+ All Categories
Home > Documents > NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J....

NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J....

Date post: 14-Jul-2020
Category:
Upload: others
View: 5 times
Download: 0 times
Share this document with a friend
13
NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST FOR AD MEMORY AND ATTITUDE TOWARD THE AD? NADER T. TAVASSOLI Ass/sfanf Professor ot Management Massachusetts Institute of Technology CLIFFORD J. SHULTZ, II Assistant Protessor of t<Jlarketing Arizona State University West GAVAN J, FITZSIMONS Assistant Protessor ot Marketing University ot California, Los Angeles An experiment found that viewer involvement with a television pro- gram has a positive effect on ad memory and attitude toward the ad (Aaj) as program involvement increases from low to moderate lev- els. However, further increases in program involvement beyond moderate to high levels lead to decreases in ad memory and A^^- This inverted-U relationship between program involvement and ad memory and A^j is explained by theories of arousal and has several implications for the slotting and design of commercials. A dvertising effectiveness is determined largely by xvho is reached and by hoxv the message contents are processed. Yet, decisions con- cerning ad placements are driven primarily by the former, by audience size and composi- tion. Ad placements, however, result not only in the selection of an audience but also in the se- lection of an advertising context. This context, in turn, affects how a message is processed. Indeed, understanding the impact of context on advertising effective- ness has been rated a top re- search priority among advertis- ing practitioners (see Chook, 1985) and has stimulated a grow- ing body of research. In particu- lar, a number of studies have demonstrated a significant effect of involvement with the adver- tising context on both memory for and attitudes toward adver- tisements {e.g., Anand and Sternthal, 1992; Bryant and Comisky, 1978; Kennedy, 1971; Krugman, 1983; Mattes and Can- The authors thank Adrian Hitchen and ISL Marketing for their research support and Mai Kaworski and Univision for their technical assistance. tor, 1982; Pavelchak, Antil, and Munch, 1988; Pham, 1992; Sie- bert, 1978; Soldow and Principe, 1981). This paper extends this stream of research and attempts to reconcile seemingly contra- dictory findings of previous studies. Research Focus Consider the over 2 billion television viewers of the 1994 World Cup final—the largest global TV audience in history (Reid, 1994). Ardent soccer fans were glued to their television screens and on the edge of their seats, brimming with excitement; their involvement with the game was very high. While generally enjoying soccer, other viewers did not become quite as excited by the game; their involvement with the game was moderate. Finally, many viewers watched the game out of curiosity or boredom, or simply because their friends and family watched it. These viewers were not ex- cited by a brilliant pass or a goal-scoring opportunity; their involvement with the game was relatively low. Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995 61
Transcript
Page 1: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

NADER I . TAVASSOLICLIFFORD J. SHULTZ, IIANDGAVANJ. FITZSIMONS

PROGRAM INVOLVEMENT:ARE MODERATE LEVELS BEST FOR ADMEMORY AND ATTITUDE TOWARD THE AD?

NADER T. TAVASSOLIAss/sfanf Professor

ot ManagementMassachusetts Institute

of Technology

CLIFFORD J. SHULTZ, IIAssistant Protessor

of t<JlarketingArizona State University West

GAVAN J, FITZSIMONSAssistant Protessor

ot MarketingUniversity ot California,Los Angeles

An experiment found that viewer involvement with a television pro-gram has a positive effect on ad memory and attitude toward the ad(Aaj) as program involvement increases from low to moderate lev-els. However, further increases in program involvement beyondmoderate to high levels lead to decreases in ad memory and A -This inverted-U relationship between program involvement and admemory and A j is explained by theories of arousal and has severalimplications for the slotting and design of commercials.

Advertising effectiveness isdetermined largely byxvho is reached and by

hoxv the message contents areprocessed. Yet, decisions con-cerning ad placements aredriven primarily by the former,by audience size and composi-tion. Ad placements, however,result not only in the selection ofan audience but also in the se-lection of an advertising context.This context, in turn, affects howa message is processed. Indeed,understanding the impact ofcontext on advertising effective-ness has been rated a top re-search priority among advertis-ing practitioners (see Chook,1985) and has stimulated a grow-ing body of research. In particu-lar, a number of studies havedemonstrated a significant effectof involvement with the adver-tising context on both memoryfor and attitudes toward adver-tisements {e.g., Anand andSternthal, 1992; Bryant andComisky, 1978; Kennedy, 1971;Krugman, 1983; Mattes and Can-

The authors thank Adrian Hitchen andISL Marketing for their research supportand Mai Kaworski and Univision fortheir technical assistance.

tor, 1982; Pavelchak, Antil, andMunch, 1988; Pham, 1992; Sie-bert, 1978; Soldow and Principe,1981). This paper extends thisstream of research and attemptsto reconcile seemingly contra-dictory findings of previousstudies.

Research Focus

Consider the over 2 billiontelevision viewers of the 1994World Cup final—the largestglobal TV audience in history(Reid, 1994). Ardent soccer fanswere glued to their televisionscreens and on the edge of theirseats, brimming with excitement;their involvement with the gamewas very high. While generallyenjoying soccer, other viewersdid not become quite as excitedby the game; their involvementwith the game was moderate.Finally, many viewers watchedthe game out of curiosity orboredom, or simply becausetheir friends and family watchedit. These viewers were not ex-cited by a brilliant pass or agoal-scoring opportunity; theirinvolvement with the game wasrelatively low.

Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995 61

Page 2: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M E N T

The above scenario is typicalfor many television programsand sporting events in particu-lar: groups of viewers are differ-entially involved with the pro-gram and experience differentlevels of arousal or excitement{Pavelchak, Antil, and Munch,1988; Pham, 1992). Moreover,one difference between televi-sion programs, such as action-packed dramas, documentaries,or comedies is that they differen-tially are involving or arousing{Bryant and Comisky, 1978;Kennedy, 1971; Krugman, 1983;Mattes and Cantor, 1982; Sie-bert, 1978; Soldow and Principe,1981). These differences in in-volvement raise a number of im-portant questions. Do viewers atdifferent levels of program in-volvement have the same abilityto process commercials in depth?Will their memory for the brandsadvertised be the same? Andhow is their attitude toward thead {Agj) affected? Moreover, weexamine the slotting of adver-tisements in television programsand explore if it is better to ad-vertise during the more excitingportions of a program {e.g., dur-ing the game) or when the ex-citement is lower {e.g., at half-time or following the game).

Previous research on programinvolvement is divided onwhether invoivement facilitatesor hinders the ability to processcommercials shown. Specifically,prior research has supportedboth a positive effect of programinvolvement on ad recall {Sie-bert, 1978; Singh and Churchill,1987) and A^^ {Kennedy, 1971),as well as a negative effect on adrecall {Bryant and Comisky,1978; Pavelchak, Antil, andMunch, 1988; Soldow and Princ-ipe, 1981) and A^ (Murry,Lastovicka, and Singh, 1992; Sol-dow and Principe, 1981). Ourstudy attempts to reconcile theseseemingly contradictory findingsand suggests that involvement

aids information processing as itincreases from low to moderatelevels, but that involvement hasa negative effect on informationprocessing as it increases beyondmoderate to high levels. In otherwords, we will argue that therelationship between programinvolvement and advertising ef-fectiveness is an inverted-Urelationship.

Literature Review

Dimensions of Program In-volvement. There are motiva-tional aspects or antecedents ofviewer involvement (Celsi andOlson, 1988). These are capturedby factors such as those de-scribed above: being an ardentsoccer fan, enjoying soccer ingeneral, or not knowing enoughabout soccer to become excitedby it. Involvement also has anumber of consequences {Celsiand Olson, 1988) and results in asubjective felt state {Mitchell,1979). In our example, theseconsequences are the excitementor arousal experienced by aviewer, thoughts generated inresponse to the game, or feel-ings produced at the outcome ofthe game, such as happiness ordisappointment. The focus ofthis paper is the arousal viewersexperience as a consequence ofinvolvement with a program andthe effects different levels ofarousal have on the processingof commercials.

It is a well-known fact thattelevision programs lead to vary-ing levels of arousal {Singh andChurchill, 1987). Similarly, in-volvement has been shown to beassociated with an increase inarousal {Mandler, 1975; Murry,Lastovicka, and Singh, 1992).While the concept of involve-ment may be quite broad, itsmain components are the direc-tion and the intensity of involve-ment {for a review, see An-drews, Durvasula, and Akhter,

1990). The involvement in ourresearch is directed toward theprogram and its intensity is theinternal state of arousal or ex-citement experienced by theviewer. This conceptualization ofinvolveinent intensity is a commonone {for a review, see Andrews,Durvasula, and Akhter, 1990;Burnkrant and Sawyer, 1983;Cohen, 1983; Day, 1973; Mitch-ell, 1979; Pavelchak, Antil, andMunch, 1988) and, in fact, pro-gram involvement has beenmeasured in terms of arousal{Bello, Pitts, and Etzel, 1983;Murry, Lastovicka, and Singh,1992).

Effects of Arousal on Informa-tion Processing. The effects ofarousal on information process-ing and behavior were intro-duced into consumer behaviorby Kroeber-Riel {1979). In brief,a stream of research based onthe Yerkes-Dodson Law {1908)postulates that arousal has apositive effect on task perfor-mance as it increases from aresting level to moderate levelsof arousal beyond which an in-crease in arousal causes a de-crease in performance. This in-verted-U relationship has beenwidely replicated (e.g., Berlyne,1960; Easterbrook, 1959; Man-dler, 1975) for a variety of tasks,such as memory tasks (for a re-view, see Eysenck, 1982).

The processes underlying theinverted-U relationship are stilldebated. The ascending part ofthe curve has been attributed toan increase in the natural speedof cognitive processing asarousal increases above its rest-ing level (Eysenck, 1982; Silkand Vavra, 1974; Srull, 1993) andis possibly due to an increase intotal attentional capacity to itshighest level at moderate levelsof arousal {Kahneman, 1973).The detrimental effects of higharousal appear to be caused by areduced ability to engage inshared or parallel processing

62 Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995

Page 3: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

R O G K A M I N V O L V E M E N T

(Eysenck, 1982), the productionof a large number of internalcues, and the focus on thearousal state itself versus thetask (Mandler, 1975), an inter-ruption in processing (Berlyne,1960), and finally, but morespeculatively, a reduction of to-tal attentional capacity (Easter-brook, 1959).

Despite the common concep-tualization of invoivement inten-sity as arousal (Andrews, Durva-suia, and Akhter, 1990; Bello,Pitts, and Etzel, 1983; Burnkrantand Sawyer, 1983; Cohen, 1983;Day, 1973; Mitchell, 1979;Murry, Lastovicka, and Singh,1992; Pavelchak, Antil, andMunch, 1988), only two levels ofprogram involvement have gen-erally been contrasted in pastresearch. This, of course, prohib-its the inverted-U relationshipfound in the arousal literaturefrom emerging. Indeed, one pa-per that measured a continuumof program involvement foundan inverted-U relationship be-tween program involvement andthe recognition of billboardsalong the periphery of a soccerfield (Pham, 1992). However,because billboards compete withthe program for viewers' atten-tion, the effect of program in-volvement was explained interms of viewers' visual focus ofattention. Arousal was conceptu-alized separately and was foundto decrease recognition (Pham,1992).

Research Objectives

The Effect of Program Involve-ment on Memory. While Kroe-ber-Riel observed that there exist"very few examples of the nega-tive effect of very strongarousal" (1979), arousal has gen-erally been found to decreaseperformance in consumer behav-ior studies. For example, in theresearch on program involve-ment the dominant position for-

warded is that loio involvementprograms lead to better brandand message recall {Bryant andComisky, 1978; Kennedy, 1971;Pavelchak, Antil, and Munch,1988; Soldow and Principe,1981). It is argued that high pro-gram involvement leads to moreeffort spent in watching a pro-gram and thereby to lower elab-oration on the embedded com-mercials. The opposite position,that commercials are more effec-tive when viewed in high-'m-volvement programs, has alsobeen forwarded {Singh andChurchill, 1987; Siebert, 1978).This and other research arguethat arousal results in a morevigilant and acute processing ofinformation which in turn leadsto an increase in memory {Silkand Vavra, 1974; Srull, 1983).

We suggest that these seem-ingly contradictory findings canbe accounted for by the in-verted-U relationship found inthe arousal literature. For exam-ple, the Pavelchak, Antil, andMunch (PAM) study (1988) usedself-selected viewers of a Super-bowl. The viewers who weresupporters of one of the partici-pating teams were more in-volved and experienced higherlevels of arousal than neutralviewers who had superior mem-ory for commercials. However,one could argue that even neu-tral viewers were at least moder-ately aroused and that the PAMstudy only captured the de-scending slope of the inverted-Ucurve. Therefore, it is possiblethat the PAM study only cap-tured the inhibiting effects ofarousal as it is increased frommoderate to high levels. Specifi-cally, we speculate that, hadviewers been included in thePAM study who had very lowinvolvement with the Super-bowl, the inverted-U relation-ship would have surfaced be-tween program involvement andmemory for the embedded com-

mercials, as typically found inthe arousal literature. Moreformally,

HI: We expect an inverted-Urelationship between mem-ory for advertised brandsand a continuum of low tohigh program involvement.

The Effect of Program Involve-ment on Attitude toward theAd. In the consumer behaviorliterature on attitude formationthe dominant view is again thatprogram involvement or arousallead to a reduced ability to pro-cess advertisements in depth.For example, a recent studyshowed that weak messages aremore persuasive in high-involve-ment programs, but that strongmessages are more persuasive inlow-involvement programs(Anand and Sternthal, 1992).Similarly, strong argumentswere found to be more persua-sive and weak ones were lesspersuasive under moderate ver-sus high levels of physiologicalarousal (Sanbonmatsu andKardes, 1988). The reasoningwas that high involvement orarousal reduces the ability tothink, thereby inhibiting the for-mation of counterarguments forweak messages versus inhibitingsupport arguments for strongmessages (e.g.. Petty and Ca-cioppo, 1986).

However, Sanbonmatsu andKardes {1988) did not includelow levels of arousal in theirstudy and themselves suggestthat low levels of physiologicalarousal may also have reducedthe ability to think. From thisstandpoint the inverted-U hy-pothesis may again be able toexplain inconsistent findings onthe impact of advertising contexton Aj j. For example, highly in-volving programs have beenshown to lead to higher A^j(Kennedy, 1971), as well as tothe opposite, namely lower A jj

Journal of ADVERTISING RESEARCH—SEPTEM BE FVOCTOBER 1995 63

Page 4: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M E N T

(Murry, Lastovicka, and Singh,1992; Soldow and Principe,1981). As discussed previously,the ability to process informationshould be highest under moder-ate versus both low or high lev-els of program involvement.And it is possible that somestudies manipulated involve-ment on the ascending part ofthe inverted-U curve (i.e., com-paring low to moderate levels ofprogram involvement) while oth-ers manipulated the descendingpart (i.e., comparing moderateto high levels).

Prior research on program in-volvement has had to rely onsung versus spoken versions ofthe same message in order toachieve differences in messagestrength via ease of processing,because "it seemed unlikely thatpractitioners couid be promptedto employ weak advertising mes-sages purposely" (Anand andSternthal, 1992). Because we usereal advertisement stimuli, it isassumed that the commercialsselected for this study do notinclude weak messages. There-fore, we expect the same in-verted-U relationship betweenA^j and program involvement asfor memory. Because moderatelyinvolved viewers should have ahigher abihty to consider themessage in depth compared toboth viewers with low and highinvolvement, their resulting atti-tudes should also be highest(Petty and Cacioppo, 1986).More formally,

H2: We expect an inverted-Urelationship between A^and a continuum of low tohigh program involve-ment.

The Slotting of Commercials.Different parts of television pro-grams are differentially arousing.For example, viewers' arousalshould be higher for televisedsports during the game itself

versus at half-time or followingthe game, that is, when arousalwears off. Therefore, a compari-son of commercials slotted dur-ing the game itself versus out-side the game offers a compari-son of viewers at the same levelof program involvement who,however, have different levels ofarousal at the time a commercialis watched. Because programinvolvement is an overall mea-sure, the arousal experienced forviewers at every level of involve-ment should be higher at the timea commercial is watclied if a com-mercial is shown during thegame compared to when it isshown outside the game. Forexample, viewers with high pro-gram involvement should experi-ence higher than optimal levelsof arousal during the game. Thesubsequent wearoff, say duringhalf-time, should benefit the in-formation processing of theseviewers. On the other hand,viewers with low levels of pro-gram involvement may alreadyexperience below-than-optimalarousal during the game anda further wearoff should bedetrimental to informationprocessing.

This reasoning translates intoan interesting additional test ofthe inverted-U hypothesis. Spe-cifically, when commercials areshown during the game the levelof program involvement thatrepresents an optimal arousallevel will be lower-than-optimalfor commercials shown outsidethe game, because this arousalwears off by then. Instead, thelevel of program involvementwhich is slightly higher-than-optimal during the game shouldrepresent an optimal level ofprogram involvement for com-mercials shown outside thegame because of the wearoff ofarousal. This should cause thehypothesized shift from an in-crease in ad memory and A^^^ toa decrease—the peak of the in-

verted-U curve—to occur at alower level of program involve-ment for a commercial shownduring the game compared to athalf-time or following the game.Specifically,

H3: We expect the shift from anincrease to a decrease in admemory and A,j—the peakof the hypothesized in-verted-U curve—to occur ata lower level of programinvolvement when commer-cials are shown during ver-sus outside of a soccergame.

The slotting of commercials isalso of interest from a secondperspective. It has been arguedthat program-interrupting com-mercials annoy viewers for alltypes of programming and maylead to lower A,,j (Steiner, 1963),especially if viewers desire acontinuation of the program-ming (Schumann and Thorson,1990). These findings may beamplified for televised soccerwhich offers no natural breaks.That is, commercials shown dur-ing the game actually force view-ers to miss part of the action.Anecdotal evidence from the1990 World Cup for which therewere commercial interruptionssuggests that many non-Spanishspeaking viewers watched thegames on the Spanish cablechannel Univision in order to notmiss part of the action. To avoidthis "boycotting" behavior andviewer dissatisfaction, the 1994World Cup eliminated commer-cial interruptions.

Moreover, we expect that thelevel of annoyance with com-mercials should be higher athigher levels of program in-volvement. If these ad-generatedfeelings influence A j (Edell andBurke, 1987; Steiner', 1963), anegative relation between pro-gram involvement and A , isexpected. However, exactly the

64 Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995

Page 5: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M E N T

opposite effect has also foundsupport, namely that arousal—which is also higher at high lev-els of program involvement—positively carries over onto thecommercials causing a positiverelationship between programinvolvement and A j (Krugman,1983). Our experiment allows usto test these competing views.

An Alternative Explanation ofExcitation Transfer. A possiblealternative explanation for aninverted-U finding can be basedon a similar idea of excitationtransfer. In contrast to Krugman(1983), Berlyne (1960) theorizedthat people find both low andhigh levels of arousal unpleasur-able. Because the excitation pro-duced by a program can be mis-attributed to the embedded com-mercial—as long as the viewer isnot aware of the source ofarousal (Mattes and Cantor,1982)—a simple process of affecttransfer could explain the hy-pothesized effect of program in-volvement (Schumann andThorson, 1990) on A,,j, That is,because the arousal-inducedpleasure may be highest at mod-erate levels of program involve-ment, attitudes may also behighest at that point.

In the case of excitation trans-fer, however, the effect of pro-gram involvement on commer-cial messages should be thesame regardless of message con-tent. Specifically, the inverted-Urelationship should surfacewhether a commercial presentsarguments that can be centrallyprocessed or whether it containsonly the brand name and a shortslogan—as for sponsorship mes-sages. Excitation transfer wouldaffect both commercials andsponsorship messages alike. Wepropose that, because there areno arguments in sponsorshipmessages that viewers could dif-ferentially process in depth(Petty and Cacioppo, 1986), pro-gram involvement should not

have an effect on their evalua-tion. A finding of the predictedeffect of program involvementon commercials (H2) contrastedwith a lack of an effect of in-volvement on sponsorship mes-sages would rule out the excita-tion-transfer hypothesis {Schu-mann and Thorson, 1990).

H4: We expect that there is noeffect of program involve-ment on the evaluation ofsponsorship messages.

Research Design

Based on the finding that theobserved variance can be greaterwithin programs than betweenprograms (Yuspeh, 1979), weoperationalize arousal as in-volvement with a single pro-gram. We selected a televisedsoccer game and viewers rang-ing along a continuum of low tohigh program involvement. Theviewers were not supporters ofeither team. By using this opera-tionalization we hoped to avoidthe myriad of potential con-founds inherent in the compari-son of programs as different ascomedies and suspense thrillers(e.g., Kennedy, 1971). Moreover,soccer is the world's most popu-lar sport and is currently experi-encing tremendous growth inthe United States {US Nezvs andWorld Report, 1994). And withthe wide viewership of sportingevents in general, we believethat this type of programmingis particularly relevant toadvertisers.

Experimental Design. Thestudy was conducted in 1992and has two experimental condi-tions. In both conditions subjectssaw a summary of the first-halfhighlights of the 1990 WorldCup game between Germanyand The Netherlands followedby the entire second half and apost-game show, including high-lights and interviews. In the in-

trusive condition five sets ofthree commercials interruptedthe game and made viewersmiss the action for the durationof the commercials, as in the1990 World Cup. There were nocommercials shown outside thegame. In the nonintrusive condi-tion the identical commercialswere shown in the same orderas in the intrusive condition(The Abdominizer, Mars, OldSpice, Players Club, the Army,TNT Movies (sponsorship), Gato-rade (sponsorship), Santa Fe(sponsorship), Micatin (sponsor-ship), Budweiser, NBA, Citizen,Coca-Cola, NFL, and Tums).The game, however, was shownwithout any commercial inter-ruptions—as in the 1994 WorldCup—and one set of commer-cials was shown at half time be-tween the first half summaryand the second half; a secondset was shown between thegame and the post-game show.Program involvement and aildependent measures were col-lected following the post-gameshow in both conditions.

Subjects. Eighty-six universitystudents—forty-three were ran-domly assigned to each condi-tion—were recruited to partici-pate in the approximately 80-minute experiment. Participationwas a class requirement in orderto attract viewers with low levelsof program involvement as well.The subjects included active soc-cer players. Subjects received $5and a chance to win $200 fortheir participation. The $200 wasawarded based on a lottery.

While the choice of a studentsample may restrict the general-izability of the findings, studentsmay be particularly appropriatefor the purposes of this study.As many soccer enthusiasts areaware, soccer's tremendousgrowth in the United States isonly a recent phenomenon. Forexample, the 16.4 million soccerparticipants in 1993 represent a

Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995 65

Page 6: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M K N T

77 percent increase over 1984participation levels {US News andWorld Report, 1994). In addition,a vast majority (85.8 percent in1992) of U.S. soccer players areunder the age of 25 (The 1993National Soccer ParticipationSurvey). These figures indicatethat the current bulk of soccerenthusiasts are college age oryounger. A wider sample of re-spondents would have thereforebiased the low-involvementgroup toward an older agebracket. Coupled with the choiceof real advertisements thiswould have made an interpreta-tion of the results extremely dif-ficult. The relative homogeneityof our student subjects thereforeallows for both a broad range ofprogram involvement and amore rigorous test of the hy-potheses (Calder, PhiUips, andTybout, 1981; Kruglanski, 1975).

Independent Variables. Condi-tion was coded as a dummyvariable in the subsequent re-gression analyses (1 = intrusive;0 = nonintrusive). Program in-volvement was measured usinga scale consisting of eight state-ments each followed by 7-pointLikert-type scales anchored by:1 = strongly disagree and7 = strongly agree. This scalewas developed to capture boththe enduring antecedents of pro-gram involvement (Celsi andOlson, 1988) ("I am a soccerfan," "I will watch many WorldCup '94 games on TV," "I canplay soccer well," and "Howmany World Cup '90 games didyou see on TV?" [1 = none,7 - six or more]) and its conse-quences (Celsi and Olson, 1988),or subjective felt state of pro-gram involvement (Mitchell,1979) ("I think watching thegame was exciting," "I enjoyedwatching the game," "At times Ireally 'got into' the game," and"So far, I regret having partici-pated in this study"). The reli-ability and dimensionality of the

scale were assessed using Cron-bach's alpha and exploratory fac-tor analysis. Confirmatory factoranalysis loadings are used tosum the items as a weighted in-dex of program involvement.

Dependent Variables. Sub-jects' mood ("1 am presently in agood mood") and overall annoy-ance with the commercials ("Thecommercials I saw were annoy-ing"), as well as attitude towardadvertising ("Advertisers' effecton soccer in general is very help-ful") were measured using thesame 7-point scales. Followingthese scales was a free-recall taskand a recognition task includingdistractors (NHL, Becks, Snap-Up, HBO, Pepsi, Brut, M&M's,Rolaid, ESPN, Tinactin, Soloflex,the Hair Club, the Navy, andTimex). The A^^ semantic differ-ential scale (anchored by goodand bad and coded from 7 to 1)was the final task. While single-item attitude scales are expectedto be less reliable, a recent re-view and meta-analysis of theA d literature (Brown and Stay-man, 1992) found no significanteffects of using multi-item scalesversus single-item scales on thevalidity of the relationship. Thesum of the attitude scores acrossall commercials was used as anindex of A^ .

The attitude scores toward thesponsorship brands (Gatorade,Micatin, and TNT Movies) weresummed to create an attitudeindex for sponsorship messages.Due to an oversight on the ques-tionnaire, attitude toward one ofthe sponsorship brands (SantaFe) was not measured.

Results

Involvement. The validity ofthe involvement scale was as-sessed prior to testing the hy-potheses. These tests indicatethat the scale reliably measureda single construct, namely "in-volvement." The involvement

scale including all eight mea-sures was reliable as measuredby Cronbach's alpha (a = .94)with no improvement possibleby dropping any of the items. Apost test (tt = 77) also showedour program-involvement scaleto be highly correlated with Za-ichkowsky's (1985) personal in-volvement inventory (p = .71;p < .0001). Moreover, an explor-atory factor analysis using alleight measures identified onlyone factor with an eigenvaluegreater than one (5.7) on whicheach item loaded positivelyabove .74 (.84 average). A confir-matory factor analysis was sub-sequently performed on theeight program-involvement mea-sures, each of which was foundto load positively and signifi-cantly on the involvement con-struct (all f-values exceeded 7.0).Standardized coefficient loadingsranged from 0.68 to 0.93 withassociated R-squared measuresof 0.46 to 0.86. The standardizedfactor loadings obtained in theconfirmatory factor analysis wereused to weight each of the eightitems. These weighted scoreswere summed to create an indexof overall program involvement.

The program involvement in-dex covered the possible rangefrom a low of 6.5 to a high of45.8 with an overall mean of29.5. The individual means forinvolvement with the game itselfdid not differ statistically be-

Memory performanceincreased for both the recall

and recognition asinvolvement increased fromlow to moderate levels and

then decreased asinvolvement increasedfurther to high levels.

66 Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995

Page 7: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M E N T

tween the two conditions[p > .7). Therefore, a shift in thepeak of the inverted-U curve(H3) can be attributed to awearoff in the level of arousalalong the continuum of programinvolvement rather than a prioridifferences between theconditions.

Overview of Findings. Thememory and attitude measuressupport the hypothesis that theability to process informationvaries along the continuum oflow to high program involve-ment in terms of an inverted-Urelationship (HI). Memory per-formance increased for both therecall and recognition as involve-ment increased from low tomoderate levels and then de-creased as involvement in-creased further to high levels.The finding that the peak of thiscurve occurs at lower levels ofprogram involvement for recall(but not recognition) in the in-trusive condition {H3) supportsthe theoretical basis proposed toexplain this effect, namely anarousal-driven explanation (seeFigure 1).

The same inverted-U relation-ship also surfaced for A ,j alongthe involvement continuum(H2). Viewers had the most pos-itive attitudes under moderatelevels as compared to both lowand high levels of program in-volvement. In comparison, atti-tudes toward the sponsorshipannouncements—which havevery little message content thatcould be differentially pro-cessed—were not affected byprogram involvement (H4).Therefore, both the memory andattitude measures support thenotion that advertising effective-ness is highest under moderateas compared to low and highlevels of program involvement.These findings are discussed inmore detail below.

Recall. Subjects' recall of the15 brands (overall mean — 4.9)

Figure 1

Recall of Intrusive and Nonintrusive Commercials as a Function ofProgram Involvement

Recall

1 1

Nonintrusive

Intrusive

14 22 30

Program Involvement Index

38 46

was influenced by both the levelof program involvement and theviewing condition. The in-verted-U relationship for recallalong the continuum of programinvolvement (HI) was confirmedin a regression analysis (see Ta-ble 1) by a positively signed andsignificant coefficient for involve-ment [p < .0001) and a nega-tively signed and significant co-efficient for involvement squared{p < .0001). The shift of the peakof the inverted-U curve to a

lower level of program involve-ment in the intrusive condition(H3) is suggested by a significantand negatively signed interactioncoefficient between conditionand involvement {p < .05).^ The

' The authors were unable to find a testwhich directly examines the differencebetween the peaks of two quadraticfunctions (this test would require testingthe difference between two Cauchy dis-tributed variables).

Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995 67

Page 8: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M E N T

peak in recall along the involve-ment index was at 26.1 for intru-sive compared to 31.5 for nonin-trusive commercials.

The significant interactionterm also indicates that the in-verted-U curve for the nonintru-sive condition is "flatter."'^ Thiscan be explained by the probablewearoff in arousal in that condi-tion. That is, because arousallevels induced by the game con-verge outside the game, the re-duced range in arousal results insmaller differences in recallamong viewers with differentlevels of involvement. There wasalso an overall recall advantagefor intrusive over nonintrusiveplacements of commercialsip = .005). This effect was notanticipated and may have oc-curred because intrusive com-mercials had an element of sur-prise or simply because theywere shown in smaller sets ofcommercials. The effects are tab-ulated in Table 1 and showngraphically in Figure 1.

Recognition. For the recogni-tion measure (overallmean = 11.2 out of 15) the re-gression (see Table 1) againdemonstrated the inverted-Urelationship (HI) with a signifi-cant and positively signed coeffi-cient for involvement {p ^ .05)and a significant and negativelysigned coefficient for involve-ment squared {p ^ .05). Themain effect for condition{p = .0001) again showed thatintrusive commercials are betterremembered overall. There wereno significant interaction effectsfor the recognition measure, andH3 was not confirmed for recog-nition. These sHghtly weakerresults on the recognition mea-sure may be due to the overallhigh level of recognition which

^ We would like to thank one of the re-viewers for pointing this out.

could have resulted in a ceilingeffect.

Attitude toward the Ad. Be-cause of missing values on theAgd scale five observations hadto be dropped in the regressionanalysis of the A^ index. Theinverted-U relationship (H2)emerged for A ^ (overallmean = 48.9 out of a possiblerange of 11 to 77) along the con-tinuum of program involvementwas demonstrated by a signifi-cant and positively signed coeffi-cient for involvement [p < .01)and a significant and negativelysigned coefficient for involve-ment squared {p = .01). More-over, this relationship did notemerge for the sponsorshipbrands (p's > .77) lending sup-port to H4. The regression re-sults for A d are summarized inTable 2.

Subjects who were more in-volved with the game also re-ported being in a better moodip < .001) in both conditions.However, neither mood nor "at-titude toward advertising" hadany significant effects on recallor Agjj as simple or second-orderterms. This further rules out thecompeting explanation of trans-fer of affect (Schumann andThorson, 1990). There was, how-ever, some support for thepower of feelings on A^j (Edelland Burke, 1987). Annoyancewith the commercials was lin-early influenced by program in-volvement, as well as by condi-tion {R^ = .38). Subjects weremore annoyed in the intrusivecondition {p < .005) and moreannoyed the higher their level ofprogram involvement{p = .0001). This annoyance hada marginal negative impact(p < .1) on A d (see Table 2).

General Discussion

Limitations. As discussed ear-lier, the choice of student sub-

jects seemed particularly appro-priate for the study. However,student subjects can limit thegeneralizability of the findings tothe general population. For ex-ample, students may generallybe more cognitively engaged ormay try harder to "do well."However, the purpose of theresearch was to demonstrate adifference for advertising effec-tiveness at different levels ofprogram involvement, and thereis no reason to believe that thisinverted-U relationship shouldnot hold for other segments ofthe population (e.g., Calder,Phillips, and Tybout, 1981; Krug-lanski, 1975).

Another limitation is that thefindings rely on an audience thatis not self-selected. Audiencesmay not always include the fullrange of viewers from low tohigh levels of program involve-ment, especially for differenttypes of programs. Moreover,advertising effectiveness is af-fected by behaviors such as talk-ing, leaving the room, muting ofcommercials, or "channel surf-ing." Subjects in our study werediscouraged from talking or leav-ing the room. In practice the au-dience for a single program maynot be as "well behaved."

Finally, memory and attitudemeasures were taken after only ashort break which included thepost-game show and involve-ment scales instead of after themore common 24-hour delayused by advertisers. It thereforeremains unknown whether the

. . advertising effectivenessdecreased as programinvolvement increased

beyond moderate to highlevels.

68 Journal ot ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995

Page 9: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M E N T

Table 1Regression Results for Memory Measures

RegressionVariable coefficienl f-value p-value

Recall^Intercept

Involvement (I)

Involvement^

1.4469

0.4363

-0.92 .4

4.10 .0001

-0.0069 -4.01 .0001

Condition' (C) 3.1938 2.91 .005

Interaction (IxC)

Recognition'^Intercept

Involvement

Involvement^

Condition'

0.0749

7.6285

0.2039

-0.0034

1.8784

2.21

5.60

2.02

-1.95

4.55

.03

,0001

.05

,05

.0001

Note: Overall Models: "F = 5.2, p = .0008, R - ,21; " f - 7.2, p = .0002,'Condition is coded as 1 = Intrusive, 0 -• Nonintrusive.

= .21.

differences in memory and atti-tudes persist for longer periodsof time. It is also possible that atime delay can influence mem-ory and attitudes differently. Forexample, the effect of annoyancewith commercial interruptionsmay be dissociated from thememory for a brand. Such dijfer-eutial decay has been proposed toexplain the "sleeper effect"(Pratkanis et a l , 1988) andwould suggest that the reductionin attitudes based on annoyancemay only be short term.

Conclusions. The results ofthe experiment demonstrate aninverted-U relationship betweenviewer involvement with a tele-vision program, and ad memoryand A^,j. An increase in involve-ment from low to moderate lev-els of program involvement re-sulted in higher advertising ef-fectiveness. However,advertising effectiveness de-creased as program involvementincreased beyond moderate tohigh levels. This finding sug-gests that consequences of in-

volvement or involvement inten-sity can be understood in termsof arousal, thus supporting itscommon conceptualization assuch in the literature (Andrews,Durvasula, and Akhter, 1990;Bello, Pitts, and Etzel, 1983;

Burnkrant and Sawyer, 1983;Cohen, 1983; Day, 1973; Mitch-ell, 1979; Murry, Lastovicka, andSingh, 1992; Pavelchak, Antil,and Munch, 1988). This argu-ment is partially supported bythe shift in the peak of the in-verted-U curve between intru-sive and nonintrusive commer-cials for recall. Interestingly,even though the game selectedfor the experiment was high inaction, it v^as not live nor wereany of the viewers avid fans ofeither team. Therefore the detri-mental effect of program in-volvement can be expected for arange of sports.

From a research standpoint,the results indicate the impor-tance of measuring program in-volvement as a continuum ver-sus contrasting just two levels.In this light, the inverted-U rela-tionship can offer an explanationfor why prior research has sup-ported both a positive effect onad recall (Siebert, 1978; Singhand Churchill, 1987) and A^^(Kennedy, 1971), as well as anegative effect on ad recall (Bry-ant and Comisky, 1978; Pavel-chak, Antil, and Munch, 1988;

Table 2Regression Results for

VariableRegressioncoefficient f-value p-value

go Commercials'

Intercept

Involvement

Involvement^

37,36 5,83 ,0001

1.24 2.71 .008

-0.02 -2.60 .01

Annoyance

Aad Sponsorships"Intercept

Involvement

-1.05

12.28

-0.007

-1.95

6.18

-0.05

.06

.0001

.96

Involvemenr 0,0007 0 29 77

Note: Overall Model: "F = 5.1, p = .008, fl^ = .12; "F = 1.1, p = .3,'Condit ion is coded as 1 = Intrusive, 0 = Nonintrusive.

= .03.

Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995 69

Page 10: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M E N T

Soldow and Principe, 1981) andA-ad (Murry, Lastovicka, andSingh, 1992; Soldow and Princ-ipe, 1981) for heightened pro-gram involvement. The findingsalso qualify the practice of "cliffhanging" from a cognitive pro-cessing standpoint. Becausearousal can be both too low andtoo high, moderately excitingportions of television programsshould result in the optimal ad-vertising effectiveness.

More speculatively, we sug-gest that commercials shownduring exciting programs, suchas playoff sports or movie thrill-ers, should be less complex incontent to allow highly involvedviewers to fully process the mes-sage. The lack of an effect ofprogram involvement on A^^ forsimple sponsorship messagessupports this view. This tacticseems particularly appropriatebecause target audiences areusually those that have the high-est program involvement. A re-lated tactic previously suggestedis the use of peripheral cues foraudiences with high versusmoderate arousal (Sanbonmatsuand Kardes, 1988). Future re-search should explore whetherperipheral cues such as expert orcelebrity endorsers are also moreeffective for audiences with lowversus moderate arousal. More-over, different types of televisionprograms can result in differentlevels of program involvement.Those that stimulate moderatelevels of arousal or excitementmay be most appropriate venuesfor cognitively demanding com-mercials. However, arousal gen-erated is only one factor in-volved when comparing pro-grams as different as thrillersand comedies (e.g., Kennedy,1971) and a myriad of other fac-tors need to be considered aswell.

Finally, the slotting of com-mercials also extends beyond the

arousal generated by a program.For soccer in particular, the ef-fects of feelings of annoyancegenerated by the commercials onAgj are of interest. In most partsof the world commercials are notbroadcast during a game becausesoccer has no natural breakssuch as time-outs. With the ex-ception of World Cup 1994, di-rectors of American sports tele-casts, on the other hand, havetended to schedule intrusivecommercials during the gameitself forcing the viewer to misspart of the action. Viewers obvi-ously resent such intrusive com-mercials and this negatively af-fects their attitudes (Schumannand Thorson, 1990; Steiner,1963). The advertisement gener-ated feeling (Edell and Burke,1987) of annoyance is particu-larly important for program in-volvement because it is moreextreme at higher levels ofinvolvement.

In conclusion, the researchfindings suggest that advertisersshould consider the level of in-volvement an audience haswith a television program. It issuggested that both low andhigh levels of viewer involve-ment are suboptimal for bothad memory and A ^ . These find-ings have implications for theslotting of ads within a singleprogram or between differenttypes of programs which maylead to low, moderate, or higharousal. And with the rise incontrol over advertising place-ments, particularly for cabletelevision, arousal appears tobe a potentially important vari-able to consider for both mem-ory and attitudes. Finally, if asfor the World Cup, the targetaudience is likely to be highlyinvolved then the design ofcommercials may need to besimplified or the increased useof peripheral cues should beconsidered. •

NADER T. TAVASSOLI is assistant protes-sor in the marketing group o1 the SloanSchool oi Management at the Massachu-setts Institute of Technology He receivedhis PhD and M Phil, degrees from Colum-bia University and his M B A degree fromSyracuse University His current researchtakes a cross-cultural perspective on con-sumer behavior issues such as the integra-tion cf multimedia information Professor Ta-vassoli has been published in such journalsas the Journal of Consumer Psychology andthe Journal ol Consumer Research.

CLIFFORD J. SHULTZ. II, is a marketingfaculty member at the Arizona State Univer-sity West. School of Management He re-ceived his Ph D in social psychology tromColumbia University and has a primary re-search interest in cross-cultural issues rele-vant lo advertising, marketing, and con-sumer behavior He has published articles inthe Columbia Journal ot World Business.Psychology and Marketing. Business Hori-zons, The Journal ol Spiled Social Psychol-ogy. Research in Consumer Behavior. Con-temporary Southeast Asia, and other jour-nals and conference proceedings.

GAVAN J. FITZSIMONS is assistant profes-sor at the John E Anderson GraduateSchool ol Management at UCLA He re-ceived his Ph.D and M Phil, in marketingfrom Columbia University and his M.B.A. inmarketing and B.Sc in chemistry from theUniversity of Western Ontario His primaryresearch interests are in the areas of con-sumer satisfaction, consumer decision mak-ing, and consumer response to constrainedchoice environments He is a consultant toand executive teacher for firms in a widevariety of businesses ranging from market-ing communications to not-for-profit man-agement.

References

Anand, P., and B. Sternthal."The Effects of Program Involve-ment and Ease of MessageCounterarguing on AdvertisingPersuasiveness." journal of Con-sumer Psychology 1, 3 (1992): 225-38.

Andrews, J. C; S. Durvasula;and S. H. Akhter. "A Frame-work for Conceptualizing andMeasuring the Involvement Con-struct in Advertising Research."Journal of Advertising 19, 4 (1990):27-40.

Bello, D. C ; R. E. Pitts; andM. J. Etzel. "The Communica-tion Effects of Controversial Sex-ual Content in Television Pro-grams and Commercials." Jour-

70 Journal ot ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995

Page 11: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M 1 N \' O L \' L M E N

ual of Advertising 12, 3 (1983);32^2.

Berlyne, D. E. Conflict, Arousaiand Curiositi/. New York: Mc-Graw-Hill, 1960.

Brown, S. P., and D. M. Stay-man. "Antecedents and Conse-quences of Attitude toward theAd: A Meta-analysis." Jourml ofConsumer Research 1% 1 (1992):34-51.

Bryant, J., and P. W. Comisky."The Effect of Positioning a Mes-sage Within Differentially In-volving Portions of a TelevisionSegment on Recall of the Mes-sage." Human Communication Re-search 5 (1978): 63-75.

Burnkrant, R. E., and A. G.Sawyer. "Effects of Involvementand Message Content on Infor-mation Processing Intensity." InVisual Information Processing Re-search ill Advertising, R. J. Harris,ed. Hillsdale, NJ: Lawrence Erl-baum, 1983.

Calder, B. J.; L. W. Phillips; andA. M. Tybout. "Designing Re-search for Application." journalof Consumer Research 8, 2 (1981):197-207.

Celsi, R. L., and J. C. Olson."The Role of Involvement in At-tention and Comprehension Pro-cesses." journal of Consumer Re-search 15, 2 (1988): 210-24.

Chook, P. H. "A ContinuingStudy of Magazine EnvironmentFrequency and Advertising Per-formance." journal of AdvertisingResearch 25, 4 (1985): 23-33.

Cohen, Joel B. "Involvementand You: 1000 Great Ideas." InAdvances in Consumer Research,Vol. 10, R. P. Bagozzi and A. M.Tybout, eds. Ann Arbor, MI:

Association for Consumer Re-search, 1983.

Day, G. S. "Theories of AttitudeStructure and Change." In Con-sumer Behavior: Theoretical Sources,S. Ward and T. S. Robertson,eds. Englewood Cliffs: PrenticeHall, 1973.

Easterbrook, J. A. "The Effectsof Emotion on Cue Utilizationand the Organization of Behav-ior." Psychological Review 66(1959): 183-201.

Edell, J. A., and M. ChapmanBurke. "The Power of Feelings."journal of Consumer Research 14, 3(1987): 421-33.

Eysenck, M. W. Attention andArousal. New York: SpringerVerlag, 1982.

Kahneman, D. Attention and Ef-fort. Englewood Cliffs, NJ; Pren-tice-Hall, 1973.

Kennedy, J. R. "How ProgramEnvironment Affects TV Com-mercials." journal of AdvertisingResearch 11, 1 (1971): 33-38.

Kroeber-Riel, W. "Activation Re-search: Psychobiological Ap-proaches in Consumer Re-search." journal of Consumer Re-search 5, 4 (1979): 240-50.

Kruglanski, A. W. "The HumanSubject in the Psychology Exper-iment: Fact and Artifact." In Ad-vances in Experimental Social Psy-chology, L. Berkowitz, ed.. Vol.8. New York: Academic Press,1975.

Krugman, H. E. "Television Pro-gram interest and CommercialInterruption." journal of Advertis-ing Research 23, 1 (1983): 21-23.

Mandler, G. Mind and Emotion.New York: Wiley, 1975.

Mattes, J., and J. Cantor. "En-hancing Responses to TelevisionAdvertisements via the Transferof Residual Arousal from PriorProgramming." journal of Broad-casting 26 (1982): 553-56.

Mitchell, A. A. "Involvement: APotentially Important Mediatorof Consumer Behavior." In Ad-vances in Consumer Research, Vol.6, W. L. Wilkie, ed. Ann Arbor,MI: Association for ConsumerResearch, 1979.

Murry, J. P.; J. L. Lastovicka;and S. N. Singh. "Feeling andLiking Responses to TelevisionPrograms: An Examination ofTwo Explanations for Media-Context Effects." journal of Con-sumer Research 18, 4 (1992): 441-51.

Pavelchak, M. A.; J. H. Antil;and J. M. Munch. "The SuperBowl: An Investigation into theRelationship Among ProgramContext, Emotional Experience,and Ad Recall." journal of Con-sumer Research 15, 3 (1988): 360-67.

Petty, R. E., and J. T. Cacioppo.Communication and Persuasion:Central and Peripheral Routes toAttitude Change. New York:Springer-Verlag, 1986.

Pham, M. T. "Effects of Involve-ment, Arousal, and Pleasure onthe Recognition of SponsorshipStimuli." In Advances in Con-sumer Research, Vol. 19, J. F.Sherry and B. Sternthal, eds.Provo, UT: Association for Con-sumer Research, 1992.

Pratkanis, A. R.; A. D. Green-wald; M. R. Leippe; and M. H.Baumgardner. "In Search of Re-liable Persuasion Effects: III. TheSleeper Effect Is Dead. LongLive the Sleeper Effect." journal

Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995 71

Page 12: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

P R O G R A M I N V O L V E M E N T

of Personality and Social Psychology53 (1988): 203-18.

Reid, S. M. "U.S. helps WorldCup make global TV history."The Atlanta journal and Constitu-tion, July 19, 1994.

Sanbonmatsu, D. M., and F. R.Kardes. "The Effects of Physio-logical Arousal on InformationProcessing and Persuasion."journal of Consumer Research 15, 3(1988): 379-85.

Schumann, D. W., and E.Thorson. "The Influence ofViewing Context on CommercialEffectiveness: A Selection-Pro-cessing Model." Current Issuesand Research in Advertising 12, 1(1990): 1-24.

Siebert, D. E. "The Effect of Pro-gram Context on CommercialRecall." Paper presented at themeeting of the Association ofNational Advertisers TelevisionWorkshop, March 1978, NewYork City.

Silk, A. J., and T. G. Vavra."The Effect of Program Context

on Commercial Recall." Paperpresented at the meeting of theAssociation of National Advertis-ers Television Workshop, March1978, New York City.

and "The Influ-ence of Advertising's AffectiveQualities on Consumer Re-sponse." In Buyer I Consumer Infor-mation Processing, G. D. Hughesand M. L. Roy, eds. Chapel Hill,NC: University of North Caro-lina Press, 1974.

Singh, S. N., and G. A. Chur-chill. "Arousal and AdvertisingEffectiveness." journal of Adver-tising 16, 1 (1987): 4-11.

The Soccer Industry Council ofAmerica. "The 1993 National Soc-cer Participation Survey." NorthPalm Beach, Florida, 1993.

Soldow, G. F., and V. Principe."Responses to Commercials as aFunction of Program Context."journal of Advertising Research 21,2 (1981): 59-65.

Srull, T. K. "The Impact of Af-fective Reactions in Advertisingon the Representation of ProductInformation in Memory." In Ad-vances in Consumer Research, Vol.10, R. P. Bagozzi and A. M. Ty-bout, eds. Ann Arbor, MI: Asso-ciation for Consumer Research,1983.

Steiner, G. A. The People Look atTelevision. New York: Knopf,1963.

U.S. News & World Report. "ForKicks." May 30, 1994.

Yerkes, R. M., and J. D. Dod-son. "The Relation of Strengthof Stimulus to Rapidity of Habit-Formation." journal of Coin pa ra-tive Neurology and Psychology 18(1908): 482-95.

Yuspeh, S. "The Medium versusthe Message." In A Look Back, aLook Ahead, G. B. Hafer, ed. Chi-cago: American Marketing Asso-ciation, 1979.

Zaichkowsky, J. L. "Measuringthe Involvement Construct."journal of Consumer Research 12, 3(1985): 341-52.

72 Journal of ADVERTISING RESEARCH—SEPTEMBER/OCTOBER 1995

Page 13: NADER I. TAVASSOLI ARE MODERATE LEVELS BEST FOR AD …gavan/bio/GJF...NADER I. TAVASSOLI CLIFFORD J. SHULTZ, II AND GAVANJ. FITZSIMONS PROGRAM INVOLVEMENT: ARE MODERATE LEVELS BEST

Recommended