□ Blumler, Jay G., Dennis Mc Quail & J. R. Brown: 'The Conduct of Exploratory Research into the Social Origins of Broadcasting Audiences'

Particip@tions Volume 1, Issue 1 (November 2003)

Chapter 5

The Development of an Instrument Designed to Investigate the Gratifications Sought from Television in General

Introduction 

While continuing to execute individual programme studies, the need for a more ‘general instrument’, capable of measuring the gratifications sought from the medium of television as a whole, was kept in mind.  Although this need could not be met merely by increasing the number of programme studies, the latter made a vital contribution to the development of such an instrument in two ways. First, by providing convincing evidence that different kinds of programme attracted similar and overlapping gratification structures, they tended to confirm the viability of the underlying concepts that a general instrument would be expected to operationalise. Second, they supplied the raw material for many of the items that were included in different trial versions of the desired technique.

As the project’s series of programme studies was being carried out, a methodical exploration of more general question formats was also being undertaken, and this chapter describes the main stages through which a possible instrument was developed. The culmination of this sequence of effort was a data-collection procedure that was incorporated into the third and final survey of the project’s first stage (conducted in July, 1970). Only a preliminary analysis of the resulting data can be reported at this time, but the outcome suggests that a general instrument similar to the one developed can be used to achieve the objectives of the second stage of the envisaged research.

The Place of a General Instrument in the Proposed Research

The central thesis of the project involves a rejection of the notion of the mass audience. The audience for a medium of communication consists of individuals whose relationships to it are differentially structured according to their personalities and social experiences. Of course these relationships (which the investigators have termed ‘media-person interactions’) are themselves the products of numerous interacting influences and causal factors. A paramount aim of the project is to classify some of the important audience relationships to television into their main types and to look for certain (i.e. sociological) explanations of them. Although individual programme studies have proved successful in tracing the satisfactions sought from certain media materials back to the social origins of requirements for particular gratifications, this detailed but piecemeal approach cannot for practical reasons be applied to all programmes. Neither would it provide answers to questions about orientations to the medium as a whole. Description and analysis at a higher level of generality are called for, and consequently different techniques are also required.

The difficulties involved in trying to devise a suitable approach are numerous and formidable, however, as the following pages attest. Perhaps the key problem stems from the need for a general instrument itself.  How does one ask viewers why they watch television as such, without running the two-fold risk a) of inviting references to a highly generalised image of what the medium should be used for and b) of losing all that rich concrete motivational content that evidently forges the attachment of audience members to particular programmes? The investigators’ final solution to this problem enabled respondents to relate a set of individual gratification items to a set of specific programmes, the latter having been chosen by a procedure that ensured they represented a sample of the individual’s recent viewing activities. But before arriving at this approach much exploratory work was undertaken, the four main stages of which are outlined below.

Repertory Grid Interviews

As a first attempt to grapple with the elements of a general approach to the study of TV gratifications, it was decided to try out the repertory grid interview technique. This seemed promising because it provides standardised procedures for eliciting qualitative data from respondents in a form that is amenable to statistical analysis. The outcome portrays in nearly spatial terms the relationships to each other of the key components of a person’s ideas about a subject.

Consequently, fifteen repertory grid interviews, lasting from one to three hours, were conducted in January, 1970, in the Leeds area.  In each case an identical sequence of activities was followed.  First, the respondent was asked to sort 60 cards, each bearing the name of the television programme, in such a way that his favourite programmes (aiming at approximately ten) were placed in our pile. The remaining cards were then removed, and only the favourites were used thereafter. At this point a method of triads was employed to elicit constructs from the respondent. That is, the interviewer chose three of the respondent’s favourite programmes at random and asked him to say which two of them gave him something that the third did not provide. After the dimension of distinction (or construct) underlying the respondent’s choice was elicited, the same procedure was followed with another set of three programmes, and so on until no fresh constructs emerged.[1] Finally, the respondent was asked to rank all of his ten favourite programmes on each of his constructs.

One example of the material that such an interview can yield, and of how it can be analysed, is presented here. In this case the respondent was a 40-year-old milkman, who produced the following eight constructs out of comparison and contrasts among 12 favourite programmes:

1.      Covers many things – covers only one topic.

2.      Makes me feel nostalgic – is viewed just for pleasure.

3.      Produces feelings of frustration and helpless – does not make me feel that way.

4.      Entertaining; you can sit back and enjoy it with a blank mind – tells you about things that are happening.

5.      Gives you other people’s opinions and views – gives you knowledge.

6.      Serious and heavy – light-hearted.

7.      Challenges one’s own opinions – enjoyable without having to work at it.

8.      Like best – like least.

The first step in analysing these data was to draw up the following table, showing how the respondent’s favourite programmes had been ranked on each of these constructs (using in each case the left-hand criterion in the above list).

TABLE V.1

One Respondent’s Ranking of Elements (Favourite TV Programmes) on Constructs (Features of Programmes)

Constructs

Spearman’s rho (a rank order coefficient of correlation) was then calculated to express the relationships between all pairs of constructs.  In order to extend the scale on which these associations could be visualised (converting the scoring range of rho from +1 to –1 into +100 to –100), each value of rho was squared and multiplied by 100, giving the matrix of relationship scores that appears in Table V.2.

TABLE V.2

Matrix of Relationship Scores between One Respondent’s Constructs

Constructs

When such relationship scores are summated regardless of sign, the resulting figures express the amount of common variance which is accounted for by each construct:

The next stage of the analysis aimed to identify those constructs which accounted for the greatest amount of variance while being unrelated (orthogonal) to each other.  The figures immediately above show that the first dominant construct is 4 (total relationship score = 179). Although 4 is followed in magnitude by 1 and 2, according to Table V.2 these constructs are significantly related to 4.[2] Construct 3, therefore, accounts for the next highest proportion of variance while being unrelated to 4.  Therefore, constructs 3 and 4 are presented as orthogonal axes in Figure V.1 over, while the remaining constructs could be plotted into the resulting space.

Seven of the repertory grid interviews were intensively analysed in this manner, and four main conclusions were drawn from the exercise as a whole. First, it was clear that detailed repertory grid interviewing could elicit a rich and wide range of orientations to TV output, even from viewers with a minimal education. On the other hand, it was difficult to confine the resulting constructs to the gratification survey as such (e.g. the milkman’s descriptive dimension of the number of topics covered in a programme); excessively general dimensions could emerge (e.g. the milkman’s liked best/liked least); and the procedure sometimes sacrificed logic to empiricism (e.g. the milkman’s contrast of nostalgia to pleasure). Second, the analysis of repertory grid material showed that respondents’ ideas about programmes had structured relationships – that is they were opposed to and inter-related with each other in patterned and probably stable ways. Third, it was evident that, for any individual, different programmes could be ranked on the main dimensions of judgement and that, for groups of respondents, separate profiles for each programme might conceivably be drawn. On the other hand, the latter possibility was substantially frustrated by a fourth and more fundamental feature of the data: considerable variation between respondents in the number and types of constructs that were produced.[3] Although the repertory grid analysis could facilitate an intensive exploration of the structure of an individual’s outlook, it did not lend itself to a generation of a common repertoire of constructs which could be applied across respondents. In that sense it was literally unsuited to the provision of a general instrument.

A Second Model of a General Instrument

Experience with the repertory grid technique having highlighted the problem of securing comparable data from all interviews, it was decided that greater reliance should be placed on recognition by the respondent of pre-formulated items that matched his dispositions and reactions. What was needed was a workable method of getting respondents to relate a set of individual gratification statements to a series of programmes. The first attempt to satisfy this criterion was embodied in an item that was administered in the second survey to the fans of quiz programmes. It so happens that this item was limited in its application to one programme type, for at this stage the investigators were mainly seeking answers to three questions about the practicality and utility of a particular question format:

1.      Could interviewers successfully administer the complex procedures involved in such a model of a general instrument?

2.      Would viewers respond to them without boredom, fatigue and ensuing bias?

3.      Would analysis of the generated data discriminate between groups of respondents according to the particulars of their social background?

Should these questions be answered affirmatively, the task of expanding the range of programmes under investigation could be tackled next.

Details of question-wording, format layout and instructions to interviewers may be gleaned by consulting item 14 of the Coronation Street/Quiz questionnaire in the Appendix. In essence, viewers who had ever seen at least three of four quiz programmes (Ask The Family, TV Brain of Britain, Top of the Form and University Challenge) were given a card bearing the following five reasons for watching such programmes:[4] 

A.     It helps me to pick up useful information.

B.     I like to try to pick the winners.

C.     It’s nice to see the contestants taken down a peg or two.

D.     I enjoy it when a likeable contestant does well.

E.      It is very satisfying when I get a question right.

The respondent then ranked all the programmes as vehicles for providing each of the designated gratifications in turn (e.g. Top of the Form might be best for supplying useful information, Brain of Britain next best for that purpose and so on). In addition, the respondent was asked to rank the programmes for their importance to him personally.

Interviewer reports and questionnaire inspection indicated that, once the interviewer had become versed in the procedure of administration, neither she nor the respondent encountered much difficulty.  From these points of view the exercise was deemed a success, and the format was not regarded as unduly complex for use on a wider scale. But did it discriminate between groups? To test this, the sample was divided by sex into two sub-groups. The raw data were tabulated as rank frequency distributions for each programme on each item (as well as for preferences). For each distribution the sum of ranks (Rj) and number of responses was calculated. Since there was variation between respondents in the number of programmes ranked (three or four), the maximum Rj range was not consistent.  Therefore Rj was re-scored to fall within a standardised range of 1-100 in such a way that the size of Rj, and the rank-order importance of the item for the group, were inversely related. That is, a score of 100 would indicate that all the respondents had agreed that a particular programme was least suited to providing the designated satisfaction. The results are presented in Table V.3.

PROGRAMMES

It can be seen that this procedure has managed to pinpoint a number of sex differences in the gratifications the respondents associated with individual quiz programmes. The favourite programme of the men was University Challenge, whereas the women’s favourite was Ask The Family, and these preferences were to some extent reflected in the rankings of the programmes on the various gratification statements (notably on item B – enjoying trying to pick the winners). Nevertheless, whereas women rated Ask The Family as best for acquiring useful information (item A), men found TV Brain of Britain best for that purpose, and a similar pattern emerged from the responses to item D (appreciating seeing likeable contestants do well on a programme). And whereas men most enjoyed getting the rights answers to the questions asked on University Challenge (item E), women found TV Brain of Britain most satisfying in that respect.

The reactions to item C, ‘It’s nice to see the contestants taken down a peg or two’, proved quite provocative. Both sexes agreed that University Challenge (with its supposed teams of well-educated competitors) was the most potent source of this satisfaction and that Top of the Form (featuring children) was least likely to supply it. It was intriguing to find, therefore, that when the gratification statements themselves were ranked for their importance to the respondent, the anti-contestant function of quizzes came last in the batting order for both sexes. Yet University Challenge was regarded very highly by both sexes precisely for the opportunities it offered for seeing the contestants make mistakes. This pattern strongly suggests that, whereas the direct rankings of the gratification statements had been influenced by considerations of social acceptability (people did not like to admit that they were moved by hostility to performers), more complex procedures, involving the application of such statements of named programmes, had helped to overcome this possible source of bias.

Thus this format had passed the three tests of its suitability that had been posed in advance and revealed an additional advantage as well. But two problems had still to be tackled in further work on a general instrument:

1.  Since the ranking of programmes on statements provides only a limited (and a laborious) basis for quantitative analysis, some method of generating raw data amenable to wider range of statistical calculations was needed.

2.  A way of eliciting ratings of a sample of television programmes that would be representative of respondents' more general viewing patterns had yet to be devised.

A Third Approach to a General Instrument

To meet the first requirement, a simple two-point rating scale (endorse/not endorse) replaced the ranking procedure. Involving less work for the respondent, this allowed the total number of programmes to be rated to be substantially increased.  Instead of four quizzes, 12 well-known and popular current programmes were incorporated into the next version of a general instrument. A total of 21 gratifications was also included, many of the meanings of which were drawn from previous project activities – e.g. group discussions, repertory grid interviews and the outcomes of the Dales and quiz cluster analyses.  The resulting two-page form, which was designed for self-completion and postal return, and was sent in June, 1970, to 20 adults living in the Leeds area, is reproduced in full below.

Only 13 people returned forms, and this low response, plus an inspection of endorsement patterns, helped to identify four problems that still demanded attention. First, a few gratification statements had not been endorsed while others had been applied to programmes in a stereotyped fashion. To counteract these tendencies the offending items were carefully reworded. Second, since the low rate of return was ascribed partly to a poor standard of format arrangement, an attempt was made to design a neater and more attractive version. Third, it was suspected that some respondents needed a more personal explanation of what was expected of them when completing the form. This suggested that, in the first instance at least, it should be presented to them by an interviewer and not through the post. Finally, the method of handling programmes was still unsatisfactory, for a) their selection had been determined not by the viewers’ own behaviour but by external criteria (familiarity and popularity) and b) respondents necessarily had varied in the number of listed programmes they rated, since any on the form that had not recently been seen had first to be crossed out.

Here is a list of statements that people have made about TV programmes. Across the top of the page are the names of 12 TV programmes. First please cross out the names of any programmes that you have not seen during the last few months.

Now for each statement decide whether it expresses how you feel about any of the programmes that you have seen. Put a tick in the box under all those programmes to which the statement applies.  Work through all 20 statements in this way. If a statement does not express your feelings about any of the listed programmes, leave that row of boxes blank.