face validity, convergent validity, and predictive validity for motivation to quit and futuresmoking behaviour when included as part of a composite measure.

VVaarriiaattiioonnss Variations of this measures use slightly different terms, including discussed and mentionedrather than talked about, as well as different response options, such as Never, Rarely, Sometimes, Frequently.

CCoommmmeennttss These measures provide a “deeper” measure of processing for labels and may be useful for comprehensive evaluations of labelling policies. Recommended, but not essential.

CCoonnssttrruucctt ((dd)) TThhiinnkkiinngg AAbboouutt HHeeaalltthh WWaarrnniinnggss

MMeeaassuurree “In the last month, how often have you thought about what the health warnings have to say?” (Never, Rarely, Sometimes, Often, All the time)

SSoouurrcceess Canadian Cancer Society, 2001; Hammond et al., 2003; Christie & Etter, 2004

VVaalliiddiittyy Good face validity, convergent validity, and predictive validity for motivation to quit and future smoking behaviour when included as part of a composite measure.

VVaarriiaattiioonnss “In the last month, have you ever thought about the warning labels or what they had to say when a cigarette pack wasn't in sight?” This variation of the measure requires a higherthreshold of processing than the items above.

CCoommmmeennttss These measures provide a “deeper” measure of processing for labels and may be useful forcomprehensive evaluations of labelling policies. Recommended, but not essential.

Table 5.28 Additional measures of labelling salience and processing



299

measure cognitive processinglabelled as “depth of processing.”Responses to each of the nineitems were rated using a 5-pointLikert-type format going from "notat all/never" to "all the time/a lot"and values added to create anindex. Examples of items includedwere “How carefully have you everread the messages on the outsideof a cigarette package?” and “Howoften have you thought about whatmessages on the inside ofpackages have to say?”Although the wording of items

is relatively similar across surveys,different time periods are used inboth the question and theresponse option in many cases.For example, whereas some“noticing” questions refer to thepast month, others refer to thepast three months, or use no timereference at all (Tables 5.27 and5.28). Nevertheless, findings fromthe same population are relativelysimilar across different question

wordings (Canadian CancerSociety, 2001; Hammond et al.,2004; Health Canada, 2005; Ham-mond et al., 2007a).

Contents & emissions:

Several studies have assessedthe extent to which smokersprocess emission informationprinted on the side of packages.These measures mirror the pro-cessing items used to gaugehealth warnings, although a morelimited set has been used. Bothstudies of which we are aware,indicate that smokers are lesslikely to read or look at emissioninformation than health warningson the face of packages(Thompson et al., 2006). Moregenerally, it is unclear whethersalience and processing typemeasures are as informative foremission labelling policies as forhealth warning policies. Unlikehealth warnings, which typically

include a number of rotatinghealth warnings, emission label-ling is consistent across packagesfor a given brand. As a result,there may be little reason forsmokers to read or attend to thisinformation on a regular basis. Asa consequence, we have notrecommended a specific measurein this section.

Physiological measures ofsalience and processing:

Physiological measures havebeen used in conjunction withsurvey measures to quantifyattention to and processing ofhealth warnings. These mea-sures have an advantage in thatthey are more “objective” giventhat they do not rely on self-reporting. In several cases, theyhave been used to compare thesalience of warnings with pac-kage design or within the contextof a tobacco advertisement. For

CCoonnssttrruucctt HHeeaalltthh WWaarrnniinnggss -- EEyyee TTrraacckkiinngg

MMeeaassuurree Participants wore eye-tracking equipment and viewed US cigarette advertisements withhealth warnings.

SSoouurrcceess Fischer et al.,1989b; Krugman et al., 1994

VVaalliiddiittyy Good predictive validity for recall and recognition of health warnings

VVaarriiaattiioonnss Viewing time serves as another measure of attention, where warnings are flashed on ascreen and the amount of time is recorded (Peters et al., 2007).

CCoommmmeennttss Eye tracking measures can help to identify the most salient design aspects of warning labelsand serve as an objective measure of attention; however, these measures are limited to“laboratory” based research designs.

Table 5.29 Physiological Measures of Salience and Processing



300

example, eye movements duringexposure to an ad have beenused as physiological indicatorsof attention to tobacco warnings.These measures that are directlylinked to cognitive processinghave been useful to investigatethe relationship between visualattention and a more traditional

communication measure (Krug-man et al., 1994) (Table 5.29).

KKnnoowwlleeddggee ooff hheeaalltthh wwaarrnniinnggss

Items assessing smokers’ know-ledge of health warnings areamong commonly used surveymeasures. Knowledge questions

have been asked using un-prompted recall (e.g. “Where arethe warnings located?”), as well asusing recognition tasks (e.g.“Please tell me which of thefollowing warnings appear oncigarette packages…”) (Table5.30) (Hill, 1988; Richards et al.,1989; Rootman et al., 1995;

CCoonnssttrruucctt ((aa)) HHeeaalltthh WWaarrnniinnggss -- LLooccaattiioonn

MMeeaassuurree Without looking at a cigarette package, where on the pack are the warnings or messages located?” (Open ended)

SSoouurrccee Hammond et al., 2003

VVaalliiddiittyy Good face validity.

VVaarriiaattiioonnss The same question has been asked without the prefix (“Without looking at a cigarettepackage…”), as well as with a diagram in self-completed surveys.

CCoommmmeennttss Useful measures for identifying basic knowledge about health warnings; however, it becomes complicated in jurisdictions with warnings on the inside and outside of packages. Unclear how emission and contents information should be treated, especially when provided by industry.

CCoonnssttrruucctt ((bb)) HHeeaalltthh WWaarrnniinnggss –– CCoonntteenntt

MMeeaassuurree “Without looking at a cigarette package, what specific health warning messages can youremember seeing on cigarette packages in Canada?” (Open ended)

SSoouurrccee Health Canada, 2005


VVaarriiaattiioonnss The same question has been asked without the prefix, which is typically included in telephone surveys to ensure the participant is not looking at the package during the call.

CCoommmmeennttss Useful measures for identifying basic knowledge about health warnings and, potentially, for identifying individual messages that are particularly salient. However, this measure will bedifficult to answer in jurisdictions with comprehensive health warnings, including multiple warnings on different areas of the package.

Table 5.30 Measures of Knowledge of Health Warnings



301

Borland & Hill, 1997a; Borland &Hill, 1997b; Robinson & Killen,1997; Hammond et al., 2003;Brown et al., 2005; Health Cana-da, 2005; O’Hegarty et al., 2006;Thompson et al., 2006). Measuresof unprompted recall for warninglabel content can be used to

identify which individual warningsmay be most effective. Injurisdictions with a large numberof warnings, this task can beparticularly helpful. Many of these measures have

been assessed among thegeneral population, including

among nonsmokers. Except forthe few questions that refer to arespondent’s “own” cigarette pac-kage, most measures ofawareness and knowledge appearto work equally well amongnonsmokers. Indeed, nonsmokershave been found to have

CCoonnssttrruucctt EEmmiissssiioonn SSiiddee PPaanneell –– CCoonntteenntt

MMeeaassuurree “Without looking at a cigarette package, can you name any chemicals or substances that arecurrently listed on cigarette packages in [country]?” (Open ended)

SSoouurrccee Health Canada, 2003

VVaalliiddiittyy Face validity.

VVaarriiaattiioonnss A common alternative is to ask about the quantitative level of specific emissions, such as tar“Without looking at a pack, can you tell me the tar level of your cigarettes?”

CCoommmmeennttss This measure examines basic recall of emission information printed on packages, and is often compared against “objective” data collected from other sources in order to evaluateaccuracy of self-report recall. This measure should be interpreted alongside measures onthe comprehension and use of this information (described later).

Table 5.31 Measures of Knowledge of Constituents and Emissions

Table 5.32 Measures of Affective Reactions to Health Warnings

CCoonnssttrruucctt HHeeaalltthh WWaarrnniinnggss –– AAffffeeccttiivvee RReeaaccttiioonnss

MMeeaassuurree “Have you experienced any fear as a result of the health warnings?” (Not at all / A little / A lot)

SSoouurrccee Hammond et al., 2004a

VVaalliiddiittyy Good face validity; good predictive validity for future smoking behaviour.

VVaarriiaattiioonnss Alternatives have used more comprehensive scales and asked about different affective reactions, including disgust and anger (Peters et al., 2007).

CCoommmmeennttss Affective reactions have been evaluated to a greater extent in qualitative evaluations of warning labels; however, survey-based measures may be a key mediator of downstream measures of impact.



302

surprisingly high levels of aware-ness and recall for prominenthealth warnings and picture-basedwarnings in particular (HealthCanada, 2005). However, bothrecall and recognition of particularmessages has been shown to behighly dependent on the com-plexity of the health warning and itsimplementation date. For example,virtually all Canadian smokers areaware of the health warnings onpackages, although we areunaware of any research indicatingthat smokers have correctly beenable to identify all 16 healthwarnings that appear on packages.Analyses must take into

account the consumption levelwhen assessing knowledge ofhealth warnings. Given theinevitable link between heavinessof smoking and viewing thewarning labels, knowledge is likely

to be greater among heaviersmokers. This association is likelyto be more pronounced withinsamples that include a broadrange of smokers, and are likely tobe greatest in studies thatcompare regular smokers withoccasional or nonsmokers. Theassociation between consumptionand knowledge is also likely to bestronger in jurisdictions with agreater number and complexity ofwarnings. For example, packagesin Canada carry information on theside panel, one of 16 healthwarnings on the outside ofpackages, and one of 16additional warnings on the insideof packages. In such cases, agreater number of exposures willbe required to recall variousaspects of the warnings. There are several limitations

with measures of knowledge.

First, when asking about thelocation of health warnings, oneissue is whether respondentsconsider emission information,which may be printed on the sidesof the package, as a healthwarning. Canadian data suggeststhat some smokers are aware ofthis information, but fail to cite it asa location. Second, in telephoneor web-based surveys, someparticipants may have a packvisible as they respond to thesurvey. As a result, somemeasures explicitly ask smokersnot to look at the package to avoidthis situation to the extentpossible. Third, measures ofknowledge can often be difficult tocompare across labelling policies.For example, smokers from theUSA, where a total of four differenttext warnings appear onpackages, have a much greater

Table 5.33 Measures of Avoidance

CCoonnssttrruucctt HHeeaalltthh WWaarrnniinnggss –– AAvvooiiddaannccee

MMeeaassuurree “In the last month, have you made any effort to avoid looking at or thinking about the warninglabels?” (Yes, No)

SSoouurrcceess Hammond et al., 2004a; International Tobacco Control Policy Evaluation Survey (The ITCProject)

VVaalliiddiittyy Good face validity; good predictive validity for future smoking behaviour.

VVaarriiaattiioonnss Several follow-up questions may be asked of those who respond “yes” to the initial question,above. For example, “Have you made any effort to avoid the warnings by: (1) Covering thewarnings up? (2) Keeping the pack out of sight? (3) Using a cigarette case or some other pack? (4) By not buying packs with particular labels?” (Yes, No to each question)

CCoommmmeennttss These measures can indicate the prevalence of avoidance behaviours and whether theyreduce the effectiveness of warnings. The follow-up questions are only necessary for in-depth exploration of avoidance.



303

likelihood of correctly identifyingall the messages than smokers inthe United Kingdom where 16different text messages appear onpackages. The same issue arisesin pre-post studies of a newlabelling policy. For example,

when Canada revised its labellingpolicy in 2000 to include pictures,the number of individual mes-sages doubled from eight to 16(not counting 16 additionalmessages that appeared on theinside of packages). In such

cases, neither the total numbernor the proportion of messagescorrectly identified, provide asuitable basis for comparingpolicies given that thedenominator is different. More-over, it is both time consuming

Table 5.34 Measures of Credibility and Public Support

CCoonnssttrruucctt ((aa)) HHeeaalltthh WWaarrnniinnggss –– BBeelliieevvaabbiilliittyy//CCrreeddiibbiilliittyy

MMeeaassuurree “Overall, do you believe the health warning message(s)?” (Not at all, A little, A lot)

SSoouurrccee Health Canada Youth Smoking Survey (http://www.hc-sc.gc.ca/hl-vs/pubs/tobac-tabac/yss-etj-2002/index-eng.php)


VVaarriiaattiioonnss Other alternatives refer to the accuracy, trustworthiness, credibility, believability andtrue/false nature of the warnings or the importance of information (Cecil et al., 1996; Borland& Hill, 1997a; Canadian Cancer Society, 2001; Hammond et al, 2004a; Brown et al., 2005;Health Canada, 2005; O’Hegarty et al., 2006). Some surveys have also included morecomprehensive, but also more time consuming, scales involving numerous items.

CCoommmmeennttss A useful, brief measure to examine credibility of message content. The measure can be used to examine whether different design and content features change the believability ofinformation among smokers. This question can be asked of individual health messages,such as in qualitative or experimental research, or to refer to a set of warnings, as is commonin population-based surveys. Note that responses to this item will also reflect denial, self-exempting beliefs, etc.

CCoonnssttrruucctt ((bb)) HHeeaalltthh WWaarrnniinnggss –– PPuubblliicc OOppiinniioonn//SSuuppppoorrtt

MMeeaassuurree “Do you approve of the health warnings on cigarette packages?” (Yes, No)

SSoouurrccee Borland & Hill, 1997a


VVaarriiaattiioonnss Other alternatives include measures of agreement with the warnings and references toappropriateness or desire for more information (Canadian Cancer Society, 2001; Hammond et al., 2004a; Brown et al., 2005., Health Canada, 2005)

CCoommmmeennttss This measure is a combination of previously administered questions and has yet to beadministered exactly as worded. Though measures of public support or approval may beless important as a measure of effectiveness, they are a critical measure for regulators and policy makers, and for demonstrating support for more comprehensive policies.



304

and awkward to prompt surveyrespondents for 16 differentwarnings. Finally, some knowledge mea-

sures may not work across allsurvey modalities. For example,

Krugman and Robinson presentedparticipants with diagrams ofvarious warnings in a recognitiontask (Krugman et al., 1994;Robinson & Killen, 1997). Anysuch measures, which require

visual information to be presentedto participants, must be ad-ministered either face-to-face orusing web-based modalities.

Table 5.35 Measures of Health Knowledge and Perceived Risk

CCoonnssttrruucctt HHeeaalltthh WWaarrnniinnggss –– TThhiinnkkiinngg AAbboouutt HHeeaalltthh RRiisskkss

MMeeaassuurree “To what extent, if at all, do the warning labels make you think about the health risks of smoking?” (Not at all, A little, A lot)


VVaalliiddiittyy Good face validity; good convergent validity; associated with strength of policy.

VVaarriiaattiioonnss Similar questions ask about the extent to which warnings affect the level of concern or worry about health risks.

CCoommmmeennttss A key mediator of the effectiveness of health warnings. This should be considered amongthe essential measures.

Table 5.36 Measurement of Comprehension of Emissions Information

CCoonnssttrruucctt EEmmiissssiioonnss IInnffoorrmmaattiioonn –– CCoommpprreehheennssiioonn

MMeeaassuurree “If you were to look for a safer or less harmful cigarette, would you use information about theamounts of chemicals listed on the cigarette packs to help you find a less harmful brand?” (Yes, Maybe, No)

SSoouurrcceess Gori, 1990; Health Canada, 2003


VVaarriiaattiioonnss Similar questions ask smokers to compare different tar levels of cigarettes in terms of delivery and health risks.

CCoommmmeennttss A critical measure to evaluate emission policies that include quantitative emission levels.The question can also be used to refer to specific emissions, such as tar or nicotine. Thismeasure is essential in any survey that also asks about recall or awareness of emissionnumbers on packages. The current wording can be used to refer both to descriptive (i.e. text-based) and quantitative emission information.



305

Table 5.37 Measures of Light, Mild, and Brand Descriptors

CCoonnssttrruucctt ((aa)) LLiigghhtt // MMiilldd DDeessccrriippttoorrss –– CCoommppaarraattiivvee RRiisskk

MMeeaassuurree “Light cigarettes are less harmful than regular cigarettes.”(Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly Disagree)

SSoouurrccee The ITC Project

VVaalliiddiittyy Good face validity; good convergent validity (Borland et al., 2004).

VVaarriiaattiioonnss This question can be adapted to refer to other descriptors, such as mild or smooth. In somecases, the terms light and mild are used in the same question.

Alternatives ask smokers about differences in the “tar” or “nicotine” of light versus regular cigarettes (Smokers of light cigarettes take in less tar than smokers of regular cigarettes). These measures have been widely used, but require a basic familiarity with tar and nicotine,which may not exist in all smokers in some jurisdictions (Kozlowski et al., 1998b; Shiffmanet al., 2001; Hamilton et al., 2004).

Other alternatives have asked smokers to report the number of light cigarettes that wouldneed to be smoked to equal the harm from 10 regular cigarettes; however, this approach requires a level of numerical literacy beyond the capacity of smokers in many jurisdictions (Kozlowski et al., 2000; Shiffman et al., 2001).

CCoommmmeennttss:: This is an essential construct, although there is no single “gold standard” question for itsmeasurement. The recommended measure has been selected because is it the most direct and may be most appropriate for smokers in low- and middle-income countries.Nevertheless, the question may need to be preceded by a general awareness questions (e.g. “Have you ever heard of light cigarettes?”) in some markets or rural areas. There arealso issues with the interpretation of this measure in jurisdictions where light and milddescriptors have been prohibited.

CCoonnssttrruucctt ((bb)) LLiigghhtt//MMiilldd DDeessccrriippttoorrss –– AAddddiiccttiioonn

MMeeaassuurree “Light cigarettes are less addictive than regular cigarettes.”(Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly Disagree)


VVaalliiddiittyy Good face validity; good convergent validity (Borland et al., 2004).

VVaarriiaattiioonnss This question can be adapted to refer to other descriptors, such as mild or smooth. In somecases, the terms light and mild are used in the same question.

CCoommmmeennttss A straightforward question with the same format and response options as above. Arecommended question to address perceptions of light and mild cigarettes, but not asessential as the comparative risk question, above.



306

Constituents & emissions:

A number of studies haveexamined whether smokers canrecall the emission informationcommonly printed on the sidepanel of cigarette packages(Table 5.31) (Chapman, 1986;Cohen, 1996b; Health Canada,2003; O'Connor et al., 2006c).These items typically askparticipants to name the emi-ssions printed on packages usingunprompted recall tasks, or askthem to report the numberassociated with a particularemission (usually “tar”). The dataindicates that many smokers havea general awareness that tar and

nicotine numbers may be printedon the package, but few are ableto recall the tar or nicotine levelsprinted on their usual brand ofcigarettes. To our knowledge, nomeasures have been developedto measure smokers’ knowledgeof tobacco contents.

Affective reactions to health warnings:

Research in the field of healthcommunication indicates thatmessages with emotionallyarousing content are more likely tobe noticed and processed bysmokers (Witte & Allen, 2000).

Strong emotional responses tomessages are also associatedwith greater behaviour changewhen supportive or “efficacy”related information is also pre-sented. To date, several studieshave used measures of affectivereactions to assess the impact ofwarnings labels (Environics Re-search Group, 2000; Elliot &Shanahan Research, 2002;Environics Research Group, 2003;Hammond et al., 2004a; HealthCanada, 2006; Peters et al., 2007).These measures are common inqualitative evaluations of individualwarning labels and have beenparticularly influential in develop-ment of picture-based warnings in

Table 5.38 Measures of Brand Appeal

CCoonnssttrruucctt BBrraanndd AAppppeeaall –– HHeeaalltthh WWaarrnniinnggss

MMeeaassuurree “Do you think the new warnings make cigarettes packages look less attractive, moreattractive, or has it made no difference to their attractiveness?” (Not at all, A little, A lot)

“How often have you put your cigarette package away because you didn’t want others to seethe warning on the package? Have you done this?” (Never, Sometimes, Often)

SSoouurrccee Canadian Cancer Society, 2001

VVaalliiddiittyy Face validity.

VVaarriiaattiioonnss Alternatives refer to quality of advertisements with and without warnings, whether youthwould want to “use” the product, intentions to purchase the product in the future, and ameasure of perceived economic values of brands (Hyland & Birrell, 1979; Brubaker & Mitby,1990; Canadian Cancer Society, 2001; Willemsen et al., 2002; Thrasher et al., 2007).“Attractiveness” scales have also been used (Loken & Howard-Pitney, 1988).

CCoommmmeennttss These measures provide a straightforward evaluation of whether health warnings havealtered the general appeal of packaging. The second of the two measures has a higher threshold and represents a more distal measure of appeal, which may also tap into social norms. Both of the measures are recommended for surveys that wish to provide acomprehensive evaluation of warnings, but are not essential.



307

Table 5.39 Measures of Behavioural Outcomes

CCoonnssttrruucctt ((aa)) CChhaannggeess iinn FFoorreeggooiinngg –– HHeeaalltthh WWaarrnniinnggss

MMeeaassuurree “In the last month, have the warning labels stopped you from having a cigarette when youwere about to smoke one?” (Never, Once, A few times, Many times)

SSoouurrcceess Borland & Hill, 1997a; Hammond et al., 2007a

VVaalliiddiittyy Good face validity; convergent validity; associated with strength of policy.

VVaarriiaattiioonnss Similar measures have referred to not smoking when tempted.

CCoommmmeennttss This question has a lower “threshold” than other measures that assess the behavioural effects of health warnings.

CCoonnssttrruucctt ((bb)) RReedduuccttiioonnss iinn SSmmookkiinngg –– HHeeaalltthh WWaarrnniinnggss

MMeeaassuurree “Are you smoking any less or more as a result of the new warnings, or are you still smoking the same amount?” (Less, Same amount, No difference)


VVaalliiddiittyy Good face validity; convergent validity; associated with strength of policy.

VVaarriiaattiioonnss Similar measures have referred to not smoking when tempted.

CCoommmmeennttss The wording “as a result of the warnings” needs to be emphasized when asking thisquestion. This item is not intended to provide a precise measure of changes in consumption as a result of the warnings; changes in consumption happen in response to a wide range of related factors. However, this question does provide a good general measure of the extentto which smokers have been affected by the warnings.

CCoonnssttrruucctt ((cc)) LLiikkeelliihhoooodd//MMoottiivvaattiioonnss ttoo QQuuiitt

MMeeaassuurree “To what extent, if at all, do the warning labels on cigarette packs make you more likely toquit smoking?” (Not at all, A little, A lot)


VVaalliiddiittyy Good face validity; convergent validity.

VVaarriiaattiioonnss Alternatives refer to motivations to quit and thinking about quitting, with some differencesbetween response categories.

CCoommmmeennttss The recommended wording refers directly to the likelihood of quitting smoking, which issomewhat broader than motivation alone. In practice, however, there appears to be few differences with regards to how these measures perform in practice given the consistency of findings from similar samples. The question has the potential to provide a very good summary measure of the self-reported impact of health warnings and should be consideredwithin the core set of items to evaluate labelling policy.


several jurisdictions. Measures ofnegative emotions, including fearand disgust, have also been usedin population-based surveys andshown to predict future cessation-related behaviour (Table 5.32).Overall, measures of emotion haveconsiderable promise as a proximalmeasure of effectiveness whichcan be used in both qualitative andquantitative research.

Avoidance:

Warnings that result in unpleasantemotions may lead some smokersto avoid the warnings. Indeed,several studies indicate that aconsiderable portion of smokersmake some attempt to avoid thewarnings, including covering orhiding the warnings, using anothercase, or requesting different packsto avoid particular warnings. Insome jurisdictions, tobacco

manufacturers have been ac-cused of marketing coversspecifically intended to cover pic-ture-based warnings, promptingcalls for regulatory bans on thesale of such covers (Table 5.33)(Wilson et al., 2006). Although avoidance behaviours

may be undesirable to someextent, these examples of fearcontrol behaviour do not neces-sarily reflect an adverse outcomeor inherent weakness of packagewarnings. Research has demon-strated that avoidant behavioursand attempts at thought sup-pression often have the oppositeeffect of increasing the presenceof the unwanted thoughts(Wegner, 1994). In the context ofthe warning labels, avoidantbehaviour might be more rea-sonably interpreted as a measureof effectiveness. Indeed, if thewarnings were ineffective in

communicating the threateningconsequences of smoking therewould be no reason to avoid them.Furthermore, one study found thatsmokers who attempted to avoidthe warnings were no less likely tosee the warnings, think aboutthem, or engage in cessationbehaviour at a 3-month follow-up(Hammond et al., 2004a).

Credibility & public support:

In order to be effective, the healthinformation presented in warningsmust be credible. The credibility ofwarnings relates not only to thehealth information contained in awarning, but also to its design andsource or attribution. Some haveeven speculated that there may bea trade-off between the vividnessof the information in healthwarnings and its credibility amongsmokers. In others words, if

308


Table 5.39 Measures of Behavioural Outcomes

CCoonnssttrruucctt ((dd)) QQuuiitt AAtttteemmppttss && AAbbssttiinneennccee

MMeeaassuurree “To what extent, if at all, were the following reasons for your current quit attempt…warning labels on cigarette packages?” (Not at all, A little, A lot)


VVaalliiddiittyy Good face and convergent validity.

VVaarriiaattiioonnss Alternatives have also asked about the effect of the warnings on staying quit in the future.This question can be asked as part of a list of reasons for quitting, which provides some useful context on the relative influence of other potential influences on quitting.

CCoommmmeennttss Retrospective measures, such as this, should be interpreted with caution given that they are subject to recall biases, particularly as the time since the quit date increases. In addition,smokers often cite a number of complementary reasons for quitting and endorsement ofthis item does not mean that the quit attempt is solely attributable to health warnings.



309

pictures and text become toostriking or graphic, smokers maybegin to question the accuracy ofthe information and become moreresistant to the messages. Although some validated

scales have been used toevaluate the believability of healthwarnings (e.g. Beltramini, 1988;Loken & Howard-Pitney, 1988;Cecil et al., 1996), many studieshave used single questions withface validity (Borland & Hill,1997a; Canadian Cancer Society,2001; Hammond et al., 2004a;Brown et al., 2005; HealthCanada, 2005; O’Hegarty et al.,2006; Peters et al., 2007).Together, the findings suggestthat health warnings represent acredible source of information,particularly when attributed to awell-respected department ofhealth, or a well-respected non-governmental authority, such as acancer society (Guttman & Peleg,2003; Health Canada, 2003; BRCMarketing & Social Research,2004). The levels of credibility donot appear to be associated withthe type or design of warninglabels; just like for text-basedwarnings, smokers report highlevels of believability for graphicpicture-based warnings as well. Several studies have also

sought to assess general mea-sures of public support for healthwarnings (Borland & Hill, 1997b;Brown et al., 2005; Hammond etal., 2004a; O’Hegarty et al., 2006).To our knowledge, two items havebeen developed to examine sup-port among smokers for emissionlabelling (“Overall, do you believethe health warning message(s)?”

and “Do you approve of the healthwarnings on cigarette packages?”)(Health Canada, 2001; HealthCanada, 2003). Public opiniondata may be particularly effectivefor policy makers in gaugingpolitical support for new or existinglabelling polices (Table 5.34).

HHeeaalltthh kknnoowwlleeddggee && ppeerrcceeiivveeddrriisskk

The primary objective of cigarettewarning labels is to communicatethe health effects from smoking.Thus, measures of health know-ledge and perceived riskrepresent critical components inany evaluation of health warnings(Table 5.35). To date, studieshave taken two main approachesto measuring the impact ofwarnings on health knowledge.One approach is to ask par-ticipants to self-report whetherhealth warnings have changed theextent or frequency with whichthey think or worry about thehealth effects of smoking. Alter-natively, some studies haveassessed health knowledgedirectly and examined changesover time or across jurisdictions inlevels of knowledge. Given thenumber of health effects causedby smoking, we are unaware ofany study that has attempted tomeasure a complete list. How-ever, studies typically measurebeliefs about a range of specifichealth effects to determineknowledge levels. Some studieshave included “bogus” healtheffects in the list in order to identifyresponse bias. Most lists include“major” health effects, such as

lung cancer and heart disease, aswell as health effects onnonsmokers, and lesser-knownhealth effects. Including lesser-known health effects can beparticularly effective in attributingchanges in knowledge to specificlabelling policies. Ideally, longi-tudinal studies, assessingchanges in health knowledge,would also select the healtheffects based upon the effects thatare targeted in the warnings. Inother words, studies shouldinclude health effects that: a) arealready included on packages atbaseline (before policy change)and will remain on packages atfollow-up; b) health effects that arenot on packages at baseline, butwill appear at follow-up; and c)health effects that are not onpackages at either baseline orfollow-up. This type of designprovides a measure of specificitywith respect to changes inlabelling policies. A similar approach has been

taken with respect to emissioninformation. At least one study hasexamined whether knowledge ofthe emissions in tobacco smoke ishigher in jurisdictions where theyare printed on the package(Hammond et al., 2006a). As withhealth effects, lists should includeemissions that are, and are not,printed on packages, in order toexamine the specificity of the effect.Overall, research conducted to

date suggests that increases inthe size, number, and content ofwarnings are associated withgreater thoughts about the healthrisks of smoking (Health Canada,2005; Hammond et al., 2007a).



310

More prominent warnings havealso been associated withincreased knowledge for specifichealth effects (Borland &Hill,1997a; Hammond, 2006a).Most of these findings derive frompopulation-based surveys, al-though one study reportedsignificantly higher beliefs abouthealth effects following presen-tation of graphic versus textwarnings within an experimentalsetting (O’Hegarty et al., 2006).

Constituents & emissions:

A number of studies have soughtto examine the extent to whichsmokers understand and interpretquantitative cigarette emissioninformation (Table 5.36). Thesestudies ask smokers to reporteither the “meaning” of thenumbers, or the extent to whichthe numbers translate into dif-ferences in exposure fromdifferent brands (Gori, 1990;Cohen, 1996a; Health Canada,2003; Thompson et al., 2006).Other questions ask smokers topredict the health consequencesof different tar levels, withoutexplicit reference to labelling poli-cies (Gori, 1990; Cohen, 1996a).Indeed, a number of studies on thistopic were conducted in the USA,where there are no mandatoryrequirements to print emissionlevels on packages, they appear onpackages less than 15% of the time,and are at the discretion of themanufacturer (Davis et al., 1990). Regardless of the jurisdiction

or the labelling policy, the findingsindicate that smokers have verylittle or no understanding of the

meaning of the emission levels,although a substantial proportionassociate health benefits withlower numbers. This type of datais critical to place measures ofknowledge into context; prominentlabelling that succeeds in in-creasing knowledge of emissionlevels is of little value if smokersdo not understand the meaning ofthese numbers. Indeed, the dataappear to indicate that com-municating quantitative emissionlevels promotes erroneous per-ceptions about exposure levelsand health risks that can beexpected from different products.In general, this set of findingsunderscores the importance ofassessing more than basic recallof information (Figure 5.30).

Light & mild descriptors:

A variety of surveys haveexamined perceptions of “light”and “mild” brand descriptors(Kozlowski et al., 1998b; Kozlow-ski et al., 2000; Ashley et al., 2001;Shiffman et al., 2001; Etter et al.,2003c; Borland et al., 2004;Hamilton et al., 2004). Bothquantitative and descriptive mea-sures have been used to assessthe health consequences ofsmoking “light/mild” cigarettes.Several studies have askedsmokers how many light or ultra-light cigarettes would need to besmoked to inhale the equivalentlevel of tar as regular cigarettes.Some of these measures used “10cigarettes” as a reference point,whereas others were open-ended.Smokers have also been asked tomake comparisons between “light/-

ultra-light” and “regular” brandsusing qualitative or descriptivecategories to describe exposurelevels and health risks. Thesequalitative response categorieshave also been used to compareperceived sensory properties andaddiction levels of “light/mild”cigarettes compared to “regular”brands. At least one study com-bined items to create a “sensory”index and a “health effects” index(Shiffman et al., 2001). Overall,both qualitative and quantitativemeasures appear to yield similarfindings, and indicate that asubstantial proportion of smokersperceive health benefits fromcigarettes with “light” and “mild”descriptors (Table 5.37). At least one study, the Inter-

national Tobacco Control PolicyEvaluation Survey (the ITC Pro-ject) (Borland et al., 2004), hasadopted an alternative approachto comparative estimates. Ratherthan asking smokers to compare“regular” and “light” cigarettes,participants were asked tocompare their “usual” brand withregular cigarettes (e.g. “Do youthink that the brand you usuallysmoke, [current brand], might be alittle less harmful, no different, ora little more harmful, compared toother cigarette brands?”).Separate items were used tocollect the name, descriptors, andrelevant attributes of participants’“usual” brand. This approach hasthe benefit of personalizing thequestion, and is particularly usefulto implement following theremoval of “light” and “mild” terms,at which point questions withdirect reference to “light” and



311

PPOOLLIICCYY--SSPPEECCIIFFIICCMMEEDDIIAATTOORRSS


IInnddiirreecctt eeffffeeccttssIndustryresponse

PPOOLLIICCYY

EEmmiissssiioonn iinnffoorrmmaattiioonn

OOUUTTCCOOMMEESS

SSaalliieennccee &&pprroocceessssiinngg

Awareness ofinformationKnowledge ofinformation


Changes inconsumptionQuit attemptAbstinence

BBrraanndd sswwiittcchhiinngg

BBeelliieeffss aabboouuttccoonntteennttss aannddeemmiissssiioonnss

GGeenneerraall ppeerrcciieeiivveedd rriisskk

Perceptions of relativerisk between products

QQuuiitt iinntteennttiioonnss

MMooddeerraattoorrss•• PPuubblliicc eedduuccaattiioonn aabboouutt rraattiioonnaallee aanndd mmeeaanniinngg

•• VVaarriiaanntt ssmmookkeedd

FFiigguurree 55..3300 CCoonncceeppttuuaall FFrraammeewwoorrkk ffoorr tthhee EEvvaalluuaattiioonn ooff EEmmiissssiioonnss aanndd CCoonntteennttss LLaabbeelllliinngg PPoolliicciieess



312

“mild” cigarettes become awkwardand confusing. The question alsohas a broader frame of reference.This is an advantage in the sensethat it captures the effect of otherpotential misleading descriptors orproduct elements. The dis-advantage is that information onthe respondents’ own brand mustalso be available (see Section5.4), and there is less specificitywith respect to the brand elementsthat underlie differences inperceptions of risk. A similarconceptual approach has recentlybeen taken with respect toevaluating print advertisements.Rather than asking smokers tocompare the risks implied by theexpressions “light” versus “regu-lar” cigarettes, respondents wereasked to rate the perceived risk totheir health derived from adver-tisements for different products,and the ratings for advertisementsof “light” versus “regular” ciga-rettes were compared (Hamiltonet al., 2004). In most cases,follow-up questions may benecessary to identify whichspecific elements underlie per-ceptions of reduced harm.Descriptors other than “light/-

mild” are likely to receiveincreased attention in the comingyears, particularly within juris-dictions where “light/mild” termshave already been prohibited. Toour knowledge, only one studyhas developed measures toevaluate health perceptions basedon other brand descriptors,including the words “smooth” and“ultra” (Thompson et al., 2006).Furthermore, studies with a focusupon brand descriptors in juris-

dictions that have banned “light”and “mild,” may wish to consideradditional measures that examinethe substitution of terms in theirplace. Market-based research,such as cataloguing the infor-mation printed on packages, canprovide “objective” data on thesubstitution of terms which may behelpful in interpreting self-reportedbrand data (see Section 5.4). Largely, the selection of

measures in this area may dependupon the current state of policymore so than other areas (Figure5.31).

BBrraanndd aappppeeaall

Health warnings target psycho-social variables other thanperceived risk and health know-ledge. More recent labellingpolicies include themes of addic-tion, industry manipulation,aesthetic costs, financial costs,and cessation beliefs, amongothers. A range of psychosocialmeasures have been developedto assess each of these con-structs, although these measureshave rarely been used to evaluatewarning labels.One area that has been ex-

plored is the impact of healthwarnings on measures of brandappeal (Table 5.38). In theory,replacing brand imagery withhealth warnings has the potentialto change perceptions of thecigarettes and packaging. Todate, the limited findings in thisarea appear to support thishypothesis, although it has yet tobe explored in much depth withrespect to warnings on packages

(Hyland & Birrell, 1979; Loken &Howard-Pitney, 1988; Brubaker &Mitby, 1990; Hammond et al.,2004b; Thrasher et al., 2007).Future research might alsoexplore whether larger graphichealth warnings undermine thevisual appeal of cigarette displaysat retail outlets.

BBeehhaavviioouurraall oouuttccoommeess

There are several approaches topredicting “downstream” cessa-tion-related outcomes from healthmodels. As with health effects,some studies have used mea-sures of processing andknowledge of the warnings, andmodelled their effects on moti-vation to quit and patterns ofsmoking behaviour (see Section3.1 for measures of tobacco useand Section 3.2 for psychosocialoutcomes). This has producedsignificant findings in longitudinalstudies to date (Hammond et al.,2003). However, this approach issomewhat limited when it comesto evaluating changes in healthwarnings. Unless both surveywaves are conducted when thesame set of health warnings is onthe pack, the baseline measuresof processing or knowledge relateto the “old” warnings, whereas anycessation-related activity at follow-up presumably reflects the impactof the “new” warnings. An alternate strategy that can

also be used in cross-sectionalstudies is to ask smokers todirectly report the extent to whichwarnings have influenced theirmotivation to quit and smokingbehaviour (Borland & Hill, 1997a;



313

PPOOLLIICCYY--SSPPEECCIIFFIICCMMEEDDIIAATTOORRSS


IInncciiddeennttaall eeffffeeccttss

Industry use ofalternativedescriptors

PPOOLLIICCYY

RReemmoovvaall ooffddeessccrriippttoorrss ssuucchh aassLLiigghhtt aanndd MMiilldd

OOUUTTCCOOMMEESS

SSaalliieennccee &&pprroocceessssiinngg

Awareness ofdescriptors


Changes in consumptionQuit attemptAbstinence

BBrraanndd sswwiittcchhiinngg

BBeelliieeffss aabboouuttddeessccrriippttoorrss//bbrraannddss

AddictionHealth risks

TTaasstteeSSeennssoorryy pprrooppeerrttiieess

GGeenneerraall ppeerrcceeiivveeddrriisskk

Perceptions ofrelative risk between

productsQQuuiitt iinntteennttiioonnss

MMooddeerraattoorrss

• IInnffoorrmmaattiioonn pprroovviiddeedd bbyy hheeaalltthh ggrroouuppss aanndd ggoovveerrnnmmeenntt oonnrraattiioonnaallee ffoorr rreemmoovvaall

• Brands smoked

FFiigguurree 55..3311 CCoonncceeppttuuaall FFrraammeewwoorrkk ffoorr tthhee EEvvaalluuaattiioonn ooff BBrraanndd DDeessccrriippttoorr PPoolliicciieess



314

Canadian Cancer Society, 2001;Health Canada, 2005; Koval et al.,2005; Willemsen, 2005; O’Hegartyet al., 2006). This approach doesnot have the same validity in termsof measuring actual changes insmoking behaviour, although itcan be used to examine changesacross labelling policies. A third alternative is to examine

changes in prevalence rates, orpopulation-based cessation acti-vity, before and after the imple-mentation of new warnings. To ourknowledge, this approach hasbeen used in only one study todate: Gospodinov & Irvine (2004)reported no discernable changes inprevalence rates, and a reductionof two cigarettes per week amongsmokers in the months followingthe implementation of pictorialhealth warnings in Canada. How-ever, as described earlier in thissection, there are serious problemsin attributing changes in nationallevel trends to changes in healthwarnings, or any other individualpolicy measure. Indeed, asGospodinov & Irvine note, therewere significant changes in priceover the same period of time, aswell as considerable sub-nationaltobacco control activity over thesame time period.Yet another approach to mea-

suring the impact of warnings oncessation behaviour has been tolook at changes in the use ofcessation services as they relateto information on warnings labels.Research conducted in the UKand the Netherlands has exa-mined changes in the usage ofnational telephone helplines afterthe contact information was in-

cluded in package healthwarnings. Each of these studiesreports significant increases in callvolumes (Willemsen et al., 2002;Department of Health, 2006). Finally, several items have

been created for use amongformer-smokers. Typically, theseitems ask about various reasonsfor quitting, including whether thehealth warnings either motivatedthem to quit or have helped themto remain abstinent (CanadianCancer Society, 2001; Hammondet al., 2004b; O’Hegarty et al.,2006; Thompson et al., 2006).These measures are, however,subject to recall bias and shouldbe interpreted with particularcaution (Table 5.39).

FFoorrmmaattiivvee rreesseeaarrcchh

Formative research is oftenconducted to help identify thecontent and design of new healthwarning policies. Regulators mustdecide what health effects tocommunicate, how many, andhow to present this information tosmokers on the package. Al-though population-based surveysmay help to guide these decisions,qualitative research is typicallyundertaken as part of the policydevelopment process. The most common approach

has been to conduct a series offocus groups (i.e. semi-structuredinterviews conducted within agroup setting). Focus groups havetwo important advantages overpopulation-based surveys: 1)participants can be presented withvisual stimuli, including examplesof health warnings in a way that is

not possible with telephone basedsurveys; and 2) focus groups arewell suited to open-ended ques-tions and allow for more in-depthdiscussion than structured sur-veys. In many cases, focus groupsare also used as a way to evaluatethe effectiveness of healthwarnings on sub-groups, includingyounger smokers and those fromlower socio-economic groups. Theprimary disadvantage of focusgroups is that the findings can behard to summarize in a systematicfashion, which complicates com-parisons across groups andsettings. As a result, conventionalvalidity tests for quantitative datacan not be conducted with focusgroup findings. Nevertheless,qualitative findings help to com-plement quantitative research inthis area, and represent animportant step in the developmentof new labelling policies. Qualitative research has

examined many of the samethemes as population-based sur-veys, and other quantitativemethods. These include generalknowledge of the warnings, suchas the content and location, theemotional impact of warnings, aswell as their general salience andnoticability (Environics ResearchGroup, 2000; Elliot & ShanahanResearch, 2002; CRÉATEC, 2003;BRC Marketing & Social Research,2004; Health Canada, 2006). Inmany cases, these studies havepresented different health warningsto participants in order to makedirect comparisons betweenlabelling policies. These designshave proven particularly effective atcomparing the emotional reactions


315


elicited by picture versus textwarnings, for example (EnvironicsResearch Group, 2000; Elliot &Shanahan Research, 2002; BRCMarketing & Social Research,2004). Focus groups have alsoprovided critical informationregarding the meaning andcomprehension of the informationcommunicated in labelling policies.For example, focus groupmeasures developed, on behalf ofHealth Canada, have helped todemonstrate that, even thoughmost Canadian smokers are awareof emission information on the sideof packages, very few understandthe actual meaning of the infor-mation (Environics ResearchGroup, 2003). Indeed, judging bythe findings of the focus group,most Canadian smokers aremisusing the emission infor-mation. Thus, carefully con-structed focus group measures canprovide “deeper,” more com-prehensive measures of meaningthat are difficult to ascertainthrough structured population-based surveys.

IInndduussttrryy ddooccuummeennttss

Internal tobacco industry docu-ments represent a potentially richsource of information about theeffectiveness of tobacco controlpolicies. There are several infor-mative reviews of industryactivities and documents on pro-duct labelling, including manyrelated to brand descriptors suchas “light” and “mild” (Slade, 1997;Pollay, 2001; Pollay & Dewhirst,2002; Wakefield et al., 2002;Chapman & Carter, 2003;

Alechnowicz & Chapman, 2004).However, to date, no com-prehensive review of packagingissues related to labelling policieshas been undertaken.

SSuummmmaarryy

Few of the measures used toevaluate warning label policieshave undergone formal psycho-metric analyses. Much of theliterature in this area has beenconducted on behalf of regulators,which may account for the lack of“formal” tests of validation morecommon to academic research. Inaddition, different studies haveused different measures to assessthe same construct. In manycases, measures differ in thewording of questions and in thetime references used in mea-sures, such as noticing andawareness. This complicatescomparisons across surveys andacross labelling policies. However,most measures have high facevalidity and several have showngood predictive validity for down-stream outcomes, includingknowledge of health effects andself-reported motivation to quit,and cessation behaviours. Inaddition, the consistency of thefindings across studies and surveymodalities suggests that thedifferences in the measures haveonly a modest effect on outcomesof interest. Nevertheless, virtuallyall of the constructs would benefitfrom further developmental work,including the standardization ofthe wordings across surveys.

Implications for study design &analysis:

No single study research design isadequate to evaluate the impact oflabelling policies. Given the chal-lenges inherent in evaluatingnational level policies, individualstudies are inevitably subject to arange of limitations. However,when taken collectively, the rangeof designs constitute a persuasivebody of evidence demonstratingthe effectiveness of compre-hensive health warnings. Quali-tative methods, including focusgroups, are essential for informingthe early stages of design andgenerating new insights intolabelling policies. Experimentalresearch is best suited to drawingdirect comparisons acrosswarnings and to isolating theeffectiveness of individual designand content features. For thisreason, experimental researchprovides the highest level ofinternal validity. Alternatively,population-based surveys havethe highest external validity andmay provide the most com-prehensive measures of effec-tiveness given adequate designs.External validity is particularlyimportant in the case of warninglabels, which operate overrepeated exposures that are tiedto smoking behaviour. The patternof exposure is the defining featureof product warnings and one thatis impossible to replicate in a“laboratory” environment. As aresult, the central question ofwhether labelling policies in-fluence beliefs, attitudes, andbehavioural change can only be



316

assessed with population-basedsurveys. The inferences that canbe made from these surveys areconsiderably enhanced withinlongitudinal and quasi-experi-mental designs, as discussed inSection 2.1.

Priorities for future work:

As countries begin to implementrestrictions on misleading pac-kaging elements, research mustbegin to examine elements otherthan “light” and “mild” brand des-criptors. These include otherpotentially misleading elements,such as the use of colour-codingand package designs that falselyconvey differences in strength. Todate, very limited work has beenconducted outside the tobaccoindustry on these issues. There isan immediate need to developmeasures that can examine theseissues within population-basedsamples, especially within juris-dictions where “light” and “mild”descriptors have already beenprohibited.A second priority for future

research is to examine contentsand emission information moreclosely. Up to now, much of theexisting research has focussedupon awareness and under-standing of ISO tar and nicotinenumbers. There is an urgent needfor measures to evaluate newapproaches to communicatingcontents and emission infor-mation. Population-based studiesshould be conducted withinjurisdictions that have developednovel policies, such as com-municating emission information

using descriptive, rather thanquantitative means. Greaterexperimental and qualitative workmust also be undertaken toexplore how smokers interpretand use this information, and tocompare different approachesmore systematically. These issuesare directly relevant to the ongoingdebate regarding how tocommunicate the risks of com-bustible versus non-combustibletobacco products. Historically,emission information has beenused by smokers to evaluate therelative risks of different products.As emission and content labellingpolicies are developed for the fullrange of tobacco products,regulators will need to considerthe delicate issue of whatfundamental message they wishto communicate to smokers.Quantitative emission and contentinformation will inevitably beinterpreted as indicators of risks,unlike descriptive information thatis uniform across products.In addition to developing new

survey measures, existing mea-sures must be administered morewidely, as a greater number ofcountries prepare to implementthe provisions within Article 11 ofthe FCTC. In particular, few of themeasures reviewed in this sectionhave been assessed amongsmokers in low- and middle-income countries. Finally, measures should be

developed to examine the impactof the cessation information that isincluded in many labelling policies.Cigarette packages are amongthe most prominent vehicles fordisseminating cessation services

and efficacy-related information.These measures may includesurvey based measures, as wellas indicators from other datasources, such as usage rates fromtelephone quitlines or web-basedservices.

RReeccoommmmeennddaattiioonnss

Comprehensive evaluations ofhealth warning labels shouldinclude recommended items fromeach of the key constructs (seeabove). Population-based sur-veys, seeking a more limitedevaluation of health warnings,should include proximal measuresof noticing, along with inter-mediate measures of perceivedrisk or health knowledge. Althoughmeasures of general awarenessand knowledge of health warningscan be informative, these mea-sures should be used with cautionfor the purpose of comparinglabelling policies. Evaluations of brand des-

criptors, and other packagingelements, should represent apriority for tobacco control policy. Inaddition to examining “light” and“mild” descriptors, research shouldconsider other potentially mis-leading terms, as well as brandelements such as colour andpackage design. Unlike healthwarnings, these policies require theremoval of information from thepackage and present challenges inthe wording of survey measures.There is an immediate need todevelop measures that canaddress these issues as morecountries implement recomm-endations under Article 11 to


317


prohibit misleading package ele-ments. Policies to communicate emis-

sions and content information viapackages, also present uniqueevaluation challenges. Unlikehealth warnings, measures ofsalience and processing for thistype of information are of limited

value. Rather, evaluations shouldfocus upon the meaning and useof emission and content infor-mation. Given the lack of researchin this area, and the lack ofconsensus regarding the bestpolicy approach, there is a par-ticular need for formative researchin this area.

Overall, the selection of mea-sures to evaluate tobacco labellingpolicies will depend upon themethod and scope of the evalu-ation, as well as the specific policycontext.

section5.5janvier13:Layout 1 13/01/2009 10:07 7

5.5 measures to evaluate the effectiveness of tobacco...

Documents