Predicting Selection and Activity in Television Genre Viewing

Requests for reprints should be sent to Robert P. Hawkins, School of Journalism andMass Communication, University of Wisconsin–Madison, 821 University Avenue,Madison, WI 53706. E-mail: [email protected]

Predicting Selection and Activity inTelevision Genre Viewing

Robert P. HawkinsSchool of Journalism and Mass Communication

Suzanne PingreeJacqueline Hitchon

Life Sciences CommunicationUniversity of Wisconsin–Madison

Bradley W. GorhamDepartment of Communication

Syracuse University

Prathana KannaovakunDepartment of Western Languages

Prince of Songkla University

Eileen GilliganDepartment of Journalism and Technical Communication

Colorado State University

Barry RadlerSchool of Journalism and Mass Communication

University of Wisconsin–Madison

Gudbjorg H. KolbeinsFaculty of Social Sciences

University of Iceland

Toni SchmidtSchool of Journalism and Mass Communication

University of Wisconsin–Madison

MEDIAPSYCHOLOGY, 3, 237–263.Copyright © 2001, Lawrence Erlbaum Associates, Inc.

2 3 8 HAWKINS ETAL.

“Active” television viewing has meant (among other things) selectiveexposure to types of content, attention to that content, and several differentkinds of other activities during viewing itself. This study argues that suchmeanings are differently predicted by three types of predictors (individuals’gratifications sought from different television genres, their expertise withthese genres, and their need for cognition), and also vary by genre. Twodifferent instrumental reasons for viewing (mood and content preference) bothpredicted selective viewing and thinking while viewing, but only contentpreference predicted attentive viewing. Casual reasons for viewing wererelated to less viewing and more channel surfing behavior. Need for cognitionwas unrelated to variation in genre viewing, but it was related in differing butsensible ways to attention to different genres.

These results support the utility of genre in differentiating processes intelevision viewing and further argue for making a number of distinctions inresearch: between genres, between gratifications, between gratifications andother predictors, and between selective viewing and during-viewing activities.

A long-continuing focus of the mass communication enterprise is the search forenduring individual characteristics and beliefs predicting mass media use.Beyond simple nonexplanatory demographics, predispositions studied haveincluded personality traits, cognitive styles, beliefs, and, probably mostimportantly, the whole constellation of beliefs and values comprising mediagratifications (Blumler, 1979; Rubin, 1994). However, “media use” itself carriesa weight of multiple meanings, many implying purposefulness and activity, mostcommonly selective exposure and a wide variety of cognitive and behavioralactivities that may occur during media use (Hawkins & Pingree, 1986; Levy &Windahl, 1984).

Given the diversity of meaning in each case, the goal here is to examine theintersection of the two sets as applied to television viewing. Furthermore, we willargue that the medium as a whole is too broad a subject for some kinds oftheoretical statements, for which the construct of genre allows clearer and moreprecise hypotheses. Stating the problem most simply, do predispositionsdifferentiate who selects what television content, or do they instead determinewhat happens during viewing, or do the same predispositions predict bothequally? Furthermore, do some individual characteristics predict selection,whereas others determine activity during exposure? These questions require usfirst to distinguish both among the actions of different predispositions and amongaspects of media use.

Far more research has addressed predictions of media exposure than activity,of course. For example, much research simply predicts that people with a certaincharacteristic or belief will watch more television or be more likely to readnewspapers than those without it. Such relationships are typically weak (Wenner,

1986), partly because so many interpersonal and situational factors alsocontribute to exposure (Webster & Newton, 1988). But one should alsoremember that the reason for making such a prediction in the first place is thenotion of selective exposure, rather than simple exposure per se. To the extentthat selectively exposing oneself to a medium or content can be distinguishedfrom simple exposure, predicting selection clearly is much more to the point.

Much of the research on predispositions and media use (both as selection andas activity) has come from the uses and gratifications tradition, which has alwaysacknowledged that significant differences exist within the viewing audience.According to this perspective, individuals actively seek out mass mediaexperiences that can meet their cognitive, social, or emotional needs. Thus,audience members are relatively aware of their needs, reasonably active in howthey select mass communication, and presumably confident that the media theyselect will in fact meet their particular needs. If particular media or messages areperceived as meeting the need, this perception should lead to patterns of mediause, as certain needs should lead to using some types of media and not others (cf.Palmgreen, Wenner, & Rosengren, 1985), as well as to differences in activities(Blumler, 1979; Lin, 1993). Research in this area has thus looked at the varioustypes of needs and motivations that seem to drive media use (e.g., instrumentalvs. consummatory as overarching categories, with multiple specific gratificationswithin each), as well as the patterns of use associated with these motivations (cf.Rubin, 1994).

An issue facing this work, however, is just what unit is appropriate. Somestudies suggest the importance of medium-level gratifications. After collectingdata on television-viewing motivations and television program viewing, Rubin(1983) concluded that individuals who watch television to seek entertainment,amusement, or simply to pass the time (i.e., consummatory reasons) have greataffinity for the medium of television without being specifically interested in anyparticular types of content. Those who reported that they watched television forinformational reasons, on the other hand, showed an affinity for watching news,game shows, and talk-interview shows.

Thus, despite showing that asking reasons for “television viewing” or“newspaper reading” are important (Robinson & Levy, 1986; Rubin, 1983), suchstudies also raise the question of whether further distinctions might be necessary.Some have suggested (e.g., Katz, 1996) that any overall characterization of eithera medium or its users almost inevitably stereotypes a medium’s content or itsusers’motives, or both. That is, casual conversation may refer to newspaper usebeing driven by a need for information or characterize television viewing asinactive, but some people use the newspaper, or at least some sections of it, for

TELEVISION GENRE V I E W I N G 2 3 9

2 4 0 HAWKINS ETAL.

entertainment. Likewise, some television content is complex and stimulating,and even formulaic entertainment television can be watched thoughtfully. Thus,making judgments about the motivations to use specific content based onmedium-level data is problematic at best.

To continue pursuing such relationships between motives and content, whatis needed is a more specific level at which viewers’ attitudes and expectationscould guide use and activity. Obviously, programs are such a more specific level,but programs present serious problems: There are so many, and individualinstances constantly come and go. As an alternative, Preston and Clair (1994)laid out several compelling reasons why genre should interest researchers whoexamine selective viewing and during-viewing activities. First of all, as Woberand Gunter (1986) pointed out, preferences for specific shows and particularepisodes may vary, but preferences for genre remain relatively stable. Secondly,genre may be meaningful and a useful marker for viewers, given cleardifferences among viewers in their perceptions of the instrumental utility ofvarious genres (Atkin, 1985). Finally, genre may influence the way mediacontent is interpreted. Preston and Clair pointed to reception studies by Jensen(1987) and Morley (1980) suggesting that audiences use genre conventions to“decode” and interpret specific television messages, although the exact strategiesemployed by viewers vary. Thus, genre is a meaningful concept not just forresearchers interested in classifying content, but also for audiences attempting tointerpret that content.

Examination of the role that genre may play in uses and gratificationsresearch is not new. Webster and Wakshlag (1983) suggested that “program typepreference” (i.e., genre) probably plays a role in influencing preferences forspecific shows. After examining the literature, they conclude that “conventional,‘common sense’ program types (such as drama, situation comedy, and so on)bear some systematic relationship to program preference” (p. 436). In theirexamination of motives for using television news, Rayburn and Palmgreen(1984) noted a hierarchical structure of gratifications sought and obtainedbetween television news (a genre) and the respondent’s most-watched televisionnews show (a specific show). They concluded that beliefs at one level ofabstraction probably influence beliefs at the other level of abstraction. That is,beliefs about a genre probably influence specific program choice, and theconsumption of a specific show may also influence the perceptions of thatgenre’s gratification-related attributes.

Several uses and gratifications studies have focused on motivations forviewing specific genres, focusing on those of interest because of viewer loyaltyor civic importance. Instead of asking respondents about their motivations for


“viewing television,” Rubin and Perse investigated motivations for viewing soapoperas (1987a; see also Rubin, 1985). Rubin and Perse also investigated uses andgratifications associated with television news (1987b; see also Perse, 1990), andRubin (1981) has even investigated the motivations for watching a particularshow, 60 Minutes. These studies have contributed greatly to our understanding ofthe varied motivations for watching these genres, but they don’t tell us muchabout the relationships between viewer predispositions and selective patterns ofviewing.

One recent study (Kim & Rubin, 1997) tested college students’gratificationsfor soap opera viewing as predictors of selective exposure and several potentialmeasures of audience activity. Three instrumental gratifications (ritualgratifications were not tested) were unrelated to selective exposure, measured asthe proportion of all viewing devoted to soap operas. However, thesegratifications were related in varying ways to a larger set of “activities.” One canquarrel with some of Kim and Rubin’s operationalizations of these activities(some seem more properly perceptions of self or media or genre), but enough areclearly activities to establish a link between gratifications and during-viewingactivities.

A few studies have reversed the process by looking at specific motivationsand seeking the types of content that seem to relate to those motivations. Potts,Dedmon, and Halford (1996), for example, investigated the relationship betweensensation seeking and viewing preference and found that high sensation seekerswatched more music videos, documentaries, and talk shows while watchingfewer newscasts and drama programs than low sensation seekers. Weaver (1991)found that viewers high in neuroticism expressed greater interest in dramatelevision than those low in neuroticism, whereas low neuroticism viewersfavored situation comedies.

Beyond the question of genre- or medium-specific measures, however,methodological issues also hamper generalizations about predispositions andmedia use from currently available studies. For example, a number of the usesand gratifications studies that seek to link viewing motivations to media use relyon measures referring to people’s favorite programs. Rubin (1983) askedrespondents to list up to three programs they would attempt to watch wheneverthose programs were aired. The genre-specific studies discussed earlier (e.g.,Rubin & Perse, 1987b) asked respondents about their viewing of a favorite showof the genre yesterday and on a typical weekday. And a recent study by Perse(1998) that investigated genre differences in channel changing behavior (hermeasure of selectivity) asked respondents to report on any program they wished,and almost all reported on a favorite program.

2 4 2 HAWKINS ETAL.

These methods help link motivations to use of favorite programs, but onemust wonder how generalizable they are to overall viewing. That is, Rubin’s(1983) results tell us that viewing to seek entertainment is unrelated to the typeof favorite program, but it is not at all clear whether this applies to viewingnonfavorite programs. Perse’s (1998) research tells us that adults are unlikely tochange channels during a favorite program, and that the genre of that programmakes little or no difference in that during-viewing behavior, but it does notinform us about the rest (the majority) of their viewing. In fact, if genre matters,it should be much more important for nonfavorite programs, where genre-relatedexpectations (rather than program-specific ones) should be considerably moreimportant in guiding selection. That is, whereas one presumably knows a favoriteprogram in all its idiosyncratic detail, one’s selection and viewing ofnonfavorites cannot draw on such detailed program-specific knowledge and ismore likely to utilize knowledge and expectations of the genre as a whole. Thus,researchers interested in examining genre’s link to media use should attempt tomeasure respondents’ overall television viewing rather than just that of favoriteprograms (e.g., see Preston & Clair, 1994).

Given all of the above, the basic questions of this study concern how wellindividuals’ predispositions and perceptions predict both their selective viewingof genres and their activities during that viewing—both what and how they view,taken at the genre rather than the medium level. To pursue further differentiation,the hypotheses and research questions that follow cover an intersection of threekinds of individual characteristics (genre-specific gratifications, genre expertise,and need for cognition) with several aspects of television use (genre selectiveviewing, genre attention, thinking while viewing, and surfing), as illustrated byFigure 1.

Television “Use” MeasuresGenre Attention Thinking

selective to genre aboutIndividual predispositions viewing television Television SurfingGenre-specific gratifications H1 H3, H4 H6 RQ3Need for cognition RQ1 RQ2 H7 RQ3Genre expertise H2 H5

Figure 1. Intersection of individual predispositions and television use.

PREDICTING SELECTIVE GENRE VIEWING

In predicting selective viewing of genres (a higher proportion of one’s viewing,not just a larger amount), the frequent distinction between ritual (orconsummatory) and instrumental gratifications may be less crucial, becauseprevious research suggests that ritual gratifications such as entertainment orpassing the time were associated with use of the medium rather thangratifications of particular content, either genres or programs (Rubin, 1983). Forthat reason (and as in Kim & Rubin, 1997), these gratifications were notmeasured in this study. However, reasons for viewing each of three genres(situation comedy, dramas, and news) were collected as part of a larger study ofvisual attention to a sample of television containing these three genres. Althoughthese genres are stereotypically quite different in their content, measuringgratifications at the genre level allows one to make similar predictions for allthree. Thus, watching in order to catch particular content or programs (aninstrumental reason) should be positively related to selectively viewing thatgenre (but not to selective viewing of other genres, unless they are substantiallysimilar in the gratifications they offer). Similarly, watching a genre for any otherinstrumental reason, such as emotion management (Bryant & Zillmann, 1994),should also be associated with selectively viewing that genre. In contrast,typically watching a genre for reasons that are essentially accidental (“someoneelse turned the television on,” “the television just caught my attention,” “I hadnothing better to do”) should be uncorrelated with viewing that genre. To bemore explicit, although these accidental reasons are often lumped with ritualgratifications, we think they warrant a different prediction.

Hypothesis 1: Of the various gratifications reported for viewing a genre, onlyhigher levels of instrumental gratifications (here, content and emotionmanagement) for a genre should predict selective exposure to that genre.

One other genre-related characteristic of individuals, self-perceived expertiseabout television genres, should also predict selective genre viewing. By expertisewe do not mean experience with a genre, which of course would merely be ahistorical measure of viewing and would correlate with any current measurementof viewing. Indeed, because most American adolescents and adults are wellversed in all common television genres (see Gerbner & Gross, 1976, for asummary of this argument), actual expertise probably does not vary much, anddegree of experience is not actually the point. Instead, the point of expertise isperceiving oneself to know the techniques and conventions of a particular genre


better than those of other genres, which very likely is closely associated withpreference for and appreciation of a particular genre. Although this is not theprogram-specific fandom studied by Kim and Rubin (1997) or Perse (1998), itdoes provide an analogue at the genre level. Thus, those who considerthemselves experts on a particular genre should selectively watch it as a greaterproportion of their total television diet.

Hypothesis 2: Perceived expertise with a genre should be correlated withselective exposure to that genre.

We measured one additional individual characteristic that may predictselective genre viewing. Need for cognition, or the general preference forthoughtful, analytical cognitive activity (Cacioppo & Petty, 1982), is aninformation-processing and problem-solving style, as well as an orientationtoward information-processing that should guide selection. At the medium level,need for cognition would seem unlikely to be positively associated withtelevision viewing, at least as viewing is often conceptualized as essentiallypassive overall (Kubey & Csikszentmihalyi, 1990; however, see also Katz,1996). For specific genres, one could make an argument that need for cognitionleads to selective viewing, but that argument rests first on a stereotyping ofcontent. For example, the range of subjects covered and their inherent degree ofunpredictability (what events have happened) in news and news magazines maypresent more opportunities for cognition, so that those with high need forcognition select these genres.

One might conversely expect situation comedies to provide little opportunityfor cognitive activity, and thus their viewing should be unrelated (or negativelyrelated) to need for cognition. Predictions for dramas would be less clear.Research has shown that viewers of soap operas often report cognition-basedenjoyment stemming from applying characters’ long histories to currenthappenings and making predictions about plot outcomes (Cantor & Pingree,1983), and a similar relation could obtain if those who enjoy thinking selectdramas in order to predict resolution of their plots. One must be at least a littleskeptical of the extension from soap opera research, however, because actionadventure dramas are often so formulaic that they may in fact provide littlecognitive reward. Furthermore, the analogy between soap opera and drama maybe entirely inappropriate, because soap opera viewers are much more likely to befans than are viewers of prime time dramas (Canton & Pingree, 1983). Beyondthe problems associated with stereotyping genres, predicting relationshipsbetween need for cognition and genre selective viewing further ignores viewer

2 4 4 HAWKINS ETAL.

flexibility and idiosyncrasy, so that news can be background noise to some andsituation comedy a source of careful study for others (cf. Livingstone, 1992).Given these uncertainties, relationships between need for cognition and genreselective viewing must be regarded as research questions instead of hypotheses.

RQ1: Does need for cognition predict selective exposure to news magazines,situation comedies, or drama?

PREDICTING VIEWING ACTIVITY

Beyond the relation of predispositions and genre selection, similar issues recurin predicting three different kinds of measures of activity during use. Self-reported attentiveness during viewing each of the three genres speaks to amountof mental effort or engagement with the content (Hawkins & Pingree, 1986).Self-reports of particular kinds of thoughts and reactions speak instead to themental content of the viewing experience, and channel changing or watchingmultiple programs indicates behavioral activity, although they could have arelation (of ambiguous direction) with amount of attentiveness as well.

Considering gratifications and activity, Rubin (1994), for example, noted thatuses and gratifications research has repeatedly shown that instrumental viewingrepresents a more active approach to television viewing, whereas ritualizedviewing is associated with passive, low-involvement processing. We suggest thatthis characterization of instrumental gratifications may be too simple forpredicting during-viewing activity, even though it served to predict selectiveviewing. Although content-specific gratifications for watching a genre shouldlead to greater attention during viewing of that genre, satisfying emotional needsthrough viewing does not require and may not benefit from careful attention tothe content. And rather than being irrelevant (as they were for selectiveexposure), watching a genre for accidental reasons should be associated withlessened attention to it.

Hypothesis 3: Higher levels of content-specific gratifications from a genreshould predict greater attention to that genre, but emotion managementshould not.

Hypothesis 4: Higher levels of accidental gratifications for a genre should beassociated with lower attention to that genre.


Expertise with a genre should also be associated with greater attention to it,given previous research on expertise and cognitive capacity. As with selectiveviewing, we pose a research question rather than a hypothesis for need forcognition, even though content stereotypes again suggest that it would bepositively correlated with attention to news magazines, but not to situationcomedies, with the prediction uncertain for dramas.

Hypothesis 5: Greater genre expertise will be correlated with greater attentionto that genre.

RQ2: Will greater need for cognition be more associated with attention tonews magazines than to situation comedies?

Our other measures of how individuals watched television were not specificto genres, thus making genre-specific predictions impossible, but one can stillpropose some relationships with individual characteristics. Instrumentalgratifications should also be positively related to thinking while viewingtelevision overall, but much less strongly than with genre-specific measures ofselection. Thinking about television while viewing should be more likely forthose expressing a need for cognition.

Hypothesis 6: Instrumental reasons for viewing all three genres should bepositively correlated with thinking while viewing.

Hypothesis 7: Need for cognition will be positively correlated with thinkingwhile viewing.

One other measure of during-viewing activity produces less clear predictions,because research has not yet been able to clarify its meaning. Surfing channelsor watching multiple programs at once could reflect a greater need for cognition,seeking a higher level of stimulation, or watching for ritualized or casual reasonscould be reflected in frequent channel changes (which occurred even in Perse’s,1998, constrained sample). Thus, relationships here will be examined as researchquestions.

RQ3: Is television channel “surfing” related to need for cognition oraccidental reasons for viewing?

2 4 6 HAWKINS ETAL.

Finally, it is possible that these influences of predispositions on during-viewing activities (Hypotheses 3–7) are not direct, but instead act throughselective viewing. As an example, correlations of gratifications (or expertise orneed for cognition) with attention (or thinking or surfing) could stem from ane ffect of selective viewing on activities, thus making the influence ofpredispositions indirect. That is, it could be that the act of viewing selectivelyitself produces attention or thinking. Although we think a direct effect process ismore plausible, we will test this alternative by controlling for selective exposure.

Hypothesis 8: Predisposition effects on during-viewing activities will bedirect rather than mediated by selective exposure.

METHOD

The data reported here were collected as part of a larger study on attention totelevision and involved three different components collected at different pointsin time. Participants were undergraduates enrolled in two semesters of anintroductory media effects class at a large midwestern university whoparticipated for class extra credit. A total of 174 students (116 women and 58men) participated in all three phases of data collection.

Our student sample presents both advantages and disadvantages. Claims thata student sample represents the general population in any individualcharacteristic (e.g., hours of television viewing, degree of instrumental viewing)deserve to be greeted with skepticism. A better case can be made when theresearch focuses on relationships between characteristics, even if thecharacteristics are more or less prevalent than in the general population (seeCourtright, 1996; Sparks, 1995; and other articles in these two journal issues fora fuller discussion of the use of student samples in communication research).

As lighter and more selective viewers of television than the population oreven their age group as a whole, college student viewing habits probably containproportionately less casual, time-filling viewing (Pingree et al., in press), andthus less error variance from the point of view of tests for selective, activeprocesses. This makes them a good sample within which to understand selectionand activity processes, and results would also be comparable to much of theprevious research on genre and gratifications, which also used student samples.However, it must be acknowledged that the balance between individual andenvironmental factors in selective viewing may be different in the studentenvironment than in other life situations. Sharing a room or an apartment withpeers is different than being an adolescent at home, a young adult living alone,


or a spouse/parent in a family situation. A college student’s selective behavior ismore constrained than in some of these other life situations, and less constrainedthan in others. Still, the effects of these environmental differences on selectivebehavior are probably much less than they used to be, given much better accessto television, both communally and individually, in student living situations. Inany event, the research questions here are fundamentally about relationshipsbetween variables, and not the absolute levels.

In addition, because television viewing was here measured by a week-long diaryseparated by time, location, and function from the survey measures, relationshipsbetween them are unlikely to be artifacts that might occur within a single surveyinstrument, where viewing is often assessed by a small set of recall-based questions.

Early in the semester, all students in the class kept a diary of their televisionviewing for 1 week. These data were used as background for extra credit studiesconducted using class volunteers, but the diary was also designed to providestudents with information about their individual and collective viewing habits.Students were given an instruction sheet and a computerized form on which torecord their data, and they were instructed to record the genre of any televisionthey viewed for each half hour between 6:00 a.m. and 2:00 a.m. for the 7 days ofthe study period. They were also instructed to leave blanks for the times whenthey did not watch television.

Students were asked to classify the television content they watched into one of10 genres, using a classification scheme similar to content categorization schemesused in previous research (e.g., Potts et al., 1996; Preston & Clair, 1994; Rao, 1975;Ta n g n e y, 1988). Previous research from the professional literature on these“common sense” program types (Webster & Wakshlag, 1983) indicates that genreis a reasonably valid and reliable construct to use for categorizing television content(e.g., Frank, Becknell, & Clokey, 1971; Gensch & Ranganathan, 1974; Rao, 1975).To aid reliability, students were also given examples of particular genres, usingshows the students were likely to be familiar with. The 10 genres (and theirexamples) were News, Sports, Soap Opera (e.g., Days of Our Lives, All MyC h i l d re n), Situation comedy (e.g., Friends, Seinfeld, The Simpsons), Drama (e.g.,E.R., Party of Five, Star Trek Vo y a g e r, Walker Texas Ranger), News &Entertainment Magazines (e.g., Dateline NBC, 60 Minutes, Entertainment To n i g h t) ,Game & Talk Shows (e.g., J e o p a rd y, Late Night With David Letterman, The RosieO’Donnell Show), Music Videos, Reality T V (e.g., Cops, A m e r i c a ’s Funniest HomeVi d e o s), and Other Television (e.g., QVC, Discovery Channel documentaries,A & E ’s B i o g r a p h y, how-to shows, infomercials). Televised (and videotaped) movieswere not included in this assignment, and so the overall television viewing figuresmay underestimate student’s actual time spent in front of the television, although itis important to note that this should not influence the genre-specific viewing figures.

2 4 8 HAWKINS ETAL.

Additional background data were collected through a brief in-class survey,announced as pretest data for extra-credit studies to be done later in the semester.The survey mainly tapped attitudes toward products prominently advertised ontelevision, but it also asked about activities engaged in while viewing television.Using a 5-point scale ranging from 1 (not at all) to 5 (a lot), respondents ratedhow much they did certain things while watching television. A set of items abouttelevision-viewing behaviors asked respondents how often they changedchannels, changed from watching broadcasts to a videotape, muted sound, fast-forwarded or rewound a videotape, watched more than one program at a time,and “surfed” channels looking for something to watch. Principal componentfactor analysis using varimax rotation revealed two factors (the “mute sound”item did not load very strongly on either factor). One factor, which we callsurfing, is made up of changing channels, watching more than one program at atime, and surfing around TV looking for something to watch. The second factor,which we call taping, included the two videotape items but was not used in thepresent analysis. A factor score was constructed for the surfing dimension.

The mental activity items asked respondents how often they “talk back” orargue with something on television, think hard about something on television,see connections to their own life, think about what they would do in thatsituation, are uncertain about what something meant or what was going on, try topredict what would happen, and imagine how a character was feeling. Factoranalysis revealed that all but the uncertainty item loaded onto a single factor, sothe uncertainty item was dropped from further analysis, and the remaining itemswere combined to form an index of “thinking” behaviors (α = .73).

The final portion of data collection for this study came from questionnairescompleted as part of a “leisure time study” students could participate in for extracredit. Students came to a university office building in pairs at an appointed timeand filled out a background “Media Use and Opinions” survey before beingobserved in a media use situation. The questionnaires included many questionsabout media-use habits and patterns, but several sets of items are pertinent forthis study. Self-reported attention to television was measured by items that askedparticipants to estimate the amount of attention they pay when they watch showsof particular genres. Participants used a 5-point scale of less than 10%,11%–33%, 34%–66%, 67%–89%, or more than 90% to estimate the amount ofattention they usually pay when they watch situation comedies, drama, news &entertainment magazines, news, soap operas, and music videos (the latter threenot used in these analyses).

Need for cognition was assessed using a five-item index adapted from the 18-item version of this scale (Cacioppo & Petty, 1982). Because the 18-item scale is


used to produce a single dimension, we saw no reason to impose this degree ofparticipant burden, when our own pilot work showed that the five items loadinghighest in previous research also formed a reliable scale. These were the items:“I like to have the responsibility for handling a situation that requires a lot ofthinking”; “I would rather do something that requires little thought thansomething that is sure to challenge my thinking abilities”; “Thinking is not myidea of fun”; “I prefer complex to simple problems”; and “I try to anticipate andavoid situations where it is likely that I will have to think in depth aboutsomething.” These items used a 5-point scale from 1 (strongly disagree) to 5(strongly agree). With negative items reversed, factor analysis confirmed onlyone factor, with all five items loading at over .70, so a summative index wasformed (α = .79).

We constructed a survey item to assess respondents’ perceived genreexpertise, or the perceptions of their own level of sophistication with a particulargenre. The item read as follows:

Everyone knows how to watch television. And most of the time, we understandeverything we see. But for each of us, there are one or two kinds of programs thatwe really understand: appreciating the subtleties, predicting what will happennext, catching things most people would miss, etc.

Participants were then asked to choose one of the following genres with whichthey felt they were most expert: situation comedies, drama, news &entertainment magazines, soap operas, or none of these. Dummy variables werecreated for sitcom expertise, drama expertise, and no expertise, and the soapopera and news & entertainment responses were combined (because of relativelyinfrequent response to each) to form a variable for “other expertise.”

To allow us to test predictions about genre-specific selection and activity,three sets of items assessed the motivations for viewing each of three genres oftelevision: dramas, situation comedies, and news and entertainment magazines.We chose not to simply use past standard gratification items, most of which weredesigned to apply to media in general, even though these medium-level itemshave then been applied to individual genres in other research regardless ofa p p l i c a b i l i t y. Instead, these items were based on focus group reports ofadolescents describing recent experiences with media in detail. However, toallow comparability across genres, we used only items that could be answeredfor each. For each genre, respondents were asked to rate how true each of thestatements is of them when they watch that genre. The complete wording forthese items is given in Table 1.

2 5 0 HAWKINS ETAL.


2 5 2 HAWKINS ETAL.

Table 1 summarizes the factor analyses of the uses and gratifications items runseparately for these genres. The analyses produced similar (but not identical)factor structures, so that although we have named these factors identically, moredetailed examination is warranted. For all three genres, the first factor (labeledmood) reflected mood-based (either to change or maintain mood) and escapeselection, with excitement also important for drama and news magazines. For thelatter, this factor also seems to reflect a need for additional mental stimulation,including “needing to have more to occupy one’s mind.” The second factor(labeled content pre f e re n c e) seemed to reflect long-term content-basedpreferences, because it contains wanting to find out what is happening (perhapsin particular shows), and seeking a particular program, as well as reporting thatwatching this genre was “just a habit” (note that a habit of watching a genre hasa different meaning than a habit of watching television indiscriminately). Fordramas, the item “I needed to relax” also loaded onto this factor. The third factor(labeled caught) reflected casual viewing in which the television caught one’sattention. It was made up of items “The TVwas already on and something caughtmy attention” and “I didn’t have anything better to do,” with “I needed to havemore than what I was doing to occupy my mind” also included for drama andsituation comedy, but not news magazines. Factor scores were constructed foreach of these dimensions.

RESULTS

Overall, these university students reported watching about 13 hours of televisionduring the diary week, with situation comedies and drama most watched. Sports(for men), soap operas (for women), and game shows were the next most heavilywatched categories. Even within this relatively light overall viewing, televisionnews viewing was rather low, averaging only 3/4 hours per week. News andentertainment magazines, music videos, and reality programs were watched lessthan an hour a week each. The residual “other” category (e.g., documentaries,shopping channels) accounted for only about 1/2 hour a week, indicating that thegenre categories named represent the bulk of what these respondents viewed. Ofthe 10 genres, men watch significantly more (p < .05) sports, news, and gameshows, whereas women watched significantly more drama and soap operas.

Table 2 shows correlations between the predisposition measures andselectively viewing the various genres. Because heavier viewers are likely towatch more of any genre, these correlations examine selectively viewing a genreby controlling for all other viewing (i.e., correlations for each genre are partialedfor total viewing time minus time viewing that genre. Thus, correlations reflect

viewing more or less of that genre than one’s overall level of viewing wouldpredict, conceptually similar to a measure of proportion of viewing; Hawkins &Pingree, 1981). To allow a more efficient table, the gratification and expertisecolumns show the correlations for the variable matching the row of the table (i.e.,we do not show the correlations between drama gratifications and situationcomedy viewing).

TABLE 2Partial Correlations Between Predispositions and Genre Viewing, Controllingfor Overall Television Viewinga

Predispositions

Media Use Gratification

Selective genre Content Expert in Need forviewing Mood Preference Caught Genre Cognition

Drama .19* .25** –.19* .14 –.16Sitcom .19* .06 –.11 .26*** .00News Mags .15 .18* –.13 — –.09

*p < .05, **p < .01, ***p < .001.aN = 153–156.

Hypothesis 1 predicted that viewing any of these three genres formood/stimulation or content-preference reasons would be positively correlatedwith viewing that particular genre. Both predicted positive correlations weresignificant for drama, as well as one (content preference) for news magazinesand one (mood) for situation comedy. Also as predicted, viewing for incidental(“caught”) reasons was not positively related to viewing each genre. However,the significant negative correlation for drama viewing may be meaningful,because it is mirrored by similarly signed nonsignificant correlations forsituation comedy and news and entertainment magazine viewing. It may well bethat if one usually watches a genre for such incidental reasons, that reflects a lackof interest in the genre that produces some avoidance of it overall.

Too few respondents regarded themselves as experts in news magazines, soHypothesis 2 was tested only for dramas and situation comedies. Those whoregarded themselves as experts in situation comedies watched significantly moresituation comedy than would be expected based on their total amount of viewing,but the corresponding correlation, although positive, was not significant for drama.


Although Research Question 1 was motivated by common assumptions aboutthe potential presence of stimulating content in some genres and its lack inothers, Table 2 shows both little correlation between need for cognition andselection, and little differentiation between genres. Need for cognition did notpredict greater selective viewing for either news magazines or drama, and ifanything was somewhat negatively correlated. And in that context, the lack ofcorrelation between need for cognition and situation comedy cannot be taken assaying much about any lack of stimulation in situation comedy.

The top half of Table 3 presents relationships between the predispositions andattention to the three genres. In contrast to selective viewing, which bothinstrumental gratifications were expected to predict, Hypothesis 3 predicted thatcontent preference but not mood gratifications would predict attention. This wasconfirmed, as all three content preference gratifications were positively related toattention to that genre (and more strongly than to selective viewing in Table 2),whereas none of the mood correlations were significant. And Hypothesis 4 wassupported in two of three cases, because watching a genre because one’s attentionwas caught by it was negatively related to attention to drama and situationcomedies, although it was unrelated to attention to news magazines. Hypothesis 5was partially supported, because expertise with dramas was correlated withattention to drama, but the correlation was not significant for situation comedies.There was some indication (Research Question 2) that need for cognition wasd i fferentially correlated with genre attention. Need for cognition was significantlynegatively related to attention to dramas and situation comedies, while positivelybut not significantly related to attention to news magazines.

The top half of Table 4 displays the correlations between predispositions andthinking or surfing activities, assessed during television viewing generally.Hypothesis 6 predicted that the two kinds of instrumental gratifications,regardless of genre, would be associated with greater thinking about television,and this was confirmed in four of six tests. Interestingly, however, thinking waspredicted by mood management and content preference for dramas and situationcomedies, but there was no correlation between seeking these gratifications fromnews magazines and thinking. For surfing channels or watching multipleprograms simultaneously (Research Question 3), it was unclear whether topredict positive or negative correlations from the instrumental gratifications, andin fact only one of the six was significant. However, watching dramas orsituation comedies because they caught one’s attention was significantlycorrelated with surfing. Perhaps surprisingly, need for cognition was unrelated tothinking during viewing.

2 5 4 HAWKINS ETAL.

TABLE 3Zero-Order Correlations Between Predispositions and Self-reported AttentionPaid when Watching Genrea

Genre-Specific Predispositions

Attention Content Expert in Need forto: Mood Preference Caught Genre Cognition

Drama .07 .44*** –.26*** .32*** –.21**Sitcom .04 .40*** –.18* .14 –.15*News Mags –.05 .24** .01 — .11

Partial Correlations Between Predispositions and Self-reported Attention Paidwhen Watching Genre (Controlling for Selective Viewing of Relevant Genre)

Genre-Specific Predispositions

Attention Content Expert in Need forto: Mood Preference Caught Genre Cognition

Drama .02 .41*** –.18* .29*** –.18*Sitcom .04 .40*** –.19* .13 –.13News Mags –.03 .21** .02 — .12

*p < .05, **p < .01, ***p <. 001. aN = 153–156.


2 5 6 HAWKINS ETAL.

TABLE 4Z e ro - O rder Correlations Between Predispositions and Behaviors During Vi e w i n g

Predisposition Thinking Surfing

MoodDrama .19* –.01Sitcom .18* –.04News Mags .06 .16*

Content PreferenceDrama .17* .06Sitcom .23** .13News Mags –.09 .12

CaughtDrama –.13 .22**Sitcom –.18* .23**News Mags .15 .06

ExpertiseDrama .10 .00Sitcom –.02 .06

Need for Cognition –.02 –.06

(Controlling for selective viewing of relevant genre)

MoodDrama .14 –.01Sitcom .18* –.02News Mags .00 .21*

Content PreferenceDrama .22** .10Sitcom .32*** .15News Mags .07 .16

CaughtDrama –.15 .31**Sitcom –.19* .25**News Mags .12 .06

ExpertiseDrama .12 .04Sitcom –.16 .00

*p < .05. **p < .01. ***p < .001. N = 134–139

Finally, an examination of Hypothesis 8 can be made by comparing thealready-reported results of the top halves of Tables 3 and 4 with the bottomhalves, in which selective viewing of a genre (the residual) is removed by partialcorrelation. (No partial correlation is shown for need for cognition in Table 4,because no genre-specific selective viewing bears on its relationships withthinking or surfing.) If these during-viewing activities resulted from selectiveexposure, this control should greatly reduce correlations with predispositionvariables. Examination of Tables 3 and 4 will show that this is clearly not thecase, because correlations generally do not shrink significantly. Two of thelargest changes actually are increases (e.g., from .23 to .32 for situation comedycontent preference and thinking), but even these are not statistically significantchanges (z < 1 by test for difference of correlations; Hays, 1963). Hypothesis 8is therefore rejected, and it appears that predispositions affect both selectiveexposure and during-viewing activities directly.

DISCUSSION

In trying to predict what television college students selectively viewed and howthey watched when they viewed, we expected some consistent processes tooperate. First, we predicted that instrumental reasons for viewing particulargenres (here drama, situation comedies, and news and entertainment magazines)would be associated with more selective viewing of these genres, but that casualreasons would be unrelated. We also expected that self-perceived expertise witha genre would be associated with viewing that genre. We also examined whetherneed for cognition was associated with viewing some genres more than others,because some genres (news, and to a lesser extent, dramas) may require andreward more thought. Because all of these are implicitly about selectiveprocesses directing viewing decisions, our tests of these issues always controlledfor other factors that differentiate heavy and light viewers overall.

In examining how carefully students attended to different genres while theyviewed them, we made similar predictions, but there were also differences that tookinto account the difference between selecting a genre and attending to it. A l t h o u g hperceived genre expertise and need for cognition were expected to be related toattention just as they had been to selective exposure, only one of two instrumentalgratifications was expected to predict attention (content specific but not moodmanagement). Both instrumental gratifications were expected to predict thinkingduring viewing, although perhaps with weaker correlations, because thinking wasmeasured for television in general rather than any specific genre. Thinking wasalso predicted to be strongly related to one’s overall need for cognition.


The story in these results was one of differentiation in several ways: mediumversus genre, between genres, selection versus activity, and betweengratifications. Our results confirm the importance of genre in two ways. Mostbasically, specifying the genre of a predisposition and activity appeared to locatestronger relationships, although these comparisons involved different variables.Nonetheless, Table 2 showed a number of significant relationships for genre-specific predispositions (gratifications and expertise) but not for more generalneed for cognition. And in terms of dependent variables, relationships predictinggenre-specific attention (Table 3) tended to be larger than those involvingmedium-level thinking for channel surfing (Table 4).

More particularly, relationships between predispositions and media use varyconsiderably between genres. Recall that medium-level gratification measureshave had relatively weak correlations with media use, and that those weakcorrelations are often for ritual rather than instrumental gratifications. Oneprevious study that actually compared genre-specific instrumental gratificationswith selective exposure (Kim & Rubin, 1997) found no relationship for soapopera viewing. Here, using three different genres, we found a number ofsignificant relationships, and although many relations were similar, some variedacross genre. Selective exposure due to instrumental reasons was less consistentfor situation comedy than for drama or news magazines, and perceived expertisewith the genre predicted attention to drama but not situation comedy.

It is conceivable that the weaker results for situation comedy may reflectgreater variance in situation comedy viewing for this two-thirds female sample,given Meadowcroft and Zillmann’s report of an association of comedy use formood management and the menstrual cycle (1987). Although cycle position waslikely randomly distributed across the women in this sample, such variationcould be reflected in the two smaller correlations of Table 2. However, it is alsoworth pointing out that there were no overall gender differences in amount ofsituation comedy viewing.

Need for cognition, a measure of general cognitive style that we had expectedto predict selective viewing, attention, and thinking, despite not being genre-specific, did not confirm expectations. It was largely unrelated to selectiveviewing, negatively related to attention to two of three genres for which norelationship had been predicted (dramas and situation comedies), and onlynonsignificantly positively related to attention to news magazines, where apositive correlation was predicted. These results may reflect an underlying truthin an old stereotype of television: There may not be enough that is mentallystimulating in most television to reward the thoughtful. Greater attention to orthinking during television viewing may be a situational response reflectinggreater interest, but it probably does not reflect this long-term cognitive style.

2 5 8 HAWKINS ETAL.

In predicting media use, these results with multiple genres also demonstratedthe need to distinguish between selective exposure and during-viewing activities,and between different predispositions. Although both instrumental gratifications(both mood/stimulation and content preference) were related to selectiveviewing, only content preference was related to attention, but this correlation wasstronger than for selective viewing and present for all three genres. That is,content-preference reasons for viewing were related more strongly to watchingattentively than they were to selecting that content, whereas mood/stimulationapparently led to viewing but did not increase attention during that viewing.Although both mood/stimulation and content preference treat viewing as a meansto an end, careful attention to content is also part of that means for contentpreference, but is unnecessary to mood-management functions. Thus, althoughboth gratifications are instrumental, they may well have very diff e r e n tconsequences on different aspects of television use.

Given this, however, it is surprising that both instrumental gratifications forboth entertainment genres were positively correlated with thinking duringviewing. One would have expected instead that thinking would be closelyassociated with attention, and thus not predicted by a mood-managementgratification. It could be that our measures of thinking during viewing actuallytapped amount of engagement with television programs. That is, almost all theitems contained in the measure describe various reactions to and connectionswith programming, and only one is merely “thinking hard.” These variousactivities may enumerate low-effort cognitive activities that result from anynoncasual engagement with television. And this explanation is then consistentwith thinking being unrelated to need for cognition, which taps more effortfulkinds of thinking. Further research will have to make finer distinctions amongduring-viewing activities to sort this out.

Casual or accidental viewing, not surprisingly, was associated with lesserattention during viewing, and perhaps also with less selective viewing as well.And watching a genre for these accidental reasons was positively correlated withfrequent channel changes or dividing attention between multiple programs.

Given these complexities, these results clarify something important aboutgratifications. The overall distinction between instrumental and ritualgratifications is simple and conceptually appealing, and it may sometimes betheoretically useful as well. But attempts to apply the simple distinction must beconstantly alert to the likelihood that differences between individual gratificationswithin these overall categories may be as or more important. In the current study,clearly not all instrumental gratifications are the same. One must then wonderwhether this is true for some applications of ritual gratifications as well.


F i n a l l y, finding that predispositions apparently affect during-viewingactivities directly rather than working through their effect on selective viewing isinstructive. These results should send us back to reconsider our general maps ofthis problem, as in the classic model of the uses and gratifications approachoutlined by Rosengren and used fruitfully by so many others (Palmgreen,Wenner, & Rosengren, 1985). This model specifies relationships between manyindividual, societal, and media characteristics, while outlining the essentialargument that gratifications sought lead to media exposure, activities, andvariations in effects. Our results suggest elaborating the model by showing adistinction within the “communicative behaviors” step to separate exposure fromduring-exposure activities, both of which have direct and separate links frommedia expectations (gratifications), perhaps as well as links from longer termpersonal characteristics directly to activities bypassing expectations.

REFERENCES

Atkin, C. (1985). Informational utility and selective exposure to entertainmentmedia. In D. Zillmann & J. Bryant (Eds.), Selective exposure tocommunication (pp. 63–91). Hillsdale, NJ: Lawrence Erlbaum Associates.

Blumler, J. G. (1979). The role of theory in uses and gratifications studies.Communication Research, 6, 9–36.

Bryant, J., & Zillmann, D. (1984). Using television to alleviate boredom andstress: Selective exposure as a function of induced excitation states. Journalof Broadcasting, 28, 1–20.

Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal ofPersonality and Social Psychology, 42, 116–131.

Cantor, M., & Pingree, S. (1983). The soap opera. Beverly Hills, CA: Sage.

Courtright, J. A. (1996). Rationally thinking about nonprobability. Journal ofBroadcasting and Electronic Media, 40, 414–421.

Frank, R. E., Becknell, J., & Clokey, J. (1971). Television program types.Journal of Marketing Research, 8, 204–211.

Gensch, D. H., & Ranganathan, B. (1974). Evaluation of television programcontent for the purposes of promotional segmentation. Journal of MarketingResearch, 11, 390–398.

Gerbner, G., & Gross, L. (1976). Living with television: The violence profile.Journal of Communication, 26(2), 173–199.

2 6 0 HAWKINS ETAL.

Hawkins, R. P., & Pingree, S. (1981). Uniform content and habitual viewing:Unnecessary assumptions in social reality effects. Human CommunicationResearch, 7, 291–301.

Hawkins, R. P,. & Pingree, S. (1986). Activity in the effects of television onchildren. In J. Bryant & D. Zillmann (Eds.), Perspectives on media effects(pp. 233–250). Hillsdale, NJ: Lawrence Erlbaum Associates.

Hays, W. L. (1963). Statistics. New York: Holt, Rinehart & Winston.

Jensen, K. B. (1987). Qualitative audience research: Toward an integrativeapproach to reception. Critical Studies in Mass Communication, 4, 21–36.

Katz, E. (1996). Viewers work. In J. Hay, L. Grossberg, & E. Wartella (Eds.), Theaudience and its landscape (pp. 9–21). Boulder, CO: Westview Press.

Kim, J., & Rubin, A. M. (1997). The variable influence of audience activity onmedia effects. Communication Research, 24, 107–135.

Kubey, R., & Csikszentmihalyi, M. (1990). Television and the quality of life:How viewing shapes everyday experience. Hillsdale, NJ: LawrenceErlbaum Associates.

Levy, M., & Windahl, S. (1984). Audience activity and gratifications: Aconceptual clarification and exploration. Communication Research, 11,51–78.

Lin, C. A. (1993). Modeling the gratification-seeking process of televisionviewing. Human Communication Research, 20, 224–244.

Livingstone, S. (1992). The resourceful reader: Interpreting television charactersand narratives. Communication Yearbook, 15, 58–90.

Meandowcroft, J. M., & Zillmann, D. (1987). Women’s comedy preferencesduring the menstrual cycle. Communication Research, 14, 204–218.

Morley, D. (1980). The nationwide audience: Structure and decoding. London:British Film Institute.

Palmgreen, P., Wenner, L., & Rosengren, K. E. (1985). Uses and gratificationsresearch: The past ten years. In K. E. Rosengren, L. Wenner, & P. Palmgreen(Eds.), Media gratifications research (pp. 11–37). Beverly Hills: Sage.

Perse, E. M. (1990). Media involvement and local news effects. Journal ofBroadcasting and Electronic Media, 34, 17–36.

Perse, E. M. (1998). Implications of cognitive and affective involvement forchannel changing. Journal of Communication, 48(3), 49–68.


Pingree, S., Hawkins, R., Hitchon, J., Gilligan, E., Radler, B., Kahlor, L.,Gorham, B., Kolbeins, G., Schmidt, T., & Kannaovakun, P. (in press). Ifcollege students are appointment tv viewers…. Journal of Broadcasting andElectronic Media.

Potts, R., Dedmon, A., & Halford, J. (1996). Sensation seeking, televisionviewing motives, and home television viewing patterns. Personality andIndividual Difference, 21, 1081–1084.

Preston, J. M., & Clair, S. A. (1994). Selective viewing: Cognition, personalityand television genres. British Journal of Social Psychology, 33, 273–288.

Rao, V. R. (1975). Taxonomy of television programs based on viewing behavior.Journal of Marketing Research, 12, 355–358.

Rayburn, J. D., & Palmgreen, P. (1984). Merging uses and gratifications andexpectancy-value theory. Communication Research, 11, 537–562.

Robinson, J. P., & Levy, M. R. (1986). The main source: Learning fromtelevision news. Beverly Hills, CA: Sage.

Rubin, A. M. (1981). A multivariate analysis of “60 Minutes” viewingmotivations. Journalism Quarterly, 58, 529–534.

Rubin, A. M. (1983). Television uses and gratifications: The interactions ofviewing patterns and motivations. Journal of Broadcasting, 27, 37–51.

Rubin, A. M. (1985). Uses of daytime television soap opera by college students.Journal of Broadcasting and Electronic Media, 29, 241–258.

Rubin, A. M. (1994). Media uses and effects: A uses and gratificationsperspective. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances intheory and research. Hillsdale, NJ: Lawrence Erlbaum Associates.

Rubin, A. M., & Perse, E. M. (1987a). Audience activity and soap operainvolvement: A uses and effects investigation. Human CommunicationResearch, 14, 246–268.

Rubin, A. M., & Perse, E. M. (1987b). Audience activity and television newsgratifications. Communication Research, 14, 58–84.

Sparks, G. (1995). Comments concerning the claim that mass media research is“prescientific”: A response to Potter, Cooper, and Dupagne. CommunicationTheory, 5, 273–280.

Tangney, J. P. (1988). Aspects of the family and children’s television viewingcontent preferences. Child Development, 59, 1070–1079.

Weaver, J. B., III. (1991). Exploring the links between personality and mediapreferences. Personality and Individual Difference, 12, 1293–1299.

2 6 2 HAWKINS ETAL.

Webster, J. G., & Newton, G. D. (1988). Structural determinants of the televisionnews audience. Journal of Broadcasting and Electronic Media, 32,381–389.

Webster, J. G., & Wakshlag, J. J. (1983). A theory of television program choice.Communication Research, 10, 430–446.

We n n e r, L. (1986). Model specification and theoretical development ingratifications sought and obtained research. Communication Monographs,53, 160–179.

Wober, J. M., & Gunter, B. (1986). Exploring the entertainment experience. In J.Bryant & D. Zillmann (Eds.), Perspectives on media effects. Hillsdale, NJ:Lawrence Erlbaum Associates.


Predicting Selection and Activity in Television Genre Viewing

Documents