Rossiter

Embed Size (px)

DESCRIPTION

d

Citation preview

  • Does TV AdvertisingAffect Children?Yes and no. It does persuade them,

    but it doesn't make them more persuasible.

    John R. RossiterChildren are exposed to large amountsof TV advertising. The average childencounters about 20,000 TV commer-cials ayeai. (The average child watchestelevision for 3.6 hours a day, or about25 hours a week, of which just over 3hours, conservatively, are devoted tocommercials [Adler et al., 1977], Thevast majority are 30-second spots, lead-ing to the estimate of 20,000 potentialexposures per year. Actual exposures,in the sense of attention, would besomewhat fewer; the figure is hard toestimate because of the necessity ofmeasuring auditory as well as visualattention,) Parents and policy makershave expressed considerable concernabout the effects of this exposure onchildren. Part of the concern relates tothe short-run effects of specific advertis-ing practices on children. But much ofthe concern relates to the effects of chil-dren's exposure to TV advertising ingeneral. This paper reviews the evi-dence from research on the generalimpact of TV advertising on children.

    Two types of general effects may bediscerned: cumulative-exposure effectsand heavy-viewing effects. Cumulative-

    exposure effects refer to the impact ofTV advertising on children as they growolder and thus see more commercials.Heavy-viewing effects refer to theimpact of TV advertising on childrenwho, within age groups, are more heav-ily exposed to commercials than theirpeers.

    In assessing cumulative-exposureeffects and heavy-viewing effects, it isimportant to specify the types of effectswe are talking about. This paper exam-ines three categories of specific effects:(1) cognitive effects, focusing on chil-dren's ability to understand the natureand purpose of TV advertising; (2) atti-tudinal effects, focusing on children'sfeelings toward TV commercials andtheir reaction to TV advertising as asocial institution; and (3) behavioraleffects, focusing on the extent to whichchildren are persuaded to want, and toask for, advertised products.

    Cumulative-Exposure Effects

    During his or her daily three and a halfhours of TV viewing, the average Amer-

    ican child is exposed to about 55 com-mercials. This amounts to approxi-mately 20,000 commercials a year, or amassive 200,000 commercials, in total,from ages 2 to 12, What are the effectsof this cumulative exposure?

    Cognitive Effects. Perhaps the mostimportant cognitive effect is children'sability to understand commercials as afunction of cumulative experience, Ros-siter and Robertson (1976) examinedchildren's understanding of the concep-tual basis of television commercials interms of six variables that measuredchildren's ability to (1) define the differ-ence between television commercialsand television programs; (2) compre-hend the existence of an external mes-sage source or sponsor; (3) perceive theexistence of intended target audiencesfor commercial messages; (4) identifyinformative intent in commercials; (5)identify persuasive intent in commer-cials; and (6) understand their symbolicrepresentational characteristics. Totalcognitive understanding of commercialswas highly correlated with age {r = .63,p < .001). Further analysis demon-strated that age and cumulative experi-

    49

  • Journal of Advertising Researchence accounted for 40 per cent of thevariance in cognitive understanding,while differences in social backgroundaccounted for only 9 per cent. Using asimilar but partial set of measures.Ward, Wackman, and Wartella (1975)also found a highly significant increasein children s cognitive understanding ofcommercials as a function of age.

    It is clear that, for the average child,cumulative exposure to commercialshas a definite positive effect on his orher cognitive understanding of whatcommercials are and what they are sup-posed to do. This effect may be due, ofcourse, not only to cumulative experi-ence, but also, following Piaget"s the-ory, to age-related increases in chil-dren's general cognitive abilities.Nevertheless, it is reasonable to regardthese cognitive measurement results asevidence against the proposition thatchildren become more "mentally" sus-ceptible to TV advertising techniques

    John R. Rossiter Is an assistant profes-sor of marketing at the Wharton School,and a faculty member at the AnnenbergSchool of Communications, Universityof Pennsylvania. He holds a B.S. degreein psychology from the University ofWestern Australia, an M.S. in market-ing from UCLA, and a Ph.D. in commu-nications from the University of Penn-sylvania. He is co-author of a recentNational Science Foundation report onchildren and TV advertising and is cur-rently co-authoring a book on communi-cation models and advertising strategy.

    as they are exposed to more TVadvertising,

    Attitndinal Effects. Apart from thequestion of whether children understandcommercials is the question of whetherchildren like television commercials andhow these "institutional" feelingschange with age. An impressive body ofstudies has documented that children'soverall attitudes toward commercialsdecline significantly with age (James,1971; Blatt, Spencer, and Ward, 1972:Robertson and Rossiter, 1974; Fergu-son, 1975; Bever et al.. 1975). Forinstance, the percentage of children whoindicated that they liked all commercialswas shown in one study to decline from69 per cent at first grade to 56 per centby third grade and to 25 per cent by fifthgrade (Robertson and Rossiter, 1974).The decline in children's attitudes seemsto apply to specific commercials as wellas to commercials in general. Atkin(1975a) measured children's overall lik-ing for three specific commercials andfound a negative correlation of r = .35(p < .001) with age.

    Behavioral Effects. An obvious ques-tion, given the increase in children'scognitive understanding of commercialswith age and their increasingly negativefeelings toward them, is whether com-mercials have any less behavioralimpact on children as the youngstersgrow older. This is not an easy questionto answer. One problem is the distinc-tion between intended behavior andactual behavior. Commercials mayinstill intentions or desire for an adver-tised product, but the execution of theseintentions in most cases requires thechild to make requests to parents. Thisprocess introduces other variables perti-nent to the requesting behavior, andthese other variables may have little todo with the impact of the commercialsper se. Consequently, it is wise to exam-ine not only requests or actual behavior,but intended behavior as well.

    The evidence on intended behavior asa function of age is dependent on thenature of the reseai'ch measure. Forexample, Robertson and Rossiter (1974)asked children whether they wanted allproducts they saw advertised on televi-

    sion. Responses indicated the expectedage-related decline: 53 per cent said yesat first grade, 27 per cent at third grade,and only 6 per cent at fifth grade. However, a more moderate question used byWard, 'Wackman, and Wartella (1975)asked children whether they wantedmost things shown in television com-mercials. A slightly broader age rangewas sampled, making the results evenmore interesting: 66 per cent of kinder-garten children said yes, 51 per cent olthird graders said yes, and a marginallylower 49 per cent of sixth graders saidyes. Moreover, when asked whethercommercials "made them want to havethings, " the children in Ward, Wackman, and Wartella's study exhibited anapparent increase in perceived motiva-tion with age: affirmative answers bvage gi'oup were kindergarten, 67 percent: third grade, 87 per cent: and sixthgrade, 84 per cent. Evidence for ;idecline in advertising-induced intentionswith age and cumulative exposure is,therefore, slight at best.

    Evidence on the actual behavioraleffects of TV advertising as a function ofage and cumulative exposure requirescareful interpretation. Because childrendo not have the means to buy manvproducts and, therefore, must ask theirparents to buy products for them, themost widely applicable behavioraleffects measure is request frequencyRobertson and Rossiter (1974) found adecrease in request frequency with age:however, their study focused on preChristmas television advertising inwhich almost half of the advertisedproducts were toys and games. Theseitems may simply be of less interest aschildren grow older. This criticism ma\also be applied to the frequently heardinterpretation of a classic study by Wartiand Wackman (1972) to the effect tha*.children's request frequencies declinewith age. (Ward and Wacknjan in facifound only a nonsignificant tendency forrequest fi-equencies to decline with age1/ =^ .13: an / of =.16 would havibeen required for significance at the .>per-cent level].) Detailed inspection olthe product-by-product data in Wardand Wackman's study indicates

    50

  • Volume 19, Number 1, February 1979

    trends by age depending on the productin question. For example, requests fortoys declined with age, but requests forbicycles increased. Requests for prod-ucts that are television advertised andpresumably relevant to children of allages, such as snack foods and softdrinks, did not show any age-relateddecline in request frequency. This non-decline effect was substantiated in alater study by Ward, Wackman, andWartella (1975) in which requests forfood products were essentially constantacross age groups (x ,^ n.s.) and requestsfor "child-relevant" products, such astoys and games, clothing, and recordalbums, actually increased significantlywith age (x^ P < -01). Ward and Wack-man's (1972) study is much more com-prehensive in terms of product cate-gories and should probably be favoredover Ward, Wackman, and Wartella's(1975) results. Neither study, however,revealed any general decline in requestfrequencies as a function of age andcumulative exposure to commercials.

    Before interpreting request frequen-cies as data on "effects," we must con-sider the problem referred to earliernamely, that requesting behavior isunder the control of factors other thanadvertising-induced intent.

    One obvious factor is the extent towhich parents acquiesce to children'srequests and thus reinforce this form ofbehavior. Ward and Wackman's study(1972) found that acquiescenceincreases with the child's age (r = .20, p< .01). To provide two illustrativetrends for products relevant to childrenof all ages, the percentage of motherswho said they usually yield to children'srequests for snack foods was 52 per centfor 5 to 7 year olds and rose linearly to77 per cent for 11 and 12 year olds;comparable acquiescence figures forsoft-drink requests were 38 per cent and54 per cent. An altemative hypothesis isthat peer influence or other nonadvertis-ing experiences might induce requestbehavior. However, Ward and Wack-man's results indicated that the role oftelevision atlvertising in the requestingprocess was relatively constant acrossage groups {r = -.14, n.s.). That is.

    television advertising did not seem to bedisplaced by peer influence or otherfactors.

    It, therefore, appears that parentalacquiescence, a form of reinforcement,is the main factor that "confounds"request frequencies as a measure ofadvertising effects. Because parentalacquiescence increases with age, weshould discount the earlier implicationof a constant advertising-inducedrequest level. A more accurate projec-tion, adjusting for parental acquies-cence, would be a slight decline inrequest levels. Thus, the safest conclu-sions from the available evidence arethat the cumulative-exposure effect ofcommercials is to reduce children'sintentions or desire for advertised prod-ucts only slightly and to produce a simi-larly slight reduction in the frequency ofadvertising-induced requests to parents.

    How is the slight decline in behavioraleffects with age to be reconciled withthe marked increase in children's cogni-tive understanding of commercials andthe equally marked decline in theiraffective feelings toward commercials?The most parsimonious explanationwotild be abandonment of the implicithypothesis that children who under-stand commercials better and like themless will be less affected by them. Thishypothesis assumes that commercialsare in some way "bad," that they arenot supposed to persuade children towant the advertised products, or thatchildren become poorer judges of adver-tised products as they grow older. Allare questionable assumptions. I shallreturn to this point in the summary.

    Heavy-Viewing Effects

    The analysis of cumulative-exposureeffects was based on the average child'sexposure to TV advertising, using dataaveraged by age level. Although theaverage child watches TV for three anda half hours a day, this figure rangesfrom an hour or less per day for lightviewers to as many as six hours per dayfor heavy viewers (Adler et at., 1977).The range for potential exposure to

    commercials is thus about 14 to 86 perday, or about 5,400 to 32,600 per year,depending on the child's viewing habits.Do heavy viewers, within age groups,differ from light viewers in terms of theimpact of TV advertising? Once again,the answer is examined in terms of cog-nitive, attitudinal, and behavioraleffects.

    Cognitive Effects. Amount of televi-sion viewing at a given age level has noeffect on children's ability to understandthe nature and purpose of TV commer-cials. It may be recalled that Robertsonand Rossiter's (1974) study of children'scognitive understanding of TV commer-cials found significant effects by age (40per cent of the variance) and by socialbackground (9 per cent). Analysis ofcognitive understanding by viewinglevel produced a nonsignificant correla-tion of r = .20, with differences inviewing level accounting for a negligibleamount of the variance in children'scognitive understanding of TV commer-cials (4 per cent). Heavy viewers, there-fore, are no moreand no lesssophis-ticated regarding TV advertising thantheir lighter-viewing peers.

    Attitudinal Effects. Heavy viewerstend to have more favorable attitudestoward television commercials. Chil-dren's attitudes toward commercialsbecome more negative as they growolder, but heavy viewers within agegroups hold more positive attitudes thantheir peers. For example, Rossiter andRobertson (1976) found a significant cor-relation of /' = .24 (p < .05) betweenattitudes and television exposure withage held constant (the measure of atti-tudes in this study included trust, liking,and behavioral-intention scales). A simi-lar correlation of ; = .30 (p < .01) wasfound by Atkin (1975a). For older chil-dren at least, it seems improbable thatliking commercials "causes" televisionviewing. Surely, interest in programsleads to viewing. Therefore, it is a fairlysafe assumption that, for commercials,the direction of cause and effect is fromexposure to attitudes, so that heavytelevision viewing causes children tolike TV commercials.

    Behavioral Effects. An intetesting

    51

  • Journal of Advertising Researchstudy of the behavioral effects of heavyTV advertising was conducted by Rob-ertson and Rossiter (1976) in conjunc-tion with pre-Christmas TV promotionsfor toys and games. During Novemberand December, TV commercials fortoys and games comprise about 50 percent of children s advertising; thus,even normal viewers become de factoheavy viewers for this product category.Using Christmas-present choices as thebehavioral-effects measure, Robertsonand Rossiter found that children'schoice of toys and games, relative toother, less advertised items, increasedby an average of 5 per cent per childover the pre-Christmas advertisingperiod, (A conservative formula wasused to compute the 5-per-cent figure.The normal "percentage increase"interpretation would have been about 8per cent to 10 per cent per child,) Thisresult was statistically significant ij)