TEE USE OF THE SEMANTIC DIFFERENTIAL TECHNIQUE IN THE

ANALYSIS OF THE IMAGES OF THREE COLUMBUS

COMMERCIAL TEIEVISION STATIONS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy In the Graduate School of The State University

By FRANKLIN DAVID SABAH, B. A., M. A

The Ohio State University 1959

Approved by

Adviser Department of Speech ACKNOWLEDGEMENTS

Expression of sincere appreciation Is extended here to the following members of my graduate committee:

Chairman Franklin H. Knower# Wallace C. Fotheringham,

Rlohard M. Mall, Paul A. Carmack, and John E. Horrocks.

Thanks, also, go to Omar S. Goode of the Bureau of

Business Research for his assistance in the IBM tabula­ tion of the study's test data.

11 TABLE OF CONTENTS

INTRODUCTION TO THE STUDY ...... 1 The Problem • ...... 1 Hypotheses to be Tested ...... 9 The Stations ••••••«••••• 10 Organisation of the Remainder of This Dissertation ...... 11

REVIEW OF THE L I T E R A T U R E ...... 12 On the Semantic Differential • • • • 12 On Station Image ...... 18

METHODOLOGY OF THE STUDY ...... 31 The "Holistic” Pilot Test ...... 31 The c o n c e p t s ...... 31 The gradients ...... 31 Construction of the test • • • • 33 The "Analytic" Pilot Test • • • • • 34 The concepts •••• ...... 34 The gradients ...... 36 Construction of the test • • • • 36 Administration of the Pilot Tests • 38 Results ••••••* ...... •» 39 The holistic test •••••••• 41 The analytic test ...... 44

The Final Tests ...... 50 Construction of the final tests • 50 Administration of the final tests 52

RESULTS ...... 55 Stability of the Final Tests1 Sosple 55 Reliability of the Final Tests • • • 57 Validity of the Final Tests • • • . 57 General Result s •••••••••* 61 The holistic test ...... 68 The analytic test ••••••• • 71 The hypotheses ••••••••• 74

SUMMARY AND CONCLUSIONS ...... 78 Summary 78 Conclusions ...... Discussion ••••••••••* Implications for Further Study • • .

ill TABUS OF CONTENTS (Continued)

CHAPTER PAGE APPENDIX A. Preference Group Mean Responses on the Holistlo and Analytic Final T e s t s ...... • 86

APPENDIX B. Holistic Pilot Test ...... 89

APPENDIX C. Analytic Pilot Test and Questionnaire 91

APPENDIX D. Holistic and Analytic Final Tests and Questlonnal re ...... 104

BIBLIOGRAPHY...... 112

AUTOBIOGRAPHY...... 116

I t LIST OF TABXJSS

TABIE PAGE

I. List of Semantic Differential Soalea Used in Goldberg's Study . . • • 20

II* Twenty Gradients Used in Holistic and Analytic Pilot Tests ...... • • • • 32

III* Ten Features of a 's Programming Used as Concepts in the Analytic Pilot Teat ...... 36

IV. A Breakdown of Descriptive Factors of the Forty Respondents to the Pilot Tests • 40

V. F Values of All Twenty Gradients on Holistic ” Pilot T e a t ...... 42

VI. t Values of All Twenty Gradients for Eaoh — Pair of Stations on Holistic Pilot Teat 44

VII. t Values of All Ten Concepts for Eaoh Pair ” of Stations on Analytic Pilot Test • • 47

VIII. t Values of All Twenty Gradients for Each — Pair of Stations on Analytic Pilot Teat 49

IX. A Breakdown of Descriptive Factors of the One Hundred Respondents to the Final Tests «•.••••#•«•••••• 54

X. Reliability Coefficients and Fiducial Limits of the Final Tests' Sanjple • • 56

XI. Mean Values of Bach Preference Group for Bach Station on Bach Final Test • • • 60

XII* Mean Responses on the Holistlo Final Test 62

XIII* Mean Responses on the Analytic Final Test 64

XIV* Summary of F Values on Holistic and Analytic~fenal Tests and Between Pairs of Stations on Total Tests • • • 68

v LIST OF TABLES (Continued)

TABLE PAGE

XV. Kolmogorov-Smlrnov D Values of all Eight Gradients for Ea?h Fair of Stations on Holistic Final T e s t ...... 70

XVI. Mann-Whitney U Values of All Ten Concepts for Each Pair of Stations on Analytlo Final T e s t ...... 75

XVII. Slse of Audience Figures and Total Holistic Test Values for all Three Stations 77

Yi LIST OP FIGURES

FIGURE PAGE 1. Profiles of Mean Responses on the Holistlo Final Test *•••••••• 68

2. Profiles of Concept Mean Responses on the Analytic Final Test ...... 66

▼ii CHAPTER I

INTRODUCTION TO THE STUDY

The Problem

M a n y people are raising questions regarding the

Influence of radio and television on the American scene*

Such questions have been emphasized as a result of the

near saturation of the population by these broadcasting

media* As will be noted later, various writers in the

field have discussed problems of radio and television

arising In the light of the media's saturation and possible

influence* It seems desirable, then, that many studies

be devoted to these social phenomena.

The tremendous extent of radio and television as mass media In the United States may be Illustrated in part by several statistics* As of the first week of

January 1959, estimates of the population indicated that there 126,250,000 people in the United States over

12 years of age* These people spent a total of 3,595*5 million hours enjoying the broadcasting media (1,914*6 million hours watching television and 1,480*7 million hours listening to radioJ* During this same period, only

725*2 million hours were spent with the media of newspapers.

1 2 magazines, and movies.'1' The following figures further show the near saturation of the population by these media.

According to the 1958 Broadcasting Yearbook, "Radio set ownership extended to virtually 97 per cent of all

United States homes, and television saturation was 85 per cent in mid-Summer 1958. Homes equipped with one or more radio sets numbered 48.9 million. Those with television numbered 43 million.-2

Perhaps a fact which makes these statistics all the more impressive is that commercial radio has existed less than 40 years (beginning with station KDKA, Pittsburgh, T Pennsylvania on November 2, 1920), while television, al­ though experimented with and in some limited use before

1948, is generally considered to have been formally intro­ duced to the American public during that year. "Thus 1948 became a crucial year in the history of television - the year in which it emerged as a mass medium.-4

The rise of radio and television as mass media found them in a position to inform, education, entertain,

^Broadcaating, LVI (January 19, 1959), p. 40.

aBroadcasting Yearbook 1958, (Washington, B.C. t Broadcasting Publications, Inc., T958), p. B-131.

''Sidney Head, Broadcasting in America (: Houghton Hifflln Company, 1955), p . lOi •

4Ibid., p. 158. 3 and influence literally tens of millions of people at a time. Leo Bogart, In his book dealing with the Impact of television on American life, suggests that television is more than entertainment when he states that *. • • it is an endless source of ideas and information and a powerful

Influence on valuesIn this same vein, Charles S.

Steinberg states*

It has become increasingly evident that the very values and judgments by which contemporary man lives are closely connected with, and influ­ enced by, the communication media and opinion- influencing techniques. Mass media have become vital centers for the transmission of knowledge, the dissemination of facts, and the directing of various emotional appeals to Influence public opinion.® Because of the tremendous potential of the mass media, it is important that those in control of them exercise certain responsibilities. Some of the possible harmful effects of media under the control of a demagogue can be very readily envisioned. With this idea of re­ sponsibility in mind, various authors have indicated their feelings about the responsibilities that the mass media and mass communicators should assume.

6 Leo Bogart, The Age of Television (New York* Frederick Unger Publishing Company, ld5S), p. vlil.

^Charles S. Steinberg, The Mass Communlcators (New York: Harper and Brothers'^ 1958j, p. 4. Gilbert Seldes, in his book The Public Arte,7

discusses the blessings and responsibilities that the

electronic revolution has given to modern man. Writing

in Radio, Television and Society, Charles Siepmann states

that "Responsibility is the condition of liberty's con- Q cession." Probably Wilbur Schramm best sums up the

responsibilities of the mass media and the mass communl- 9 cators in Responsibility In Mass Communication. In

evolving his social responsibility theory relative to

the mass media, Schramn emphasizes the proposition that

the producer has responsibilities commensurate with his

freedom. According to Schramm the communicators' responsibilities are "• . .to turn out the highest

quality product they can; to be aware in depth and

breadth of the needs and interests or their publics;

and to maintain the free market place of ideas and the

self-righting process of truth. . • ."10

7 Gilbert Seldes, The Public Arts (New York* Simon and Schuster, 1956].

®Charles Siepmann, Radio, Television and Soolety (New York* Oxford University Press, 1056^, p. 22TT. Q Wilbur Schramm, Reap onslblllty in Mass Communication (Hew Yorkt Harper and BroTEers, 1957).

10Ibld., p. 334. 5

Broadcasters have become aware of and expressed concern about their responsibilities• There has been an increasing amount of emphasis attached to the impres­ sions that their media reflect to the public. Not only are broadcasters Interested in the type of information yielded in traditional radio-television studies, (that is, who listens to or watches what and at what time) but also these some broadcasters often have an interest in knowing In what regard their particular medium Is held in the eyes of the public. That is, to what extent does the public hold a particular attitude toward a station with regard to one or several particular standards of evaluation? One term which is often used to designate the object of concern of the broadcaster is station

"Image.** In disousslng this particular term In

Broadcasting. John E. MclSillln says "The Advertising

Catchword of 1958 was easily the over-worked word

1image.I"*1 Several other terms which are often used

Interchangeably with "image" are "inqpression" and

"personality." Thus, one often hears that a station owner or manager is curious and concerned scout the

"Image,” "impression," or "personality" that his station

11John E. HcMillln, "Commercial Commentary,” Broadcasting, LVI (January 5, 1959), p. 6. has or reflects In the community which it serves. As shall he seen In a later part of Chapter II, soma studies have been done In the broadcasting Industry in an attenpt to discover station "Image." These studies have not ex* hausted the prospect of profitable learning from this concept.

Since the station image has been so frequently considered important, the purpose of this study is one which Involves this problem. Traditionally, studies done

In the area of broadcasting have followed the pattern of marketing and public opinion polling. Commercial research organizations have ordinarily relied on such methods as the telephone coincidental, roster, diary, and mechanical recorder techniques to gather data relative to the broad­ casting media. Questions asked and information yielded from these methods often have not been subjected to pre­ cise statistical analysis. It is believed that the use of the semantic differential technique as a measuring

Instrument in broadcasting research will permit more meaningful analysis to the problem of station image under consideration. Thus, the primary purpose of this study is dual in nature (1) methodologically, to determine whether the semantic differential technique can be adapted constructively to the problem of measuring station Image with the ultimate objective being to 7 provide refined analysis of this problem as a result of the use of this measuring Instrument and statistical treatment of the data colected, and (2) to evaluate com­ paratively the three Columbus commercial television stations* This dual purpose led to hypothesis number one (see page nine)*

Another objective of this study la to determine whether or not the semantic differential technique, as a measuring Instrument In broadcasting research might pro­ vide a different analysis and evaluation of stations than that made available through typical rating services and research methods. It is also ooncelvable tnat this method of analyzing stations and listeners may point up factors not shown by other methods of broadcasting measurement and research. This question led to hypothesis number two (see page nine)*

The question of whether or not station Image might vary when considered from several points of view resulted

In the decision to study this problem from a "holistic"

(station as a whole) point of view ana an "analytic"

(station broken down Into various concepts or significant features of Its programming) point of view* It was deemed desirable to measure station image from a total or station as a whole point of view In order to arrive at some in­ dication of eaoh station's complete picture as envisioned by viewers* Also, generally speaking, it Is the sum total

of impressions reflected by a station that people refer to when asked to evaluate that station. Since stations vary from one to another In terms of specific elements and factors within their operations, it was considered de­ sirable to test whether or not some measurement of station image might point up specific areas In which stations differ* In view of this, station image measurement from an analytic point of view was undertaken* As will be noted later, those factors considered in the analytic approach dealt with features of programming only since such factors were considered most important in Influ­ encing one's attitudes toward a station* These two approaches to station image measurement, then, raised an interesting question which led to hypothesis number three (see page ten)•

An additional purpose of this study, of course, is to compare the Images of the three Columbus commercial television stations on the basis of semantic differential data collected to determine what these stations mean to one of the audiences for which they operate* lfce Images or personality profiles of these stations, when developed, should be of value (1) to the community In providing it with critical impressions of what the stations mean, connote, or reflect in terms of their operations; 9

(2) to station ownership and management In snowing them

the cultural impact of their Individual operations,

thereby enabling them to Improve their medium (total operations, programming, services) in the eyes of the

community on the basis of their relative strengths and weaknesses as Indicated by the profile impressions; and

(3) to other agencies directly connected with stations

(for example, advertising agencies) in providing them with a valuable "pictured of the stations.

Hypotheses to be Tested

It was assumed in this study that the Columbus commercial television stations do reflect some kind of

"image” to their viewing publics. Accepting this as­ sumption, this research was an attempt to measure these

Images in the light of the following hypotheses!

1. There la no difference between the Images

(impressions or personalities may be substituted here) of the three Columbus commercial television stations as measured by the semantic differential technique.

2. There 1s no difference between a station's image as measured by the semantic differential technique and its commercial ratings as determined by a rating

service. 10

3. There la no difference between a station'a image when measured from a total Image or "holistic"

(station as a whole) point of view and its image when measured from an "analytic" (station broken down into various concepts or significant features of programming) point of view.

The Stations

The three Coiumbus commercial television stations to be analyzed in this study are WjJI-C (channel 4),

WTVN (channel 6), and WBNS (channel 10).^ All tnree stations have been serving the Columbus market area approximately ten years, having begun operations in iy4y.

WLW-C is affiliated primarily wltn the National Broad­ casting Company, WTVN with the American Broadcasting

Company* and WBNS with the Columbia Broadcasting System.

1» WLW-C (TV) is owned by the Crosiey Broadcasting Corporation which also owns WLW and WUK-T (TV) * WlKf-D (TV) Dayton, WUT-A (TV) Atlanta, and WLW-I (TV) . Crosiey is a subsidiary of Avco Manufac­ turing Company. WTVN-TV is owned by tne Tart Stations' group which also owns WKRC-AM-FM-TV Cincinnati* WTVN-AM Columbus* WBRC-AM-FM-TV Birmingham* WBIR-AM-FM-TV Knoxville* and WKYT-TV Lexington* WBNS-TV is owned by the Dispatch Printing Company which publishes the Ohio State Journal and the Columbus Dispatch newspapers and also owns WBNS-AM-FM. 11

Organisation of tna Remainder or Till a Dissertation

Chapter II will be a review of background litera­ ture pertinent to this study.

Chapter III will present the sources of data and also the procedures and methodology followed In this research.

Chapter IV describes the analyses used in evalu­ ating the data and reports the results of these techniques.

Chapter V summarises the results of the study; presents conclusions derived from them, and considers the Implications of the study to current problems In evaluating station Image In the broadcasting industry and to future research. CHAPTER II

REVIEW OP THE LITERATURE

On the Semantic Differential

The notion of the semantic differential grew out

of research In the psychological experience of synesthesia.

Synesthesia, as defined by Warren in his Dictionary of

Psychologyj la phenomenon characterizing the experiences

of certain Individuals, in which certain sensations be*

longing to one sense or mode attach to certain sensations

of another group and appear regularly whenever a stimulus

of the latter type occurs."1 For exanqple, a happy person may be said to feel "high," music may be described as

"red," sadness may be experienced as "black," or "low," and so forth. Up until recently this phenomenon has been considered rare. Now, however, researchers feel that It might be the common basis of figurative use of language, and can be used as a basis for a technique of social research called the semantic differential.

The major research technique to be used In this study, the semantic differential, has been developed by

^H. C. Warren, Dlctlonyy of Psychology (Boston* Houghton Mifflin Company, 195^.

12 13 Osgood and others for the measurement of "meaning."^

The semantic differential is a combination of assocla- tional and rating scale procedures. Its purpose is to dlsoover the connotative meanings of concepts® by means of a set of seven-point scales. Each seven-point scale continuum lies between two psychologically polar terms such as "Good-Bad." Subjects mark on each scale the place where they judge the connotative meaning of a concept to fall.

Presumably, if the polar terms of a particular scale have no relat ion to the concept according to the subject doing the evaluating, he will place a mark in the center, at the neutral point of the scale. If one of the polar terms describes or applies to the concept perfectly, the subject will place the mark beside that term. By having subjects mark each concept on many scales (or gradients) it becomes possible to discover the connotative meanings of the concept as an outcome of the resulting rating profile.

Q Charles E. Osgood, "The Mature and Measurement of Meaning," Psychological Bulletin, XL (May, 1962), pp« 19*7—237 . « wConcept refers to a "stimulus" word to which a subject's checking operation is a "response." For example, "love, father, sports," etc. 14

Osgood describes the rationale of the semantic differential as follows:

Through the functioning of a generalization principle, the concept will elicit checking of that scale position whose dominant mediator component most closely matches In Intensity the corresponding component In the process associated with the concept Itself. Since the positions checked on the scales constitute the coordinates of the concept's location In semantic space, we assume that the coordinates In the measurement space are functionally equivalent with the components of the repre­ sentational mediation process associated with this concept.4

The following hypotheses have been postulated with reference to the semantic differential as a measuring instrument:

1. The process of description or Judgment can be conceived as the allocation of a concept to an experiential continuum, definable by a pair of polar terms.

2. Many different experiential continue, or ways In which meanings vary, are essentially equivalent and hence may be represented by a single dimension.

3. A limited number of such continue can be used to define a semantic space within which the meaning of any concept can be specified."

All of these hypotheses have been substantiated by research utilising the semantic differential.

4C. Osgood, G. Suci, and F. Tannenbaum, The Measurement of Meaning (Urbana: University of Illinois Press'," IPS7)7~p: ig « ®C* Osgood, oj>. olt., p. 227. 15

The criteria for evaluating almost any measuring

Instrument are basically the same* Major criteria usually

considered include objectivity, reliability, validity,

sensitivity, comparability, and utility* In evaluating the semantic differential in the light of these criteria,

it appears that the semantic differential may be con­

sidered a worth-while measuring instrument. Following is a discussion of how the semantic differential technique may be evaluated in terms of each of the six aforementioned major criteria.

Objectivity* Insofar as the procedures of measure­ ment with the semantic differential are clear and can be duplicated, the semantic differential is objective. Two investigators given the same data from a semantic differential research and using the same rules of analysis must end up with similar meanings and/or profiles for the concepts* Granted interpretation of these results may be subjective, but such is the case with nearly all research*

Reliability* This criterion concerns the consis­ tency with which similar results are yielded when research is duplicated under like conditions* Osgood reports several experiments which support the reliability of 16 g the semantic differential. Norman, in a study done

involving the notions of reliability and stability of the semantic differential, states "Consequently, the semantic differential can be recommended for use in investigations where this sort of 'meaning* is to be Tt measured over groups of subjects."

Validity. How well an instrument measures what it is supposed to measure is the concern of this cri­ terion. The semantic differential is designed to measure meaning. Thus, in order to test semantic differential validity, semantic differential scores should be cor- related with some Independent criterion of meaning.

However, since there is no commonly accepted quantita­ tive criterion of meaning, "face validity" (the extent to which an Instrument appears to measure what it's supposed to measure) is considered in evaluating the semantic differential in terms of a validity considera­ tion. Therefore, careiUl selection of scales and concepts relative to the objects to be measured becomes of con­ siderable importance.

6See Osgood, Sucl, and Tannenbaum, op. clt., pp. 126-40.

7Warren T. Norman, "Stability Characteristics of the Semantic Differential," Technical Report No. 19, The Role of Language in Behavior, University of Minnesota. 17

Sensitivity* This criterion deals with the dis­ criminatory ability of a measuring instrument. Ideally,

an instrument should provide finer distinctions than

those which can be made without the use of scientific

instruments. The semantic differential has been shown to discriminate significantly.

Cos^arablllty, To date, the uses of the semantic differential have appeared to satisfy the criterion of comparability - that criterion which asks that the meas­ uring instrument be applicable to a wide range of phe­ nomena and situations and provide for comparisons within this range of uses. The concern with comparability of the semantic differential obviously relates primarily to its uses over a wide range of subjects and concepts.

Research studies utilising this technique have dealt with broad areas (Including such areas as attitude measurement and communlcations research}; included a wide variety of subjects (for example, men, women, various educational levels, et cetera); and treated a wide variety of concepts (for example, China,

Eisenhower, myself, the Church, et cetera).

Utility. The semantic differential has been shown to be an efficient measuring technique. It is neither cumbersome nor laborious in its construction and adminis­ tration. It is economical of time and effort. 18

Finally, beyond the semantic differential's ability to meet the aforementioned six criteri *, it must be remembered that one of the purposes of the present research la to ascertain the adaptability and worthiness of this measuring technique to the problems under consideration.

On Station Image

One entire chapter In Osgood's The Measurement of Meaning is devoted to a discussion of the use of p semantic measurement In communications research. Various studies ranging from Investigations In psycholinguistics to research In comnunlcations effects are treated. At the time of this present study, however, the writer knew of only one study done In the area of broadcasting which used the semantic differential technique in an attempt to discover station image. Several other studies which concern themselves with the problem of station image but do not employ the semantic differential as a measuring technique will be discussed later.

Melvin A. Goldberg, director of research for the Westlnghouse Broadcasting Coiq>any, and Dr. Percy

Tannenbaum of the University of Illinois, conducted a

eOsgood, Suol, and Tannenbaum, op. cit., Chapter VII, "Semantic Measurement in Conmunicatlons Research.” 19 study In the Pittsburgh metropolitan area designed to get a deeper measure of a radio listener's reaction to o programming. The purpose of the study was to locate reasons "why" listeners might like one station's "sound” and not another and which of several stations studied would appear most like the respondents' conception of an ”ld6al" station. The semantic differential technique was used in this research.

A total of 266 respondents aged 13 and over who admitted to listening to the radio sometime during the

"past week" were used in the study. Each respondent was asked to evaluate five Pittsburgh radio stations and an "ideal" station. A separate identical form was used for each of the six evaluations. Respondents did not know the survey was being conducted for station KDKA or Westlnghouae. Table I lists the fourteen scales used in ail the evaluations.

Prom this study Westlnghouae learned that lis­ teners reliably describe how they would like an "ideal" station to sound and that listeners can also reveal in what areas a given station differs from this "ideal.”

Obviously the semantic differential profile of the

"ideal* station was nearest the end of the scale con­ tinue which represented the most favorable of the

®"Wew Way to Measure Stations," Sponsor XII {Kovember 23, 1957), pp. 42-43. 20

TABLE X

LIST OF SEMANTIC DIFFERENTIAL SCALES USED IN GOLDBERG'S PITTSBURGH STUDY

unpleasant - pleasant stuffy - casual unfair - fair rough - smooth tense - relaxed weak - strong worthless - valuable heavy - light stale - fresh low class - high class dull - exciting loud - soft passive - active usuax - unusual

bi-polar terms. Of the five leading Pittsburgh stations,

KDKA was found to come closest to the "ideal" station

image.

Another study in the area of broadcasting which was concerned with station image is described in "The

People Talk Back to RadioThis was a motivational research study conducted in Houston, Texas, for station

KPKC by the Institute for Motivational Research, Inc.

The puxposes of this study were to discover

1. The real needs and desires of the Houston radio audience today and how Houston stations fulfill them.

2• The attitudes of the listening audience concerning KPRC and other Houston stations.

*°"The People Talk Back to Radio" (A Motivational Research Study of the Houston Radio Audlenoe, 1959), Institute for Motivational Research, Inc., Croton-on- Hudson, Hew York. 21

3. Whether these attitudes had measurable effect on advertising carried by each station.

4. The difference, if any, between people who listen to KPRC and other Houston stations.

5. The missing elements in radio, which, if combined with the best of radio today, would give a blueprint for tomorrow's ideal radio s tat ion .H

The major research Instruments used in this study were depth interviews and projective tests. A total of

270 subjects representing both sexes and a variety of occupations, ages, educational levels, and socioeconomic statuses was used as a sample.

The findings of this study revealed that (1) KPRC is considered more reliable, believable, expert, pro­ fessional, authoritative, reputable, and educational than any other Houston station; (2) KPRC commercials are considered more reliable, more believable, and more trustworthy than commercials of any other station in

Houston; (3) more people turn to KPRC for news than to any other station; and, finally, (4) the group that listens most to KPRC contains a higher proportion or people from upper level occupations, professions, and social classes which group Includes a generally higher percentage of thought leaders, trend setters, and com­ munity active people.

U Ibid. 22

In the spring of 1957, Motivation Analysis, Inc. conducted a study for CBS Radio In an attempt to evaluate some qualitative differences between large network-owned stations and the leading competitive independent stations in the same citiesBasically, this study, like the other two heretofore discussed, was concerned with station image and/or personality. It grew out of a desire for answers to the following three questions!

1. Do listeners pay Closer attention to some radio stations than to othersT

2, Do listeners distinguish between stations, regarding some more favorably than others?

5. Are listeners more likely to believe some stations than others?*5

The study was conducted in the six cities in which

CBS-owned radio stations operate! New York, , Los

Angeles, Boston, St. Louis, and San Francisco. A total of 1,202 radio listeners were Interviewed with a semi­ structured depth interview technique.

Results of this research in terms of the three questions which were originally set up to be answered showed that (1) listeners pay more attention to CBS radio stations tnan to the leading Independents;

12 “Different" (An Evaluation of Some Qualitative Differences Between Raalo Stations, Based on a Study by Motivation Analysis, Inc., 1957), CBS Radio, New York.

l3Ibid 23

(2) listeners distinguish between stations. • • regard

CBS radio stations more favorably than the leaaing In*

dependents; and (3) listeners believe CBS radio stations

more than the leading independents. • • butn tneir pro­

grams and their advertising.

Pulse, Inc, conduoted a stuay for radio station

WWDC In Washington, D. C., designed to profile its 11 s- 14 tenlng aualence lnt,o people, not statistics. The

base for the study was one thousand people (5UO men ana

500 women) residing in ana around the District of

- Columbia. Ten questions were asked of the respondents:

1. When you first turn on the radio, what station do you tune to?

2. When you first turn on your radio, what do you listen to? (News, music, etc.)

3. What station do you first tune to for news?

4. What station do you first turn to for local news - news of the Washington area?

5. In time of an emergency, such as a heavy snowstorm, what station do you listen to for news of school closings, meeting post­ ponements, storm news, etc.?

6. What station comes closest to playing the kind of music you like?

7. If all stations but one were foroed to go off the air, what station would you like to remain?

i4"Personality Profile of a Radio Station" (A Study made of Station WWDC, Washington, D.C., July 1958), Pulse, Inc., Now York. 24

8* Which of the following words describes each of these stationsT Dull? Lively?

9. Which of the following words describes each of these stations? Old-fashioned? Modern?

10. Which of the following words describes each of these stations? Friendly? Unfriendly?

As may be noted, the last three questions are somewhat similar to the nature of a semantic differential measurement* That is, responses to bl-polar terms are used In evaluating and discriminating among stations•

It was the conclusion of this research that

WWDC has a definite personality to a very great number of people* The station Is considered lively, modern, and friendly* WWDC leads all other stations in sup­ plying the music and news the people want, in being tuned in first by more people than any other station, and in being of aid In emergencies.

Another study conducted by Pulse, Inc* in the area of station image was that done for radio station

WHDH In Boston*16 This study, like the preceding one done by Pulse for station WWDC* was conducted in 1958 and used one thousand people (500 men and 500 women)

16Ibid.

16"Radlofa 5 Years of Great Change - 1953/1958* (A Qialltative Survey of WHDH, Boston), Pulse, Inc., New York. 25 aa a baae sample. These people were residents both within and outside the Boston metropolitan area. Again, ques­ tions were asked of the respondents In an attempt to provide some answers to the study's objectives.

The objectives of the WHDH qualitative study can be summarized aa follows:

What are the reasons causing people to turn on radio? Program preferences? Awareness of network stations? Extent of out-of-home radio audience? Station preferences? Station image?17

It was found in 98.5 per cent of the replies that there is "reason" for turning on the radio. These reasons were related to the program preferences of radio lis­ teners, with music being the first choice and news and weather following. Another rinding In this study pointed up the dec Line of awareness on the part of radio lis­ teners with respect to what a network radio station Is and which stations are network affiliates. In line with these results, it was found that there has been a decline in the audience of stations and a rise ID in audience of Independent stations.

17Ibid, 18 It might be interesting to note how these net­ work stations versus Independent stations results oompare with the "Different" study results reported on pp. 22-23. 26

It alao was discovered In this research that during the paat five years there has been a substantial

increase in the size of the radio audience “out-of-home.*

In fact, three out of four respondents reported that they listened to radio away from home during the week.

With respect to the objective area of station preference, it was found that WHDH was preferred above any of the other area stations for music, news, sports, and all-around listening*

Finally, in dealing with the problem of station

Image, apparently only one question was asaedt "If you heard conflicting accounts of the same news story on different radio stations, which of these stations would you believe?”19 This study concludes by stating that

"A station’s image is at best an intangible thing, but its importance cannot be minimized. The WHDH image is one of bellevabllity and reflects the result of more than a decade of responsible management

KRNT radio and KRNT television in Des Moines,

Iowa, had a study done in depth by Central Surveys, Inc. in order to point up any evaluative differences between

^"Radio's 5 Years of Oreat Change - 1953/1958," op. clt. 20ibid. 27 those stations and any other stations in the Des Moines 21 market area. This study was done In January 1958 and is based on personal interviews with 400 heads or house­ holds in Des Moines and suburban areas.

Results of this study showed that both KRNT radio among radio stations, and KRNT television among television stations

1. have the largest listening and/or viewing audiences;

2. are considered most accurate in their news reporting; 3. are considered most rellaole, most dependable;

4. have personalities who are considered by far the most believable.22

A major contribution to the area of research dealing with station image was a study conducted by 23 Alfred Politz Research, Inc. in three different markets.

A major portion of this study waa devoted to developing the important qualitative characteristics of stations as reflected in the conqpositlon of the audience, the

2 "Des Moines In Depth" (Highlights of Media Study by Central Surveys, Inc. for KRNT and KRNT~TV), Central Surveys, Inc., Shenandoah, Iowa.

22Ibid.

25HRadlo Today - New Discoveries about the Constant Companion of the American People” (A survey in major markets conducted by Alfred Politz Research, Inc.), November 1954. 28 attitudes and opinions of people, and tne reasons under* lying the public's preferences and reactions.

The market areas surveyed were WJR, ;

WHAS, Louisville; and WGY, Albany-Troy-Schenectady.

During a three week period in November 1^54, there were a total of 1,800 personal interviews conducted in the

WJR area, 1,200 in the WHAS area, and 1,050 in the

WGY area.

Perhaps two of the generally significant contribu­ tions of this study were-*

1. The confirmation of the continuing vitality of radio in its unique role as the constant com­ panion of the American people;

2. The discovery that people are much more highly selective than ever supposed and In each market choose a particular station as their favored companion from among the many available stations In the area.24

It Is in the light of the second contribution listed above that this writer feels the Folits study treats most relevantly the problem of individual station Image.

In the three market areas under consideration, it was found that stations WJR, WHAS, and WGY represented the stations in their respective markets which

1. people listened to most frequently;

2. stood out In share of audience for the average week, average day, and every time segment of the day;

24Ibid 29

3* stood out as being best for specific types of programs Including news, music, and sports;

4. were considered most public-spirited;

5. were considered most helpful;

6. stood out In the quality of advertising carried - that Is, the most helpful, most reliable, and least irritating;

7. stood out in reliability and completeness In service programs;

0* people would tune to In case of an emergency*”®

In the fall of 1955, Alfred Politz Research, Inc• conducted a study similar to the preceding one discussed, however, this study was done for radio station WTIC in «>6 the Hartford, Connecticut market area." A sample of

1,005 personal Interviews was used In this research*

Essentially the same type information was considered in this study in relation to WTIC as in the previous

Politz study (that Is, Information relative to station watched most, station considered most reliable, et cetera)*

It was found that station WTIC held the same desirable

26Ibid.

2®*Radio Today - New Discoveries about the Constant Companion of the American People In the Intensive Listening Area of WTIC - Hartford, Connecti­ cut,* Alfred Politz Research, Inc*, November 1955* 30 position In Its market area as WJR, WHAS, and WGY entertained In their respective market areas.

In considering the studies here reported. It Is well to note that virtually little research has been devoted to the problem of station Image. Actually, as far as this writer has been able to determine, the area represents one of both limited and recent Investigation.

Also, those studies dealing with station image, except for one, have been almost wholly in relation to radio, not to television. The emphasis of this study dealt with the problem of television station image and a measurement technique for image appraisal. CHAPTER III

METHODOLOGY OF THE STUDY

Ths ••mantle differential vai used in this study to evaluate comparatively the Images of the three Columbus

commercial television stations from two points of view.

One scaling procedure was designed to measure a "holistic”

(station as a whole) image, and the other procedure was

set up to measure an "analytic” (station broken down into various concepts or significant features of its program­ ming) Image.

The "Holistic” Pilot Test

The concepts. Slnoe this test was designed to measure the stations' images from a total or wnole point of view, the three stations themselves served as con­ cepts to be evaluated by semantic differential soales or gradients.

The gradients. The gradients which were used to evaluate the stations were arrived at after surveying other research studies utilising the semantic differential teohnique (including Goldberg's radio study1) In order to get some idea of bipolar terms which had been used in

^See "Hew Way to Measure Stations*” op. clt.

31 32 past Investigations. Also, literature relative to the area of broadcasting was surveyed for purpose of noting important adjectival terns used in reference to television.

In the final analysis, the experimenter, combining inform­ ation gained from both the previously discussed methods, decided on a list of 20 gradients which it was estimated would be appropriate ror evaluating the television sta­ tions and might axso serve to point up differences, if any, existing between them. It was considered desirable to limit the number of gradients to 20. This number was deemed adequate for reliability of measurement yet small enough to provide for economical treatment of statistical data. These same gradients, as will be noted later, were used in the analytic pilot test. They are listed in

Table II.

TABLE II TWENTY GRADIENTS USED IN HOLISTIC AND ANALYTIC PILOT TESTS

pleasant-unpleasant big-small good-bad timely-stale rough-smooth colorful-colorless valuable-worthless friendly-unfriendly hones t-dlshonest serious-humorous strong-weak usual-unusual di f f ere n t - s ame Ideallstlc-practloal exciting-dull expensive-cheap authoritative-questionable humble-vain oommon-sophlsticated whol e s ome-unwho 1 e s ome 33

Construction of the test. Since from each respond­ ent taking the test a reaction was desired for each station

In the light of ail the gradients, provision for making a total of 60 reactions or respondent Judgments (3 stations times 20 gradients) was necessary. In order to make it easy to respond to the test and to economise on the length of the form, the entire test was limited to one page and set up as follows:

WLMf-C WTVN WBNS (4) (6) (10J pleasant 1 2 3 4 6 6 7 unpleasant • • • • > good 1 2 3 4 5 6 7 bad . . • rough 1 2 3 4 5 6 7 smooth • •

The order of the gradients was arranged according to Kendall's table of random numbers. The columnar positions of the stations were randomised on different copies of the test In order to prevent one station from having the advantage or disadvantage of any possible positional bias. Thus, on some copies of the test, WlW-C appeared first; on some copies, WTVN appeared first; and on otner copies, WBNS appeared first. A brief set of instructions on completing the test appeared at the top of the page. See Appendix B for a copy of the teat. 34

The '•Analytic*1 Pilot Teat

The eoneepta. Initially, It vaa the Intention of the experimenter to evaluate the atatIona from an analytic point of view in terms of any and all those factors which might affect a television station's image in the eyes of viewers. Thus, it was thought that possibly such factors as station ownership reputation, station network affili­ ation, reception of station signal, et cetera, might be included in the analytic test as concepts against which to pose gradients. However, in talking with the three local stations' managers, tneso men agreed that people are generally unaware of such factors as previously noted, that viewers watch stations with programs in mind, and that they evaluate a station largely, if not entirely, on the basis of its programming and program offerings. A decision was tnen made to limit the concepts included in the analytic test to features of a television station's programing.

Again, various station managers were consulted in order to determine their opinions regarding the sig­ nificant features of a television station's programing.

In addition to station managers, the experimenter also discussed this problem with Professors Harrison B. Summers and Riohard H. Mali of the radlo/teievialon area in the 35

Department of Speeen; Profeasor I. Keith Tyler, Director

of Radio-Television Education at Tne Ohio State University;

Professor Edgar Dale, Research Associate in The Ohio State

University's Bureau of Educational Research and Service;

and Professor Forest L. Whan of Kansas State College.

Finally, the experimenter asxed approximately 30 members

of a apeecn class during the Summer Charter of lWSb to

spend ten minutes listing all those television programming

features that they considered Important In evaluating a

television station. After getting these reactions to tne problem, the

experimenter complied all the "expert" opinions and student

opinions, and seleoted ten programming features to be used

as concepts In the analytic pilot test. These ten concepts

are listed In Table III. It was estimated that ten Im­ portant concepts would be adequate for a general evaluation

of programming features. Also, It was thought that

evaluation of these ten features of a television station's programming would reveal a meaningful Image and permit

the construction of valuable profiles of comnerclal

television stations. 36

TABLE III

TEN FEATURES OF A TELEVISION STATION’S PROGRAMMING USED AS CONCEPTS IN THE ANALYTIC PILOT TEST

News Sports Commercials Children's Programs Public (Community) Service Comedy*Variety Mys tery-Adventure-Western Music Feature Films (Movies) Program Personalities-Announcers

The gradients. As mentioned in a previous section

of this chapter, the gradients posed against the concepts

in the analytic test were the same 20 gradients used in

the holistic test (see page 32)• It was considered de­

sirable to use the same gradients in both tests for pur­ poses of making comparisons between the tests, especially

in the light of hypothesis number three on page ten.

Construction of the teat. Sacn of the ten concepts was paired with each of the 20 gradients, making a total of 200 items in the analytic pilot test. Since from each respondent taking the test a reaction was desired for each station in the light of each concept paired with each gradient, provision for making a total of 600 reactions or respondent judgments (3 stations times 10 concepts 37 times 20 gradients) was necessary. Thus, the analytic test was set up somewhat similarly to the holistic test in order to make it easy to respond to the test, to economize on the length of the form, and to provide for comparisons between the teats.

The 200 items (coneapt-gradient pairings) were arranged in random order according to Kendall's table of random numbers and set up on the pages as follows:

WTVN WBNS WLW-C (6) (10) (4)

1. NKWS strong 1 2 3 4 6 6 7 weak • • •

2. COMMERCIALS good 1 2 3 4 5 6 7 bad . 3. MUSIC timely 1 2 3 4 5 6 7 scale i .

Double-spaced, there were 20 items on each page making a total of ten pages. The columnar positions of the stations were rotated from one page to the next in order to prevent one station from having the advantage or disadvantage of any possible positional bias. Thus, on several pages, WLW-C appeared first; on other pages,

WTVN appeared first; and on still others, WBNS appeared first. The ten page test was stapled together along with a cover page of instructions on the front and an informa­ tional questionnaire on the back. See Appendix C for a copy of the test and questionnaire. 36

Administration of the Pilot Tests

All subjects used In this pilot test experimenta­ tion responded to both the holistic and analytic tests*

Instructions for both tests, which, of course, were very much alike, were read aloud to the subjects who marked the tests In groups* The holistic test was mar&ed first.

The time needed to mark both tests varied from 36 to 55 minutes•

While the experimenter read the directions aloud, the subjects read silently. Essentially, the subjects were asked to record in the columns to the right of each page that number for each item which to them best described each of the stations listed. Afterwards, the experimenter asked whether further clarification was needed. For the most part, none of the subjects needed additional explana­ tion of the manner in which their markings were to be made.

Sometimes questions were asked about the nature and use of the semantic differential technique. The experimenter answered such questions as briefly as possible, but no other explanations were given to subjects until after they had oompleted both tests.

The respondents for both pilot tests were secured from speech classes at The Ohio State University during the Fall Quarter of 1958. Obviously, since three Columbus 39 commercial television stations were being evaluated, it was considered desirable to use as subjects only those students who lived at home in Columbus or the Columbus metropolitan area and who had a television set in their home. It was usually possible to get at least three to six such students from a class and administer the tests to them at the same time during a regular class hour. A total of 40 students were used as respondents to these pilot tests. Certainly* students represented an atypical sample on which to evaluate the three television stations.

However* in view of the fact that these tests were pre­ liminary in nature - that is* their primary purpose was to find out what concepts and gradients were significantly discriminatory among stations for use in future tests - it was felt that the 40 respondents to the pilot tests represented an adequate and Justifiable sample.

Results

Table IV lists some of the basic characteristics of the sample that responded to the pilot tests.

Two of the gradients used in the pilot tests*

■ rough-smooth” and "conuon-sophisticated*" represented the less desirable polar term first* whereas all the other 18 gradients had the desirable polar term first

(for example* "good-bad*u "pleasant-unpleasant*" 40

TABLE IV A BREAKDOWN OP DESCRIPTIVE FACTORS OF THE FORTY RESPONDENTS TO THE PILOT TESTS

Sex Average age Educational level 26 male 20.3 years 7 freshman 14 female 13 sophomore 13 junior Religion 6 senior Marital status 27 ^roteatant 1 fifth year 35 single 11 Catholic 5 married 2 other

Occupation of father Number of children In family 9 professional 8 one ohlld 12 managers, officials, 16 two children or owners 9 three children 7 sales 4 four children 4 clerical 3 five or more 1 skilled 1 semi-skilled Where lived most of life 1 manual labor 38 urban 5 no data 2 rural

Economic status of family Station watched moat often 1 below average 10 WLW-C (channel 4) 24 average 11 WTVN (channel 6) 15 well-to-do 19 WBNS (channel 10)

Number hours watch television per week 18 none to five hours 7 six to ten hours 12 eleven to fifteen hours 2 sixteen to twenty hours 1 twenty-one hours or more 41

"valuable-worthless," et cetera). Thus, prior to begin­ ning any tabulation of data. It was necessary to reverse the responses to Items involving the gradients "rough- smooth1' and "oorvnon-sophlstlcated." For example, a 1 response to any item Involving "rough-smooth" became a 7 response; a 2 response was changed to a 6 response; et cetera. Naturally, a 4 response remained unchanged.

Methods of scoring the responses were compared in order to achieve economy in the scoring and analytic processes. Six rank-order correlations were computed between the ranks given groups of ten items by the statistical means of the responses and the ranks given those items by the proportion of total responses given by the subjects to the first two places on the semantic differential scale continuum. These groups of ten items were chosen randomly from the 200-item analytic test.

The resulting rho's (.87, .91, .74, .75, .95, .84) were considered to be high enough that the proportion of total responses given to the first two places on the continuum could be used to compare group station Images.

The holistic test. The first comparative analysis of stations was made with an F test. This F test, was computed on the proportion of responses given to the first two places on an item by each respondent for ail three stations on the entire holistic test. This F test, 42 which represented a two-way analysis of stations and gradients with 40 scores (40 respondents) In each cell, yielded a value of 35.06. With two and thirty-eight degrees of freedom, this value was significant at the one per cent level of confidence. The F test indicated that the three stations were given significantly dif­ ferent ratings on the holistic test.

F tests were then coiqputed on each of the 20 gradients malcing up the holistic test. Table V lists

TABLE V

f v a l u e s o f a l l t w e n t y g r a d i e n t s ON h o l i s t i c p i l o t t e s t

Gradient F Value Significance level pleasant-unpleasant 3 6 . 7 6 p < .01 good-bad 63.62 p < .01 rough-smooth .01 valuab1e-worthie ss 11.54 P < *01 hone st-dlshone s t 2.0 3 strong-weak 24.73 P < *01 different-same 4.86 exciting-dull 5.62 authoritative-questionable 8.42 P < *01 coBBon-sophlstlcated 7.56 P c .01 big-small 33 .93 p <. .01 timely-stale 11.52 P < .01 eoiorfui-ooioriess 16.48 p < .01 friendly-unfriendly 8 . 4 0 P < .01 serlous-humorous 2.37 usual-unusual 1.19 idealistic-practical 6.46 P < .01 expensive-cheap 38.08 P < .01 humble-vain 6.76 whole s cme-unwhole some .80 45 the resulting values of these tests and their significance

levels with two and thirty-eight degrees of freedom. From

this table it can be seen that 12 of the 20 gradients

differentiated among the stations at the one per cent

level of confidence.

Accordingly, t-tests were then computed in order

to determine what gradients discriminated among what pairs

of stations. Since there were three possible pairs of

stations (WLW-C and WTVN; WTVN and WBNS; and, WLW-C and

WBNS) and 20 gradients, a total of 60 t'a were computed.

Table VI lists the obtained t values and their significance

levels with 59 degrees of freedom for each pair of sta­

tions for each gradient.

The following is a summary of the results of these t-tests j WLW-C WTVN WLW-C between stations - and and and WTVN WBNS WBNS gradients 9 .01 level 11 15 1 gradients 9 .02 level 2 4 gradients 9 *05 level 1 2 1

Thus, it can be seen that a total of 14 gradients dif­ ferentiated at the .05 level of confidence or beyond

between stations WLW-C and WTVN; 15 gradients at the

•05 level or beyond between stations WTVN and WBNS; and only six gradients at the ,05 level or beyond between

stations WLW-C and WBNS. 44

TABLE VI

t VALUES OF ALL TWENTY GRADIENTS FOR Ea CH PAIR of s t a t i o n s on h o l i s t i c p i l o t t e s t

WLw-C WTVN WLW-C Gradient and and and WTVN WBNS WBNS

p leas ant -unp lea san t 5 *83* 6 . 66* 3 . 1 4 * good-bad 3*90* 5.47* 2 . 4 7 * * rough-smooth .07 •11 •06 valuable-wo rthle a b 3*03* 4.48* 1. 9 0 honest-dishonest 2*66** 1.52 •66 strong-weak 5 . 06# 5.70* 1.66 differ en t-aaste 2. 91* 2.27*** •22 axel ting-dull .91 2.80* 2 .69** au tho rl ta ti v e - qu e a tl o na ble 2.66** 3.69* 1.16 common-sophisticated 3*22* 3.40* •33 big-small 5.76* 7 .81* 1.44 timely-stale 3 . 2 7 * 3.64* 1 .11 colorful-eoloriesa 4 . 41* 4 . 7 0 * •54 f rl endly-unf rl endly 2.36*** 3 . 4 6 * 2 . 4 2 * * * a er lou a-humo roua 0 1.57 2 . 5 0 * * usual-unusual • 10 1.15 1.64 idealistic-practical 4.06* 2. 4 1 * * * 1.58 expensive-cheap 6 .82* 6 . 4 6 * .71 humble-vain •66 1.23 • 62 wholesome-unwholesome 1.32 3 . 2 6 * 2 . 7 4 * * * * value significant at •01 level of confidence ** * value significant at •02 level of confidence *** * value significant at •05 level of confidence

The analytic tort* Two F tests wore computed on the entire analytic teat* Aa in the analysis of the holistic pilot test, the proportion or reaponaea given to the flrat two pleoea on the aoale continuum of an

Item waa used. One F teat repreaented a two-way analysis of stations and concepts with 800 scores (20 gradients

times 40 respondents) In each cell* This test yielded

a value of 39.2o which was significant at the one per

cent level of confideuce with 2 and 18 degrees of free­

dom. The otner F test represented a two-way analysis of stations and gradients with 400 scores (10 concepts times 40 respondents) In each ceil. This test yielded

a value of 46.06 which was significant at the one per cent level of confidence with 2 and 38 degrees of free­

dom. These F tests Indicate that the three stations were given significantly different ratings on the analytic test.

Accordingly, t-teats were then computed In order to determine (1) what concepts discriminated among what pairs of stations and (2) what gradients discriminated among what pairs of stations. Some question was raised here over whether or not the concept and gradient values to be employed In the confutation of t values could be considered to be related. Therefore, 12 rank-order cor­ relations (six involving concept values and six Involving gradient values) were confuted between the ranks given one station based on the proportion of one and two re­ sponses for each Item and the ranks given another station based on the proportion of one and two responses for eaoh item. Twenty items were considered in eaoh of the six 46 confutations involving concept value a and ten itema were conaidered in the computations involving gradient valuea.

The reauXting rho'a (.70, *82, .68, ,20, .81, .25, .56,

•78, .30, .48, ,68, .66) were oonaldered to be high enough that the concept and gradient valuea received for eaoh of the three atations could be assumed to be related.

However, further Inspection of the diatributiona auggeated lack of homogeneity of variance. Since the presence of heterogeneity of variance will affect the t valuea, it waa necessary to apply a corrective measure to the t values received. In discussing the problem of heterogeneity of variance, Lindquist atateat

However, the heterogeneity must be quite extreme to be of any aorloua conaequenoe. While statistical teats of heterogeneity of variance are available. • . there will be relatively few situations in which any auch teat la required.

According to Lindquist, one common method of making allowance for heterogeneity of variance is to aet a higher “apparent* level of significance for the valuea received. Another common method used In dealing with this problem la to * . • .perform the t-teat in the uaual manner, but enter the table of t with one-half the number of degrees of freedom that would ordinarily

2 B. P. Lindquist, Design and Analysis of Experiment a (Houghton ElffiIn Company, 1955), p. 86. 47 3 be available." Edwards' method of maxing allowance for heterogeneity of variance was the one employed In this study.

Table VII lists the obtained t values and their significance levels, based on 4 1/2 degrees of freedom instead of 9# for eaoh pair of stations for each concept

TABLE VII

t VALUES OF ALL TEN CONCEPTS FOR EACH PAIR OF STATIONS ON ANALYTIC PILOT TEST

WLW-C WTVN WLW-C Concept and and and WTVN WBNS WBNS News 6 . 4 1 * 10.39* 6.15* Sports 6 . 4 6 * 8 . 4 6 * •60 Commercials 4.25* 5.64* 4.51* Children's Programs 3 . 1 5 * * 4 . 2 6 * 3 .86* Public (Conmunlty) Service 3 . 0 8 * * 10.27* 9.21* Comedy-Variety 5 .23* 7.17* 2.73*** Mystery-Adventure-Western 2.78*** 3.87* 6 .50* Music 5 . 8 6 * 7 . 0 4 * 4.61* Feature Films (Movies) 4 . 44* 7 . 0 7 * 6 . 00* Program Personalities- Announcers 5 .95* 7 . 9 1 * 7 . 26*

* = value significant at .01 level of confidence ** = value significant at .02 level of confidence *#* * value significant at .05 level of confidence

sAllen Edwards, Experimental Design in Psychological Research (Mew York: Rinehart and Company, l95oj,p. 17(5. 48

The following la a summary of the results of these t-tests: WDtf-C WTVN WLW-C between stations - and and and WTVN WBNS WBNS concepts W .01 level 7 10 8 concepts @ .02 level 2 concepts .05 level 1 1

Thus, It can be seen that a total of ten concepts dif­ ferentiated at the *U5 level of confidence or beyond between stations WIW-C and WTVN; all ten concepts at the .01 level between stations WTVN and WBNS; and nine concepts at the .05 ievex or beyond between stations

WLW-C and WBNS.

Table VIII lists the obtained t values and their significance levels, based on 9 1/2 degrees of freedom

Instead of 19, for each pair of stations for each gradient.

The following is a summary of the results of these t-tests: WLW-C WTVN WLW-C between stations - and and and WTVN WBNS WBNS gradients d *01 level 5 16 5 gradients 9 .02 level 5 6 gradients 0 .05 level 5 2 7

Thus, it can be seen that a total of 15 gradients dif­ ferentiated at the .05 level of confidence or beyond 49

TABLE VIII

t VALUES OP ALL TWENTY GRADIENTS FOR EACH PAIR ~ OP STATIONS ON ANALYTIC PILOT TEST

WLW-C WTVN WLW-C Gradient and and and WTVN WBNS WBNS

pleasantrunpleasant 3.37*** 7 . 5 9 * 3 . 6 9 * * good-bad 3.92** 6 . 8 4 * 2 . 9 7 * * * rough-smooth 2.76*** 5.28* 3.98** valuable -wor tJhle s a 3.48** 5.74* 3 . 1 7 * * * honest-dishonest a e fv^ft 6 . 5 7 * 4. 2 4 * * strong-weak 4.72* 6.28* 3 . 5 9 * * different-same 6.15* 5.29* 3.28*** exciting-dull 2.48 w w i r v r 3.93** authoritative-questionable 1.98 5.83* 2.85*** oosmon-sophlstlcated 3.17#** 7.45* 3. 1 7 * * * big-small 3. 6 8 * * 7 . 8 7 * 2.97*** timely-stale 3.37*** 8.93* 5.34* colorful-colorless 3 . 6 3 * * 9.69* 4.45* friendly-unfriendly 3.66** 4 . 50* 1.98 ser1ous-humo rous 7 . 5 2 * 6.72* 2.59 u sua 1-unu s ual 2.11 1.78 .20 Idealistic-practical 1.13 2.35 1.97 expens1ve-c he ap 5 . 4 8 * 8 . 1 3 * 3. 97** humble-vain 0 3 .18*** 4 .82* whole a ome -unwho le a ome 4.66* 8 . 2 5 * 3 . 0 1 * * *

• * value significant at •01 level of confidence «* s value significant at .02 level of confidence *** = value significant at •05 level of confidence

between stations WLW-C and WTVN; 18 gradients at tbs

•05 level or beyond between stations WTVN and WBNS; and

16 gradients at the .05 level or beyond between stations

WLW-C and WBNS. 50

The Fine! Teata

Construction of the final teata. It waa decided to uae in the holistic final test thoae gradients that discriminated moat among pa Ira of stations• Inspection of Table VI on page 44 reveala that the following gradients exercised the moat discriminating power in the holistic pilot teats pleasant-unpleasant, good-bad, valuable-worthless, strong-weak, common-sophisticated, big-small, timely-stale, colorful-colorless, and expensive-cheap. All nine of these gradients dis­ criminated among at least two pairs of stations at the one per cent level of confidence.

Inspection of Table VIII on page 49 reveala that the following gradients discriminated among at least one pair of stations at the one per cent level, one pair of stations at the two per cent level, and one pair of stations at the five per cent level in the analytic pilot testi pleasant-unpleasant, good-bad, rough-smooth, valuable-worthless, honest-dishonest, strong-weak, different-same, big-small, timely-stale, colorful- colorless, expensive-cheap, and whole some-unwhole some.

In terms of these arbitrarily set up standards for evaluating the discriminating power of gradients. 51

the following eight gradients were discriminating among

stations in both pilot tests *

pleasant-unpleasant big-small good-bad colorful-colorless valuable-worthless timely-stale strong-weak expensive-cheap Since these eight gradients were most discrimi­

nating among stations on both the holistic and analytic

pilot tests and since it was considered desirable to use

the same gradients in both the final testa* only these

eight gradients were retained for inclusion in the final

tests• Inspection of Table VII on page 47 indicates that

all ten of the concepts discriminated among at least two pairs of stations at the one per cent level of confidence.

In view of this* all ten of these concepts were Included

In the analytic final test.

The holistic final test was set up in the same

nmnner as the pilot teat. That ls# the test was limited

to one page* with a brief set of Instructions appearing

at the top of the page. The order of the gradients was

arranged according to Kendall's table of random numbers.

The columnar positions of the stations were randomized

on different copies of the test In order to minimize

the possibility of any positional bias. Since there

were eight items (gradients) on the test* eaoh respondent 52 would make a total of 24 judgments (3 stations times

8 gradients) In taking the test.

The analytic final test was set up In the same manner as the pilot test. Since each of ten concepts was paired with each of eight gradients, this test had a total of 80 Items. However, a total of 240 judgments would be made by each respondent to the test (3 stations times 80 conoept-gradlent pairings). The 80 Items were arranged In random order according to Kendall's table of random numbers. Double-spaced, there were a total of

20 Items on each page making a total of four pages. The columnar positions of the stations were changed from page to page In order to minimize the possibility of any positional bias.

The one page holistic final test, the five page analytic final test (which lnoluded one page of instruc­ tions), plus an Informational questionnaire, were stapled together in that order. See Appendix D for a copy of the tests and questionnaire.

Administration of the final tests. Respondents to the final tests were drawn from the population of parents of Columbus resident students enrolled in speech courses. The experimenter visited various classes and explained briefly the nature of the research. Then, those students who were Columbus area residents, living 63 with their parents and with a television set in their home, were asked whether they would enlist the cooperation of one of their parents in responding to the final tests.

Ordinarily, those students who said they were Columbus area residents agreed to ask one of their parents to answer the tests. A negligible number refused.

Since each set of tests Included directions, a questionnaire, and the holistic and analytic final tests, students were asked to have their parents read the directions before responding. The students were in­ structed to help their parents in no other way. After a student's parent completed the tests, the student was to return them to his instructor, who in turn saw that the experimenter received them. Host students returned their parents' completed tests within a week after re­ ceiving them. According to approximately 20 students who were asked, the time it took their parents to respond to both tests and to fill out the questionnaire, ranged from 12 to 20 minutes.

Of approximately 165 sets of tests handed out to students who agreed to enlist the cooperation of their parents, 116 were returned. Of this number, 12 had to be discarded because they were too incompletely filled out or else represented obvious distortions. The 54 experimenter then randomly discarded four of the remaining

104, leaving a total of one hundred to be used aa a sample.

Table IX 11 eta sane of the basic characteristics of the sample which responded to the final tests.

TAB IE IX

A BREAKDOWN OF DESCRIPTIVE FACTORS OF THE ONE HUNDRED RESPONDENTS TO THE FINAL TESTS

Sex Educational level 44 male 29 736 to 45) 55 no college 56 female 56 (46 to 50) 24 some college 55 (51 to 66) 25 college grad or more average age 4£.5

Where lived Marital status Religion most of life 03 married 67 Protestant 97 urban 7 widowed, 24 Catholic 5 rural separated, 9 other or divorced Occupation 4 professional Number of children In family 18 managers, officials, 2& one ckllS or owners 36 two children 9 sales 19 three children 7 clerical 12 four children 10 skilled 8 five or more 5 unskilled 55 housewife Station watched most often 12 miscellaneous 25 WLW-C (channel 4) Economic status of fkmUy 21 WTVN (channel 6) 56 WBNS (channel 10) 82 average 18 well-to-do Number hours watch television per week 12 none to five hours 18 six to ten hours 22 eleven to fifteen hours 17 sixteen to twenty hours 51 twenty-one hours or more CHAPTER IV

RESULTS

Stability of the Final Teats1 Sample

The final testa1 total sample of one hundred re­ spondents was divided Into two equal groups of 50 each*

Additionally, It was deemed desirable to equate the groups

In terms of "sex*1 and "station watched most" variables.

Thus, in each group there were 22 male and 26 female respondents. One group contained 12 respondents who watched WLW-C most, ten who watched WTVN most, and 28 who watched WBNS most. The other group contained 11 respondents who watched WLW-C most, 11 who watched WTVN most, and 26 who watched WBNS most. In evenly dividing the total sample on the basis of these two characteristics.

It was found that for the most part the two groups were well-equated In terms of most other characteristics (for example, age, occupation, religion, et cetera).

Six groups of elthe items each were chosen randomly from the analytic final test. Split-half reliabilities were computed between the means of the responses given to these items by each group of 50 respondents. The coefficients, corrected for length by the Spearman-Brown formula, are listed in Table X.

55 56

TABLE X

RELIABILITY COEFFICIENTS AND FIDUCIAL LIMITS OF THE FINAL TESTS' SAMPLE

Coefficient SIPIF confidence confidence

.85 .76 .91 • 95 .88 . 96 .89 .82 .94 •88 .81 .95 .92 .87 .95 • 93 .88 .96

Accordingly, the fiducial limits of these reliability coefficients were obtained by means of s' transformation scoresTable X also lists the one per cent limits (lower and upper) for each of the six coefficients. Therefore, if the final tests were ad­ ministered one hundred times to the same group of respondents and under like conditions, mean scores comparable to those received in this research could be expected 99 of the 100 times within the oonfldenoe limits indicated. These values seem to indicate a reasonably high degree of reliability or stability for the final tests' sample.

For a discussion of the prooedure for determining fiducial limits of correlation coefficients, see Allen L. Edwards, op . oit». pp. 126-51. 57

Reliability of the Final Testa

In order to establish some measure of the reliability of the final tests, two Kuder-Richardson correlations were computed - one for the holistic final test and one for the analytic final test* It was con­ sidered necessary to compute the correlations in t^rms of only one station for eaoh test since the manner in which respondents would answer the tests would be re­ lated from one station to another*

The derived Kuder-Richardson reliability coef­ ficient for the holistic test was *881; for the analytic test, *974* In order to determine the stability of these values, the fiducial limits of them were obtained by means of z 1 transformation scores* The one per cent limits of the holistic test coefficient of *881 were

.805 and *927. The one per cent limits of the analytic test coefficient of *974 were *959 and *985. These statistics suggest a fairly high degree of reliability for both the holistic and analytic final tests.

Validity of the Final Testa

As mentioned earlier in a discussion of validity in semantic differential measurement,^ "face validity"

o See page 16* 56 is usually considered in determing how well an Instrument meets this criterion. In considering the holistic end

analytic final tests, the careful selection of concepts

and gradients included in these tests plus the fact that

these tests do appear to be measuring what they are in­

tended to measure (station Image) seem to satisfy the

criterion of face validity.

There is additional evidence of the validity of

the final teats. It seems relatively safe to assume that

if a person professes to watch one particular station more than either of the other two, then his evaluation

of the stations made on the semantic differential items

should tend to favor the station he watches most. That

is, it would seem logical that a respondent's judgments

in the final tests would tend to be near the "desirable" polar term when considering the station he watches most,

and less often near that "desirable” term when considering

either of the other two stations. For purposes of de­

termining whether or not this was the case, the total

sample of one hundred respondents was divided on the

basis of the "station watched most" characteristic.

This information was readily available from the ques­

tionnaire data. 59

It was found that 56 respondents Indicated they watched WBNS most; 23 Indicated they watched WLW-C most; and 21 Indicated they watched WTVN most* Responses from the subjects of each group were tabulated. Table XI lists the mean values of each preference group for each station on each test. Actually, each value In this table Is a mean of means value since each one either represents a mean of eight gradient mean value (In the case of the holistic final test) or a mean of ten concept values (In the case of the analytic final test).

First, observe this table by considering each group's evaluations of all three stations. In other words, read the values horizontally. It may be noted that on both the holistic and analytic tests, the WBNS and WLW-C preference groups rated "their" respective stations In the most desirable light (the lower a mean value, the more desirable It Is since the "1" end of the seven point semantic differential scale was nearest the desirable polar term). The WTVN preference group considered WTVN and WBNS equally desirable on the holistic test but did not consider WTVN most desirable on the analytic teat• Thus, in five out of six situations

(eaoh of 3 stations evaluated on each of 2 tests) It was found that a group of respondents who professed 60

TABLE XI MEAN VALUES OP EACH PREFERENCE GROUP FOR EACH STATION ON EACH PINAL TEST5

Holistic Test Analytic Test (8 gradients) (10 concepts)

WLW-C WTVN WBNS WLW-C WTVN WBNS

Watch WBNS most group 2.9 3.6 2.2 3.0 3.5 2.7 (n«56)

Watch WLW-C most group 2.6 4.0 2.9 2.7 3.5 2.1 (n*23)

Watch WTVN most group 2.7 2.5 2*5 3.1 3.2 3.0 (n*21)

to watch a particular station most also tended to con­

sider that station as most or equally desirable to the

other two.

Looking at these data in yet another way gives

even greater strength to this notion of predictive

validity. Now observe the table by considering eaoh

station's evaluations by all three groups. In other words# read the values vertically. It may be noted

5See Appendix A for gradient and concept means from which these station mean values are derived. 61 that on both the holla tie and analytic tests, each station*a lowest mean value (or, moat desirable rating) waa given to It by its particular preference group. That la, the WBNS preference group gave to station WBNS the most desirable mean values that that station received on both teats. This alao was the case with stations WTVN and WLW-C.

The apparent marking of the tests by the respondents in an expected direction lends support to the notion that these final tests seem to satisfy a predictive validity criterion.

General Results

Table XII lists the mean responses of the entire sample, for each station, on individual items, on the holistic final test. Figure 1 gives a plotting of these item (or gradient) means. The resulting profiles give a clearer indication of the total sample* a image of each of the three stations with respect to the gradients used in this holistic approach evaluation. It may be noted that in no case do the lines cross. Station WBNS con­ sistently receives the most desirable rating, WLW-C the next most desirable, and WTVN the least desirable. There will be a discussion later in this chapter regarding which gradient values represent significant differences between the stations. 62

TABLE XII

MEAN RESPONSES ON THE HOLISTIC FINAL TEST

Gradient WBNS WLW-C WTVN pleasant-unpleas ant 2.2 2.6 3.1 good-bad 2.3 2.6 3.4 strong-weak 2.2 2.7 3.7 big-small 2.3 2.6 3.7 valuable-worthiess 2.6 2.6 3.5 timely-stale 2.6 2.7 3.3 oolorful-colorless 2.6 2.9 3.4 expensive-cheap 2.6 3.0 3.6 Mean of gradient means 2.4 2.7 3.5

WBNS WIW-C WTVN 1 pleasant-unpleasant. M | *\ \ I \ I good-bad . \ \ \ strong-weak • t I I big-small • i { \ {\ valuable-worthiesa • \ t I I timely-stale • i \ \ oolorful-colorlfi si • \ \ \ expensive-cheap •

FIGURE 1

PROFILES OF MEAN RESPONSES ON THE HOLISTIC FINAL TEST Table XIII lists the mean responses of the entire sample, for each station, on individual items, on the analytic final test. The individual item means are grouped by concept and the means of these means are given at the end of each concept grouping. The order of listing of items under a concept is constant through­ out the table. That is, the "pleasant-unpleasant* scale item is given first under each concept. Then the "good- bad" scale item, "strong-weak," %ig-small," "valuable- worthless," "timely-stale," "colorful-colorless," and finally "expensive-cheap.” This is exactly the same order in which the gradients were listed In the holistic test table (Table XII).

Each station's analytic image as conceived by the total sample may be seen more clearly in Figure 2, where the concept means for eaoh of the three stations were used to plot the profiles. In this figure, unlike the holistic test profiles' figure, it may be noted that in several instances the lines do cross. In general, the stations' analytic test Images are not as clearly defined or separated from one another as are the stations' holistic test images. However, even here, station WBNS consistently receives the most desirable rating on most concepts with WLW-C and WTVN following 64

TABI2 XIII

MEAN RESPONSES ON THE ANALYTIC FINAL TEST

News Sports No. WBNS WLW-C WTVN No. WBNSWLW-C WTVN 56. 2.3 2.4 3.1 80. 2.5 2.3 2.7 52. 2.1 2.2 2.9 1. 2.7 2.2 2.9 38. 2.0 2.2 3.1 18. 2.9 2.5 3.2 51. 2.1 2.3 3.2 35. 2.8 2.5 3.0 11. 2.0 2.2 3.1 2. 3.0 2.6 3.3 72. 2.1 2.2 2.8 57. 2.3 2.3 2.8 64. 2.4 2.6 3.2 10. 2.7 2.5 3.1 58. 2.6 2.7 3.4 32. 2.9 2.6 3.2 Mean 2.2 2.4 3,1 Mean 2.7 2.4 3.0

Commercials Children's Programs No. WBNS wixr-c WTVN No. WBNS WLW-C WTVN 42. 3.7 3.8 4.2 33. 3.0 3.4 3.2 20. 3.7 3.7 4.2 66. 3.1 3.3 3.5 67. 3.5 3.5 3.9 13. 3.2 3.6 3.4 75. 3.3 3.3 3.7 29. 3.2 3.5 3.5 8. 4.1 4.1 4.6 65. 3.0 3.3 3.3 46. 3.5 3.6 3.9 47. 3.2 3.5 3.6 41. 3.4 3.6 4.0 55. 3.2 3.4 3.6 54. 3.4 5.6 4.1 46. 3.4 3.7 5.8 Mean 3.6 3.6 4.1 Mean 3.2 3.5 3.5

Public (Community) Service Comedy-Variety No. WBNS WLW-C WTVN NO. WBNS wur-c WTVN 9. 2.7 3.1 3.4 68. 2.7 2.6 3.3 31. 2.4 2.7 3.2 3. 2.6 2.6 3.6 25. 2.6 3.0 5.4 30. 3.0 2.9 3.6 56. 2.6 2.9 3.4 4. 2.6 2.7 3.7 34. 2.4 2.7 5.1 17. 3.0 3.0 3.8 74. 2.5 2.7 3.3 26. 2.6 2.7 3.4 73. 3.0 5.2 5.6 28. 2.8 2.8 5.5 39. 2.9 3.1 5.6 76. 2.7 2.8 5.5 Mean 2.6 2.9 3.4 Mean 2.8 2.8 3.6 65

TABLE XIII (continued)

MEAN RESPONSES ON THE ANALYTIC FINAL TEST

Mya te ry-Adventure- Music Western

No. WBNS WLW-C WTVN NO. WBNS WLW-C WTVN 40. 2.7 2.7 2.8 62. 2.6 2.7 3.0 59. 2.7 2.9 2.9 21. 2.7 2.7 3.1 50. 3.1 3.2 3.1 44. 2.8 2.9 3.5 69. 2.8 3*0 2.9 27. 2.8 2.8 3.1 49. 3.2 3.2 3.1 43. 2.7 2.7 3.1 78. 2.9 3.1 3.1 37. 2.7 2.8 3.0 6. 2.5 2.8 2.7 79. 2.7 2.6 3.0 5. 3.0 3.0 2.9 63. 2*9 2.9 5.4 Mean 2.9 3.0 2.9 Mean 2.7 2.8 3.1

Feature Films Program Personalities- (Movies) Announcers

No. WBNS WLW-C WTVN No. WBNSWLW-C WTVN 22. 3.2 3.4 3.4 7. 2.3 2.6 3.2 12. 3.4 3.4 3.4 16. 2.6 2.8 3.6 19. 3.5 3.4 3,8 60* 2.7 2.9 3.5 70. 3.3 3.4 3.7 61. 2.7 2.8 3.6 53. 3.7 3.8 3.9 45. 2.6 2.9 3.6 14. 4.1 4.1 4.5 23. 2.9 3.1 3.4 71. 3.2 3.4 3.6 77. 2.7 2.8 3.6 24. 3.6 3.6 4.0 15. 2.9 3.0 3.7 Mean 3.5 3.6 3.8 Mean 2.7 2.9 3.5

WBNS WLW-C WTVN Mean of concept means 2.9 3.0 3.4

in that order* A later discussion will point up which concept values represent significant differences between the stations* WBNS WLW-C WTVN l yT----- 1 News

Sports

Comnerclala

Children*s Programs

Public (.Community) Service

Comedy-Variety

Mystery-Adventure- Western

Music

Feature Films (Movies)

Program Personalities* Announcers

FIGURE 2

PROFIIES OF CONCEPT MEAN RESPONSES ON THE ANALYTIC FINAL TEST 67

The signifloanee of the differences between the gradient means on the total holistic test and among pairs of stations (3 pairs) on this test, was computed by the use of the F test. Thus, four F values were calculated on this test. This same statistical technique was used to compute the differences between the concept means on the total analytic test and among pairs of stations on this test. Again, four P values were calculated.

Table XIV gives a summary of the derived F values on the total holistic and total analytic tests, and among pairs of stations on these tests. Inspection of the F values in this table makes it apparent that there are very significant differences between stations in both total tests, among all three pairs of stations in the holistic test, and among two pairs of stations in the analytlo test. The difference between WIW-C and

WBNS on the analytic test was not nearly so significant.

For purposes of further analyzing the results of the holistic and analytic final tests' data, that is, determining what differences existed between what sta­ tions on what gradients and concepts, the experimenter 4 deoided to rely on nonparametrlo statistical techniques.

*See Sidney Siegel, Honparametrlc Statistics (New York i Me Or aw Hill Book Company, I no., 1956). 68

TABLE XIV

SUMMARY OP P VALUES ON HOLISTIC AND ANALYTIC FINlL TESTS AND BETWEEN PAIRS OP STATIONS ON TOTAL TESTS

Holiatie Analytic P Value Sig. level P Value Sig. level

WLW-C and WTVN 91.20 p < .01* 18.70 p <.01**

WTVN art WBNS 110.69 p < .01* 28.94 p < .01**

WLW-C and WBNS 64.08 p < .01* 3.21 p<.10>.20*1

Total Test 91.26 p <.01*** 20.77 p < .01****

#sig. levels based on 1 and 7 degrees of freedom #*slg. levels based on 1 and 9 degrees of freedom #*#slg. level based on 2 and 14 dagrees of freedom -iHHHtsig. level based on 2 and 18 degrees of freedom

Non parametric statistics are very useful when It Is considered desirable to avoid the assuiq>tlon8 Inherent in other statistics - for example, the t-test's assump­ tions of normality of distribution and homogeneity of variance.

The holistic test. The Kolmogorov-Smirnov two- sample test*’ was found to be most useful in determining what gradients discriminated among what pairs of stations

See Siegel, o p . clt.j pp. 127-56. on the holistic test. This statistical measure tests

"whether two Independent samples have been drawn from the same population (or from populations with the same A distribution}." It is concerned with the agreement between two cumulative distributions. The D value, which results from the Kolmogorov-Smirnov technique, represents

the largest difference between one sample1s cumulative frequencies and the cumulative frequencies of the other sample.

A total of 24 Kolmogorov-Smirnov D values were computed for the three pairs of stations and eight gradients. Table XV lists the obtained D values and their significance levels for each pair of stations for each gradient.

It can be seen from these data that a total of six gradients differentiated at the .05 level of confl* dance or beyond between stations WLW-C and WTVN; all eight gradients at the .01 level or beyond between stations WTVN and WBNSj and, no gradients at even the

•05 level between stations WDf-C and WBNS.

These results may be more easily understood in the ligjit of the holistic test profiles plotted for

each of the stations on page 62. It may be noted that 70

TAB IX XV

KOLMOGOROV-SMIRNOV D VALUES OP ALL EIGHT GRADIENTS FOR EA'CH PAIR OP STATIONS ON HOLISTIC PINAL TEST

WLW-C WTVN WLW-C Gradient and and and WTVN WBNS WBNS

pleasant-unpleasant • 16 .33* .17 good-bad .27* .34* .16 strong-weak .37* • 44* • 18 big-small .34* .32* .15 valuable-worthless .20** .26* .10 timely-stale .24* •24* .07 o 00 colorful-colorless •15 •23* . expensive-cheap .26* .37* •13

* = value significant at .01 level of confidence ** - value significant at *05 level of confidence

the greatest distance between station profile lines lies between the WBNS and WTVN lines. This apparently is support for the significant differences existing between these stations on all eight gradients. Also, the smallest distance between station profile lines lies between the WBNS and WLW-C lines. This explains the lack of significant differences between these stations on any of the gradients.

If only those gradients whloh discriminated among at least two pairs of stations at the .01 level of confidence were considered, then it might be said that the following five gradients accounted for most of the differences in the holistic final testt good-bad, strong-weak, big-small, timely-stale, and expensive-cheap.

Using these gradients, a shorter (five-item) holistic test for measuring station "total image" might very well be constructed.

The analytic test. The Mann-Whitney U test for very small samples (eigit or less)^ was found to be most applicable in determining what concepts discriminated among what pairs of stations on the analytic test. This statistical measure, which is one of the most powerful of the nonparametric tests, is often used as an alterna­ tive to the parametric t-test when it is considered de­ sirable to avoid the ^-test's assumptions. The Mann-

Whltney U, like the Kolmogorov-Smlrnov technique previously discussed, may be used to test whether two

Independent groups have been drawn from the same population. In computing this statistical measure, scores from the two groups are combined and then ranked. The

U value is obtained by counting the number of scores in one group that precede each score in the second group. 72

Contrary to most statistical test values, the

lower a Mann-Whitney U value, the more significant it

is* This is due to the manner in which the U value is

computed. That is, if an almost equal number of scores

from each group precedes the other, then the smaller the

difference between the two groups, the larger the U value,

and the less significant this value Is. As the value

decreases, however, (Indicating more differences between

the groups) then its significance increases.

A total of 30 Mann-Whitney U values were computed

for the three pairs of stations and ten concepts.

Table XVI lists the obtained values and their signifi­ cance levels for each pair of stations for each concept.

The significance levels given in Siegel's Mann-Whitney U

value table** are designed for a one-tailed test. Inasmuch as there was no prediction made regarding the direction in which a difference should occur, a two-tailed test of

significance is appropriate. Therefore, the U table

values, which are significance levels for a one-tailed

test, were doubled.

It can be seen from the data In Table XVI that a

total of seven concepts discriminated at the .002 level of confidence or beyond between stations WIAT-C and WTVN;

8Ibld., p. 275. 73

TABLE XVI

MANN-WHITNEY U VALUES OF ALL TEN CONCEPTS FOR RiffH PAIR OF STATIONS ON ANALYTIC FINAL TEST

WLW-C WTVN WLW-C Concept and and and WTVN WBNS WBNS

News 0* 0# 16 Sports 0* 10*** Commercials 5** 5** 24 Children's Programs 30 4«* * — —

• » * M public (Community) Service v W W 0* 10*** Comedy-Variety 0* 0* 27 Mya tery-Adventure-Wea tern 25 24 18 Music 0* 0* 26 Feature Films (Movies) 19 IS 22 Program Personalltlea- Announcers 0* 0* 14

* s value significant at •001 level . of confidence ## * value significant at •002 level . of confidence *** s value significant at •02 level of confidence

eight conoepta at the .02 level or beyond between atat Iona

WTVN and WBNS) and only three oonCepta at the «02 level or beyond between atatIona WLW-C and WBNS.

Figure 2 on page 66* which ahowa the analytic teat profiles for each station, nay aid in understanding these reaulta. It may be noted that generally the greatest distance between profile lines Ilea between the WBNS and

WTVN lines. Thus, the largest number of significantly 74 different concepts exists among this pair of stations.

Since the smallest distance between profile lines usually falls between the WBNS and WLW-C lines, it only stands to reason that the fewest number of sig­ nificantly different concepts should be found among this pair of stations. Where the profile lines of all three stations are bunched up, especially in the case of the concepts "Mystery-Adventure-Western* and "Feature Films

(Movies,* no significant difference among any pair of stations was found. Obviously, then, these two concepts did not add much to the total test differences insofar as these three stations were concerned.

The Hypotheses

Information and analysis of data necessary for answering hypotheses one and three already have been presented in the earlier parts of this chapter. Inter­ pretation of these results in the more precise form of answers to the hypotheses and conclusions drawn, will be discussed in Chapter V. However, In order to pro­ vide a basis for investigating the seoond hypothesis originally set up in this research, it was necessary to go beyond the data provided by the hollstlo and analytic tests, Hie results of two comerolal rating services had to be seoured. 75

The Telepulse Report for the Columbus, Ohio metropolitan area, February, 1959 was one service 9 selected for this purpose. The February Pulse was chosen since it corresponded most closely with the time the holistic and analytic final tests were administered to the sample used In this study. It was considered desirable to compare the rating services1 findings with the holistic final test*s findings (instead of the analytic final test) since both these evaluations, although measuring somewhat different factors, can be considered

"general" In nature.

The first step In comparing data from these sources of television station evaluation required the conversion of the Telepulse information into the best expression of the number of television sets tuned Into each station during an average day over the course of a week. It was felt that such an average Index of station viewing audience would be most easily comparable with the three stations1 total holistic test mean values or scores.

The Telepulae Report, in one of its early pages, gives share-of-audience figures, expressed in percentages of homes-uslng-televlslon, for each station, for each day, and for each of the following three time periods -

®"A Telepulse Report" (Columbus, Ohio Metropolitan Area, February 1969)y The Pulse, Inc., New Yoric. 76

7 am to 12 noon; 12 noon to 6 pm; and 6 pm to 12 midnight.

The Columbus metropolitan area is credited with 180,400

television homes. Using these items of information, the

experimenter calculated the average number of homes apt

to be tuned in to each station during an average day.

These figures were then expressed in a percentage share

of homes-uslng-television figure for each station.

The ARB report for February, 1959 was also

selected.^0 This service gives an average share of

homes-using-televislon figure, expressed In percentage

form, for each station from sign-on to sign-off for an

average day.

Table XVII lists the Pulse and ARB sise-of-audience

figures and the total holistic test mean score for each

station. It may be noted from this table that station

WBNS Is number one among the stations with respect to slae-of-audlence figures based on both of the reports.

In terms of the holistic test results, WENS is also con­

sidered number one. However, whereas WLW-C is the second most desirable station in terms of the holistic test

results, it actually is third when considered in the

•The Columbus Television Audience,* An ARB Metropolitan Area Report, February 1959, American Research Bureau, Inc., Beltavllle, Md. 77

TABLE XVII

SIZE OF AUDIENCE FIGURES AND TOTAL HOLISTIC TEST VALUES FOR ALL THREE STATIONS

Percentage Station Home a-Using-Te1evlaIon Holistic test mean values Pulse ARB

WBNS SB .4 S7.0 2.4

WLW-C SO.6 31.2 2.7

WTVN SO.9 32*6 3.5

light of probable aize-of-audlence figures. WTVN is third among the stations when evaluated on the basis of the holistic test re suita, but second when sise-of- audlence is considered. CHAPTER V

SUMMARY AND CONCLUSIONS

Summary

The methodology used and the results obtained In this study suty be considered in their relation to (1) the eonoiuslons that might be drawn from them and (2) their lmpllcations for further study.

litis study concerned the problem of station image in broadcasting. The iiajor purposes were to determine whether the semantic differential technique oould be adapted constructively to the problem of measuring station

Image and to evaluate comparatively the images of tne three Columbus oofluterolal television stations.

The problem of station image was considered from two points of view, holistic and analytic. An atteinpt was made to measure each station's image from a holistic

(total lflwge or station as a whole) point of view and an analytic (station broken down Into various types of pro­ gramming) point of view. A holistlo pilot test, which had each station evaluated In terms of 20 gradients; and an analytic pilot test, which had each station evaluated in terms of eaoh of ten concepts paired with the sane

20 gradients, were constructed and administered to 40

78 79

Columbus resident students with television seta In their

bones. The purpose of these pilot testa was to determine

what gradients and oonoapt-gradient pairings vere dis­

criminating among the three stations.

A hoxlstlo final test, oonqpoaed of the eight

gradients found to be nost discriminating among the

stations, and an analytic final test, including these

same eight gradients paired with each of the original

ten conoepts, vere constructed and administered to one

hundred parents of Columbua resident students. It is

In the light of results obtained from these final tests

that the conclusions of this study are drawn.

Any conclusions drawn about the data presented In

this study should not be extended Indiscriminately beyond

types of conditions that prevailed In this study. It is

hoped that these conclusions may have implications for

other situations, however, such extended Interpretations

should be made only with caution.

Conclusions

The first hypothesis, that there Is no difference between the Images of the three Columbus coasnerclal tele­

vision stations as amasured by the semantic differential

technique, seems to have been rejected In view of the

holistic and analytic final teats' findings. 00 Thus, It nay bs concluded that -

1* The Columbus oomaerolal television atationa do reflect different Images. In terms of a total Image evaluation, WBNS la considoped slgnifloantly different

(Ln a desirable direction) from either WLW-C or WTVN, while WLW-C la eonaldered slgnifloantly different (in a desirable dlreotion) from WTVN* In terms of an analp- tle image evaluation, WBNS la eonaldered slgnifloantly different (In a deairable dlreotion) from WTVN and dif­ ferent, although not slgnifloantly (in a desirable dlreotion) from WLW-C* WLW-C also is eonaldered slgnlfi- oantly different (In a desirable dlreotion) from WTVN*

The aeoond hypothesis, that there Is no dlfferenoe between a station's image as measured by the semantic differential teohnlque and Its oommerolai ratings as determined by rating servloes, seems to have been rejeoted in view of a comparison of total hollstle teat values and

Pulse and ARB slse-of-audlenoe figures for eaoh station.

Thus, It may be oonoiuded that

2* The viewer Images of the three Columbua com­ mercial television stations are not all consistent with the stations' slse-of-audlenoe ratings* WBNS, whloh Is considered In terms of Its hoiistio Image to be slgnifl­ oantly different (In a desirable dlreotion) from either

WLW-C and WTVN, also has the largest slse-of-audlenoe 81 figures. However* although WLV-C la eonaldered In terms of lta holistic Image to be slgnifloantly different (In a desirable directIon) from WTVW* It actually has smaller slse-of-audlenoe figures than does WTVN.

The third hypotnesls* that there Is no dLfferenoe between a station's Image when measured from a holistic point of view and when measured from an analytic point of view* seems to have been substantiated In one sense but rejeoted In another. Differences between stations were shown to exist In the same direction when measured by the two methods* but to be more dlstinot when measured by the holistic test. Thus* It may be concluded that

3. Analysis of the three Columous oonnerolal television stations' Inmges from holistic and analytic points of view do produce confirming results Insofar as the dlreotion of differences Is concerned* however* dif­ ferences between tns stations are more distinct when measured from the holistic point, of view. Apparently*

if people tend to like a station generally* then they also seem to like most of its program offerings.

Discussion. In addition to the ooneiusions relative to the hypotheses already discussed* it might be interesting to poln* up other aspects of this study.

This dlseusslon makes no attempt to state such factors 82 with statistical significance,, taut rather prefers to

Mention them from a standpoint of Interest.

As a measurement teohnlque to be used In broad­ casting research, the semantic differential vas found to be successful in quantifying station image. For the most part, the semantic differentia^ teohnlque was not found to be a difficult one to understand by the respondents.

In this respect, however, it must be remembered that respondents to the pilot tests were college students and respondents to the final tests were parents of oo liege students who very probably represented a higher than average educational level. In every Instance but one, mean evaluations for all the stations on gradients in the hollstlo test and oonoept-gradlent pairings in the analytic teat were on the favorable end of the semantic differential soale.

If a value of 4.0 could be considered a mid-point on tbs continuum, inspection of Tables XII and XIII on pages 68 and 64-66 shows that only with respeot to the oonoept

•Comseroials" (4.1) on WTVM did a mean value exceed 4.0,

The two most probable explanations for suoh a predominance of favorable ratings are (1) people tend to avoid orltloal extremes in their evaluations, and (8) television, generally, nas met with popular acceptance among the viewers• 85 It was found that genepally# paopla who professed to watoh a particular station more than either of the other two tended to evaluate that station most desirably* or* at least not as undesirably as viewers who admitted to watohlng eltner of the other two stations most* This behavior on the part of respondents seems understandable since it might be justifiably assumed that watohlng a particular station most presupposes a Hiring for that station*

In noting the analytic final teat results more closely* the only ooneept on whloh WLW-C was given a more desirable rating than WBNa was "Sports*" It may be oonjeotured that such rating behavior was due in part to the WLW-C'a HBC network affiliation and the general thinicing that HBC is the strongest network in terms of sports programming* Also* the only concept on which WTVH was given a more desirable rating than

WLW-C was "Mystery-Adventure-Weatern*” This rating* which was equally as favorable as that given to WBMS* could be considered eaplloable in the light of WTVH's

ABU networn affiliation* In general* the ABC network is considered to have more top programs of this type than either of the other networks* especially HBC* 84

Implications for Parther Study

This study considered the use of the semantic differential teohnlque In the analysis of the Images

of three Columbus comerelal television stations.

Further study In the following areas might provide useful Infonsatlon on the problem of station Image.

1. Since this research eonaldered a total of

20 gradients initially# further study with a different group of gradients (for exasple# "interesting-boring,"

"light-heavy#" et cater*) might be conducted under this

experimental situation.

2* Further study with a different group of oon-

eepts# pernaps concepts dealing with television station variables other than programming, might be conducted under this experimental situation.

S. Consideration of Individual programs within

the general program types used as concepts might reveal why stations are strong or weak in various programming

categories.

4. A replication of the same experimental situa­

tion and variables but with a larger# more structured sample would very probably yield more meaningful results

regarding station Image. 6* Additional research of the ease nature but using different stations in another marfcet area night

reveal similarities or differences in the variables from one station to another*

6* Having respondents react to the same stlauil but with reference to an "ideal” station as well as

specific stations in a market night provlds a basis for interesting oonparlsons. APPENDIX A

PREFERENCE GROUP MEAN RESPONSES ON THE HOLISTIC AND ANALYTIC FINAL TESTS

66 87

PREFERENCE GROUP GRADIENT 1GAN3 OH THE HOLISTIC FIMiL TEST

Watch WIW-C Wateh V t n Watch WBNS ■oat group ■oat group aoat group Qradlout (»■ 23) (n? 2 1 ) (»* 5 6 ) WDT-C WTVN WBNS WUT-C WTVN WBNS WI*-C WTVN WBNS ploaoant-uuploasant 2.3 2.5 2.7 2.4 2.0 2.2 2.9 3.3 1.9 good-bad 2.4 4.1 2.9 2.4 2.L 2.3 2.8 3.4 2.0 strong-i»»ak 2.4 4.0 2.7 3.X 2.7 2.2 2.7 3.9 1.9 blg-aaall 2.3 4.0 2.6 2.8 2.6 2.7 2.7 4.0 2.0 valuablo-srorthloaa 2.9 3.8 3.5 2.4 2.2 2.1 2.9 3.8 2.4 tlaaly-atala 2.7 4.0 3.3 2.1 2.2 2*i| 2.9 3.5 2.4 oolarful-oolorlaoa 2.7 4.1 3.0 3.0 2.2 2.8 3.0 3.5 2.4

•xponoiro-ohoap 2.7 4.3 2.8 3.2 3.5 3.1 3.0 3.8 2.3

Total 20.6 31.8 23.5 21 Ji 19.8 19.8 22.9 29.2 17.3 CM o O'. Moan 2.6 • . 2.7 2.5 2.5 2.9 3.6 2.2 88

PREFERENCE QROUP CONCEPT HEINS ON THE ANALTTIC FINAL TEST

watch h u m : Watch WTVN Hatch WBNS — st group — st groqp — st group Concept 0 * 2 3 ) (a* 21) ( ■ = 5 6 ) i a - j m r m s lUOTffn is® RU-U "WfVN"WBH5

Nows 2*0 3.2 2.7 2.5 2.7 2.1 2.U 3.2 2.0

Sports 2.3 3.1 3.0 2.5 2.5 2.9 2.6 3.2 2.5

Co— oroials 3.6 lt.6 3.9 3.7 3.9 3.5 3.6 4.0 3.U

Children's Progra— 3.U 3.U 3.2 3.6 3.U 3.U 3.5 3.5 3.0 Public (Co— unity) Sorvioo 2.6 3.3 2.6 3.1 3.3 3.2 3.0 3.U 2.1*

Co— dy-Varicty 2.U 3.6 3.0 3.2 3.3 3.1 2.7 3*6 2.6 Myatory-Adrcnturc- Wostorn 2.7 3.1 3.0 3.2 2.7 3.1 3.0 3.0 2.7

Hallo 2.5 3.0 2.9 3.1 3.2 3.0 2.8 3.2 2.6 1 1 ■ >)3.U 3.7 3.8 3.2 3.3 3.1 3.6 JuO 3.5 Progi— Forooaalitlos- Anneuaoors 2*U 3.6 3.0 3.1 3 . h 3.0 3.0 3.5 2.5

Total 27.3 3U.6 31.1 31.2 31.7 30.k 30.2 3U.6 27.2

M o w 2.7 3.5 3.1 3.1 3.2 3.0 3.0 3.5 2.7 APPENDIX B

HOLISTIC PILOT TEST

69 90

DIRECTIONS

loch of the following lt«iu !• composed of two word* and a ernle of seven lumbers bat wa n than. Look at S O W word* in esich pair. In the columns to the right of the page, place tha number that, TO TOO. boat describes cjich of tha three television stations listad. Do this for aneh of tha twenty pairs of words. Tor example, if a pair of words was "fair-unfair," consider the ambers to mean so net h leg like Ihie: 1 - tbit fair 2 - quite fair 3 - slightly fair I - neither fair nor unfair; in the middle 5 - slightly unfair o - quite unfair 7 - reiy unfair Some of tha items may seam to hare little relation to the television stations, but think about each one and give your answer. Judge each itan by itself. Try not to base your judgments on your feelings toward the particular word pairs; but rather, relate the words to your feelings about each of the three television stations. Please ear* the Items as fast as you con. Record your first impression. Pe sure to mark a umber for each of the thjee television stations for each item. wens; w pleasant 2 3 4 3 e unpleasant ......

good 2 3 4 5 6 b a d ...... rough 2 3 4 * 6 smooth ...... valuable 2 3 4 5 6 worthless ...... honest 2 3 4 3 6 d i s h o n e s t ......

stxos^ 2 3 4 5 6 w e a k ......

different 2 3 4 5 6 ' s a m e ......

exciting ? 3 4 5 6 d u l l ...... authoritative 2 3 4 5 6 noamon 2 3 4 5 6

big 2 3 4 5 6 timely 2 3 4 5 6 stale ......

oolorful 2 3 4 5 6 odorless ...... friendly 2 3 4 5 6 urfrlandly ...... serious 2 3 4 5 6 humorous ......

yeutl 2 3 4 5 6 idealistic 2 3 4 5 6 expensive 2 3 4 5 6 o h e n p ...... ,

wholeness 2 3 4 5 6 bumble 2 3 4 5 6 APPENDIX C

ANALYTIC PILOT TEST AND QUESTIONNAIRE

91 HHECTIOSS

Too. ax* about to help me discover the manning of carts in voids. Each of the items on the following pages is cooposed of a significant feature of a television station’s programing and a pair of opposite terms with a scale of seven amber* betwesi tba tens. Please mark the items as fast as you can. Bacozd youx first inpress.oa. Eb not stop tc thirk it over. Do not go back to change a mark,

this is tba vcqt you do it. Here is an It at sat up just like the iteras on tha test:

WDM8*3 F B U & J B 6 related 1 2 3 4 ^67 tenae

Thirfrlng of %naen*s orogzana," record that number which you think represents the piopar distance between ’relaxed* and •tense' for each of tha three stations listed at the right of tha page. Consider the lumbers to nean something like this: 1 - tbit relaxed 2 - quite relaxed 3 - sliAtlj relaxed 4 - neither relaxed nor tense: no feeling either way; is the middle 5 - slightly tense b - quite tense 7 - veiy t<

It is quite likely that you will he.se different feelings regamlii^ how 'relaxed* or bow *tenee* "women's programs" are on said: of the three stations. Eras, it is possible to record different umbers for each of the three stations. 411 of the i t « s in the test will be dose in this Banner. Obviously, the progressing ferture being cnl-iated and the pair of opposite texxx change fro* itec to item. Sane of the items may seec to have little relation to the television stations, hut think about each one and give your answers. Judge each item by itself. try not to base your judgments on your feelings toward the particular word pairs; tat rather. relate the void pairs to your fbelings about the type of programing under qjestioc. Be cure to made a number for each of the three television stations for each ites, B D K S E B t Vfoik fast. Mkc* ycrax nubers clecr. And, be sure to stu'fy the order of sjmeumnce of the stations at the top of each page since the order will chrtge f n » page to page. 93

S * « I

g 8 • •JJtatjtaaiiiittt rr\ \r\ oo o\ nHHHHHHHHHq h m cn *t if\ u> an cr» fj o n 56* 21. saggS serious 1 2 3 4 5 6 7 fcujcrouc . , 22. SFQiBS cannon 1 2 3 4 5 6 7 sophisticated 23-H i S U C (COMJjIIti) valuable 1 234567 vorthiees 24. lUluIhbPS PHXrRtfS eomon 1 2 3 4 5 6 7 sophisticated 25. colorful 1234567 colorless . . . . , 2 6 . Oa M B C IJLS friendly 1 2 3 4 5 6 7 unfriendly . ,

27. serious 1 2 3 4 5 6 7 biaoxous ...... 23. HOBUM PPSOaaiTIlS-iWUtaCgaS cocron 1234^6 7 sophisticated . . .

29. g a w r r i i s good 1 2 3 4 5 6 7 w ...... 30. h w b l e 1 2 3 4 5 6 7 vain ...... 31- HSJJ^ authoritative 1 2 3 4 5 6 7 questionable...... 32. H U P H S H U 6 DCVJi) oonmrm 1 2 3 4 5 6 7 sophisticated . . 33- MTaaj^dUAPflfillHMflBSTKai expensive 1 2 3 4 5 6 7 c^eap . . . 34. SSL pleasant 1 2 3 4 5 6 7 unpleasant ...... 35. m a o m r m s fronts) ccicrtui 1234567 colorless . . . ?6 .iruflH valuable 1 2 3 4 5 6 7 worthless...... 37* lH IIJ/ntt*S friendly 1 2 3 4 3 6 7 unfriendly . . . . 38. fmUitfc THH5 tejfleaj expensive 1 7 3 4 5 6 7 <&sip . . . . 39. TUfflPg T O M terns) valuable 1 2 3 4 5 6 7worthless . . . 40. HBLIC (aSMMlTT) SXflflCZ owon 1 2 3 4 5 6 7 sophisticated 41. ti»ely 1234567 s t a l e ...... £2. WJifciA-TAtiliJ1 idealistic 1 234567 practical ...... 43. MJSIC pleasant 1 2 3 4 5 6 7 urjlecsart ...... 44. IXjJJia TO>C (tpvm) hocect 12345 S 7 di (bare st ...... 45. OQ^CT-TiPxfci' I pleesart 1 2 3 4 5 6 7 unpleasant ...... 46. miunE rmc Ctcms) tig 1 2345 67 suaii......

47. I B g tauable 1 2 3 4 5 6 7 ‘v e i n ...... 46. mamas nut frmss) unmi 1234567 uzusuai...... 49. CCyTTW’— friendly 1 2 3 4 5 6 7 unfriendly ...... 50. SFCBTS strong 1234567...... 51- different 1 2 3 4 5 6 7 **ne ...... 5*. H 3BUC fcemaniT) s g n a tig 1234567 s c a n ...... 53. MI8imaiPreW0»4tBTgW fri«dly 1 2 3 4 5 6 7 unfriendly . . . . 54. mSJRt TJUfS OCffXlS) antboritatiT* 1 2 5 4 5 6 7 questionable , . KIHLIC (OCMiJSIfr) SZPTia huabl el 234 5 6 7 rain...... 56. HJBLIC (COfCTITT) SBfflCS colorful 1 2 3 4 5 6 7 oclorlees . . . . 57. O ttM a O ilS hudfcle 1 2 3 4 5 6 7 rain......

58. 0PH

s 06

O

\o

i£ \

V 0

Li'\ C\ in ir\

W in M

M

t/1 tn

M m !

« 3 ^ ^ ^ 58 # & £ R P £ £ £ f£ H ££ 97

li

■ft 96

O

in \£> +H lT\ iTs uv

p-l \D lf> VO VO H 4n in tr\ ir\

r-< HM

O

4 H in vi rt H aS rt (AH ficS 2 S f* fH M H H rH 99

P

%

81 1 81

H H H H H H H H H H H H H H ...... L 5 5 v t z i i**®* sBZTim 8ffuS M i i m a s 8« w r a s g H *o5i

...... a i q w r ^ s o n * £ 9 a f t. z I M f f V S m$Tl ...... i p ® £ 9 5 r t z i m sBantawCTr-'SgiirargosgiJ wmxxs. mLn

...... £ 9 * p £ z -[ 7^3X0^ WHI 3H0STU 'ttt

...... tu>p £ 9 =, fr t z 1 3& 7V ? » * K V e X H J S » O T 5 m H 3 *lVl 'm. M. m '«»- m w (ic) 161. phxsuh p s a o m m Bs-iRJXiCTie good 1 2 3 4 5 6 7 M...... 162. pbd ® am pffisciALirns ^uworocgffi idealistic 12^45^7 prectiwl . . . 163. ODMEDT—VAHm tdiolesoae 1 2 3 4 5 ^ 7 umiioleeoee...... 164. IUSIC valuable 12 3 4 5 6 7 worthless ...... 16&. BUptUm oleasast 1 2 3 4 5 6 7 unpleasant ...... 166. 1C8IC strong 12343 6 7 ...... 167. tCBIC itoleaoM 1234 367 urwholesome...... 168. CDKBCLEiS earcitiiw 1 2 3 4 5 6 7 ill...... 169. 0CHDK»Y4EX2FY good 1 2 ^ 4 5 6 7 bod...... 170. MgTBHT^VflH.'UHL.VlSTgg bustle 12-34^67 vH in...... 171. » 8 honest 1 2 3 4 5 6 7 dishonest...... 172. IXAICJ1E n u c Ocms) olensant 1234367 uiq>lee*«t...... 173. QMIX'TlhM'I 4lff erect 1 2 34367 e a s e ...... 174- ZKX2UK FlRCaULini^ABKCFCISS valuable 1234367 worthless .... 175. MBIC enthoritative 1 2 3 4^67 quest io a h le ...... 176. w v p big 1234? 6 7 a n i l ...... 377. pRxeiH piBscMiLm^jaocarrK rough 1 2 3 4 5 6 7 saooth ...... 178. HOC ldtoleeoce 12 3 4 3 6 7 urodjolesoae...... 179. HOgUM HJSOM.ITrSS-dMUBCiafe authoritative 1 2 3 4 5 6 7 questionable 160. SHJXn*S7 B0GBA6 ideelirtic 12 3 4 5 6 7 practical ...... TOT 102

©

ITS ua U\ ua

CM fN| \D ■H IT. If

m M M

CO o •H 103

PleaM fill out the factual questloTmaire belov, Tour iw * Is needed only to correlate responses made at a future d-te. All info met ion given will be held in strictest confidence.

1. Hone ______2. Age ______3. Sex ______4* Marital statue: hfcrried _____ S'ngle _____ Divorced _____ Religious ■preference: Protestant, Catholic ______Other 6. Mucational level: College 1 2 3 4 ( year) 7. Number of ch ild re n in your Our.ily _ _ _ _ _ 8. Occupation of your father ______9. Would you soy that the economic status of your family is: Below average ______Average ______Woll-to-do _____ Vac 1 thy _____ 10. Where hrvo you liv e d lo s t of your l i f e : Urban _ _ _ _ _ Vfeich c i t i e s ! ___ _ Rjral _____ 11. Vdiich Columbus television station do you watch most often (chect only one): W W *5_ 12. Hov many hours do you watch television tsnch week; 0 - 9 ______6-1C ______1 1 -lf ______1 6 - 2 0 ______21 o r mo re ______13. Vfcat are your three favorite nrogmcie on television (indicate specific orograo*)i

14. R»nk the folloving three television stations according to your attitude towards than. Rink the station for vdiich you have the nost favorable attitude, lj the station for which you hnve t h e 1 ecst f e w in D ie ’"ttitude, 3.; and the remaining station, 2. VLW-C ____ V T V N ______w w s ____

Be sure you have signed your name on the front of aach of the two forr* you have previously filled out. Thank you. APPENDIX D

HOLISTIC AND ANALYTIC PINAL TESTS AND QUESTIONNAIRE

104 105

To nr. TV-lf o e t t t t c t i s

Lech o f the f ol Ic-'ing items is oociposed of l’i'l ir.Kf5 •'bout f ch *'f the th re e te le v is io n st>‘tto r s . Fle .se mcrk the itw s s.s fast as yew cur.. flecoxd your firrd ijsircreion.

HElZlFnr: Be sure t C IttfiW ■j, mi"-her for ecrh of the three telisviriior. st’ t tc75fi fo r arch ite r w

■*5 rle a sn n t 2 ■i G, 6 7 good 2 3 4 ✓ 6 7 b o d ...... stro rg 2 3 4 6 7 im u le big 2 p 4 G 7 scnll ......

C •7 vp livable 2 3 4 * 6 w o r t h l e s s ...... tim ely 2 3 4 5 C 7 stele ......

7 C co lo rfu l 2 4 ✓ 6 7 c o lo rle ss ...... expensive 2 3 4 5 6 7 ► To— rv-2f ITFECZICrS

Tot ax* stout to help ice discover the weeing of certrin -jcnis. lech of the it«rs on the following rentes is co^iosed of c sigr.ificant feature of a television station's proi^xwaring end a neir of opposite terns vitt a scale or K e r e n cunben between the teics.

?lease n&jfc the iteas as feet »s t o : can. Becoid your first ir.xession. Do rot step to think: it over. Do not go beck to change s .nasc. This ie the w v you do it. Her* is **n iter ret ur „-oot like tl.c iter.* on the test:

IS’hE J . W . VCMES*S ESC&ttMg relaxed 1 2 ? A H b 7 tense I . . . I . . . I. . . I

Thinking of "wxiec'e urograms,*1 record that number -f let you think represents the proper distance between "relaxed" and "tense" for each of the three station* listed «t the right cf the page. * Conkli n the* mtbere to ne^r. sons th ing like t h i s : 1 - very reTaxed 2 - quite relaxed 3 - rligfatly relaxed 4 - reither relaxed nor tense; no feeling either wey; in the niddle £ - slightly tense 0 - quite ter re 7 - very tense It is Qpite likely that you v ill have different feelings retarding hov "relaxed" or how "terte" "vwienH programs" are on each of the three stations. Thus, it is possible to record different numbers for e»ci of the three stations. JL31 of the ite&s in this survey will be done ir this aaror. The progmci irg feature bain? evaluated =nd the ncfr of opposite tern* ’-dll usu-lly ch.'TO'-p* frcr it** to iteir. Judge each item by itself. Try rot to base your judgnenbr on ycur feelings toward the pFrticulc.r void pairs, hut rather, relate the vcrd 'sairs to your feeling* a tout the typt of progrsinricg under question, ^e sure to ‘Aid; r Tuber for sect of the three television stations for erch it sc.

HM2EZP4 Work fcrt. IJake vout hum era clear. And, be sure to rtudy tie order 0.' sfpieararco of the station* nt tKe ten of each pye since tha order vill charge fro:.- page to page. 106 107

ft t m C ■8 4n 3 FM »;* r - C S r ,0 a VO ft r

r«! u v j • I) K \J? IA V IB 3 1 ■*r 3 • t tr * i£ \ r~* 1^ • w ? * 4*tt CM 8 • VO v o v O r • 3 ft •aO r- « M * •c r* XI (T'S i n u v VO •a • s 6 8 t 't •rl vo S' r-- t, Oi M 'vt M vo c m r CM VD r- VL> v o CM ■3 CM VO IN vo cn trv p-t 8 a * vO U v p p. 1r\ CN cm 5 1 P“N CN H 1 a PH Hft pH •a f I | 1 T* O 0 pH H ft f 8 y d W ri

o g | Si £ g ri & £ 8 8 0 8 0 * * • • * * • • pH M m^ U > VO r-. QJ o\ CM a p H 108

*' tT'

o M m cc t 9 k. H ■4* s I £ 4> H ■d c“<4, S 8 ±> r - II M i n ? r - 1 rHc r - vO O VO O r-» M VO lA v O Ui -g l VO f-^ U \ tr % v-r U i '4 CN ** v£> -i C M V O •3 (M T* l £ \ irv CN ■g 7 ^ n*\ . a CN u> ■3 N & U-) on r* "3- CN r - *— vj- r-. ir> cn V D r-« cm vo 1 *d 8 & CN M 5 U V un 9 *3 if" »*4 t> « CN I ► i? fc p« M 0 Jf £ PN B on B cm i S' S*. E o K CN H N ii s CN t ? £> #• -% >1 ft 8 ► I A? « 7 « u> si 4 .a* •-t I A3 i % d t (*■ B PI a >i o u o o i ^ i'.' H *H a § »| R fi § 1 tu 8 fc! 8 H« & B&s & p; liiii

< M M & CN iv « a* ft r\ ft « & £ ft

5^ W'"" Pi"

«< $ A*pH ■e S

■ * t •. • * W\ «, •*, Mff to 4 J

r- v o IX , U V c n

1 c n U ' H CM "T ' O tfv CM •vj* H o \ 11 > t \

«H C n f

cn £ n « »—i PH (T\ | Hi CM 0 +*1 H ■d *< ft u I w o r* i' M 8 M t—* 1 ? di +> 1 g o o *-« M l M a all ii • » cn IT\ VO ^ $ £ <3- &\ c\ av i' * 110

ST'

8 »4 c-*r 8 r- vXJ LiA

fn

o « ^ ^ H l ft ^ fi

^ a a & ^ f£ fi p ?t tCv v/. p v ‘ Ill

Flense fill cut the factual quart! jehjui* lelov. ^h*t infottu't ion is. in order in describe ^enoral1;' the ft*yiplr on vhicb this survey ’nr taken.

1. ±i’r ______2. Sex ______1. Marital stntnsi liwried ______Divorced, Widowed, Separated _____

4. Religious preference! Pretext.-’nt _____ inthnV.: _ _ _ _ _ ^thrr 5. Ednc-tioncil level 1 8th ;r ds or ler- ____ School _ _ College 1 2 3 4 6. Fuuber of children in yrrir f.-^aily______7. Occupation______8. Ubuld you ccy thnt the eoonooic »t t.to of your fn.nily i»! Below avrrv,e _____ Average ______Well-to-do _____ Wealthy ______9. Where have you 1 iwed oost of your lif* » Urban _____ Which cities) ______

fbra.1 _ _ _ 10, Which O olushis te le v is io n s ta tio n do you witch ...o*t of^tr (check only one) 1

'W W 155______11, How ;nny iinur* do you vntch te l cvision ascii week! 0 - 5 ______6-10 ______11-15______lt>*20______21 o r _ -a cre______12, ’,/hc,t are your three favorite pjMjmie on television (indicate specific proarrnsie) |

Nor.e (opt ioncl)

THA1K YOU VERT MUCK! BIBLIOGRAPHY u s

bibliography

Bogart, Leo. The Age of Television. New York! Frederlok Unger Publishing Company, 1956.

Broadcasting. LVI, January 19, 1969.

Broadcasting Yearbook 1958. Broadcaating Publications, Inc., Washington, D. ti., 1958.

"The Columbus Television Audience. 11 An ARB Metropolitan Area Report, February 1959, American Research Bureau, Inc., Beltsvllle, Md.

•Das Moines in Depth." Highlights of Media Study by Central Survey, Inc. for KRHT and KKNT-TV, Central Surveys, Inc., Shenandoah, Iowa.

"Different." An Evaluation of Some Qaalltatlve Dif­ ferences Between Radio Stations Baaed on a Study by Motivation Analysis, Inc., 1957, CBS Radio, Mew York.

Edwards, Allen L. Experimental Design in Psychological Research. New Yorki Rinehart ana Company, 1 9 & 0 .

Head, Sidney. Broadcasting In America. Boston! Houghton Mifflin Coup any, 1&&6.

Lindquist, E. F. Design and Analysis of Experiment s. Houghton Mifflin company, ls#S5.

MoMlllln, John E. "Commercial Commentary," Broadcasting. LVI, January 5, 1959, p. 6 .

"New Vay to Measure Stations." Sponsor. XII, November 25, 1957, pp. 42-45.

Norman, Varren T. "Stability Characteristics of the Semantic Differential," Teohnlcal Report No. 19, The Role of Language in Behavior, University of Minnesota.

Oagood, C. E. "The Nature and Measurement of Meaning," Psychological Bulletin. XL, May 1952, pp. 197-257. 114

BIBLIOGRAPHY (Continued)

Osgood, C., G. Suoi, and P. Tannenbaum. The Measurement of Meaning. Urbanes University of Illinois Press, TCsTI

"The People Talk Back to RadioA Motivational Research Study of the Houston Radio Audience, 1959, Institute for Motivational Research, Inc., Croton-on-Hudson, New York*

"Personality Profile of a Radio Station." A Study Made of Station WWDC, Washington, D. C., July 1958, Pulse, Inc., lew York.

"Radio Today - New Discoveries About the Constant Com­ panion of the American People." A Survey In Major Markets Conducted Alfred Polite Research, Inc., November 1954, Alfred Polite Research, Ino., New York. "Radio Today - New Discoveries About the Constant Com­ panion of the American People In the Intensive listening Area of WTIC - Hartford, Connecticut, November 1955, Alfred Polite Research, Inc., New York.

"Radio*s 5 Years of Great Change - 1955/1958." A Qiali- tative Survey of WHDH, Boston, Pulse, Inc., New York.

Schramm, Wilbur* Responslbi llty In Mass Commnication. New Yorkt Harper and Brothers, iWP# Seldes, Gilbert. The Public Arts. New Yorkt Simon and Schuster, 1956. Siegel, Sidney. Nonnarametrio Statistics. New Yoxki McGraw Hill Book doaqpany. Inc., 19KT7 Slepmann, Charles. Radio. Television and Society. New Yorkt Oxford University tress, \ 050.

Steinberg, Charles S. The Mkss Communicators. New Yorkt Harper and Brothers7 1988.

"A Telepulse Report •* Columbus, Ohio Metropolitan Area, February 1959, The Pulse, Ino., New York. Warren, H. C. Dictionary of Psychology. Bostons Houghton Mifflin tton^any, i9ls4 • 115

AUTOBIOGRAPHY

I, Franklin David Sabah, was born In Steubenville,

Ohio, February 20, 1954* I received my secondary educa­ tion at Steubenville High School from which I graduated in JUne, 1951* In September, 1952,I enrolled in The Ohio

State University and received the degree Bachelor of Arts

In June, 1965, and the degree Master of Arts In June,

1956 • During the final quarter of my mas ter *s program

I served as a Teaching Assistant in the Department of

Speech. After working a year in coonerclal broadcasting,

I returned to The Ohio State University and began my doctoral program in June, 1957. I have been employed as a Teaching Assistant in the Department of Speech while completing the requirements for the degree Doctor of

Philosophy.