THE EFFECTIVERESS OF DIFFERENT TYPES OF LETTERS OF REQUEST TO COLLEGE STUDERTS Thain far flu Dow” of Ed. D. MECHIGAN STATE UNIVERSITY John M. Fohr 1959 This is to certify that the thesis entitled THE EFFECTIVENESS OF DIFFERENT TYPES OF LETTERS OF REQUEST TO COLLEGE STUDENTS presented by John M. Fohr has been accepted towards fulfillment of the requirements for . ‘ [answzbzwn ‘- - 1, _£d_ol_3__degree law 6LT 1.8% Major professor Date—Emmew LIBR A R Y Michigan Stan Unrvcrsny THE EFFECTIVENESS OF DIFFERENT TYPES OF LETTERS OF REQUEST TO COLLEGE STUDENTS By John M} thr A DISSERTATION submitted to the School for Advanced Graduate Studies of Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF EDUCATION Department of Foundations of Education Higher Education 1959 John M. Fohr Candidate for the degree of Doctor of Education Data of Examination: Thursday, November 19, 1959, 3:00 P.M. Room 6A Berkey Hall Dissertation: The Effectiveness of Different Types of Letters of Request to College Students Outline of Studies 8 Major area -- Higher Education Minor areas -- Sociology and Anthropology, Guidance and Counseling Biographical Items: Birthdate -- Decenber 29, 1916. New England, North Dakota Undergraduate Studies -- Marquette University, B.S. Milwaukee, Wisconsin -- 1939 Graduate Studies -- Marquette University, Milwaukee, Wisconsin -- l9h7-19h9 Michigan State College, M.A. East Lansing, Michigan -- 1952 Michigan State University East Lansing, Michig -- 1952-1959 Experience: United States Department of Agriculture, Resettlement Administration, Milwaukee, Wisconsin, Technical and Publicity Writer, 1937-1939 United States Army Air Force, Bombardier, Navigation, Meteorology Instructor, and Technical Writer, 19hl-l9 American Appraisal Company, Milwaukee, Wisconsin, Assistant Contract Manager, l9h6-19h7 Spencerian Business College, Mflnaukee, Wisconsin, Instructor of ‘ Business and Journalism Michigan State University, East Lansing, Michigan, Instructor of Journalism and Business Writing, 191:9 to the present. ACKNWIEDGMENTS The writer wishes to express his sincere appreciation to the chairman of his guidance committee, Dr. Paul L. Dressel, for his willing assistance and guidance in carrying forward this study. He also greatly appreciates the interest and cooperation of Dr. Walter F. Johnson and Dr. Wilbur B. Brookover, the other two members of the writer's guidance committee. Special gratitude is expressed to Dr. Clyde W. Wilkinson, for his assistance in formulating the study and providing the inspiration to undertake this type of research. Sincere thanks is also due to Dr. Frank E. Ryerson, for his constant interest and helpful suggestions during the course of the study. - A great debt of gratitude is owed to Dr. E. L. Marietta, Dr. Marjorie Hunsinger, Dr. Charles H. Proctor, and Dr. Paul J. Deutschmann, for their cooperation and support. To the writer's wife, Betty Jo, and sons, John and Bruce, a special thank-you for their patience, understanding, and encouragement . iii THE EFFECTIVENESS OF DIFFERENT TYPES OF LETTERS OF REQUEST TO COLLEGE STUDENTS By John M. Fohr AN ABSTRACT Submitted to the School for Advanced Graduate Studies of Michigan State University'in partial fulfillment of the requirements for the degree of DOCTOR or EDUCATION Department of Foundations of Education Higher Education 1959 63M 1W ABSTRACT The Problem The purposes of this study were: 1) to determine the action in~ duced and reactions caused by the (a) emotional, (b) rational, and (c) combined emotional-rational approaches in persuasive letters of request to Michigan State University seniors; 2) to ascertain the effects of the use of (a) reader benefit appeal, and (b) group benefit appeal in the letters of this study. Methods and Procedures The sample consisted of six groups of 20h randomly selected seniors. Each group was mailed one of six different letters; each included one of two appeals and one of three approaches. The letters requested the students to arrange appointments for interviews with the writer. The personal benefits appeal letters offered to provide the reader with personal help in writing letters of application; the group benefits letters emphasized the help the readers could give the writer in assist» ing other members of the senior class to write letters of application. Eightyhthree students responded to the letters. In eliciting the reactions of the respondents to the appeals and approaches of the letters they received, a “disguised" interview technique was used, in which incidental and comparative reference to the letters of request was interwoven with the discussions of letters of application. No mention of the writer's personal purpose was made. A similar but shortened type of interview was employed in telephone interviews with 72 non-respondent students who were contacted. Principal Findings Action. At the .05 level of prdbability, no single letter, appeal, or approach was more effective than any of the other letters, appeals, or approaches. A significant difference was found in the higher action responses of 1) males as compared with females, and 2) upper lower-lower lower socio-economic class as compared with other socio-economic classes. Reactions. A large majority of the respondent and telephone inter- viewees had favorable reactions to the letters. The most significant factor in creating these favorable reactions was the personal help which the readers believed was available from a specialist in business letter writing. Regardless of the appeal and approach used, most of the readers viewed the letters in essentially the same way-~as an offer of personal help. The emotional approach is definitely the least favored of the three approaches. However, there is insufficient difference between the re- actions to the rational and the emotional-rational approaches to justify a statement that one of these two approaches is favored over the other. vi The criterion groups who had the most marked favorable reactions to the letters 'and indicated the greatest interest in obtaining help were: 1) students in the upper lower-lower lower socio-economic class, 2) married students, 3) males, and h) students in the 2.00-2.19 grade-4 point class (grade-point based on h.0 system). j The highest percentage of, negative reactions to the letters came from the interviewees who had had training in business letter writing from the writer. The most favorable attitudes to the letters were expressed by the readers who had had no training in business letter writing. vii TABLE OF CONTENTS CHAPTER Page ‘ I mmTIONOO0.000000°000000009000000OOOOOOOOOOOOOOOOOOOOO The PrOblemOOOOOOOOOOo00.0000000...OOOOOOOOOOOOOOOOOOO... Definition of Terms...................................... B8810 HymthOSj-BOOOOOOOOOOO00000.000000000000000000000000 Plan Of the study-00000OO00.00000000.00.00...00.00.00.000. cacaoxna to II WOFTHELITMTUIEOOO000000.00.0000000000000000000000 10 Literature OHAEffeCt Studies.................o........... 10 Literature on Appealseeeeeeeeeeeeeeeoeeeeeeeeeoeeeoeoeeoe 16 Literature on ApprOaChGSeeeoeeeeeeeeeeeeeoeeoeeeeeeeeoeee 21 III PROCEDURE AND TECHNIQUE OF THE STUDY 27 Initiation Of the SbUdY-eeoeooeeeeeoooeeeooeeeooeeeeeeeoe 27 Preparation Of the letterSooeoeeeeeeemoeoeeeooeeeeoeoeeee 27 Pre"TeSt Of the Letters...........a......o..........ooo.. 33 Sampling Procedure....................o.............o.... 36 Mailing syStem.and Appointment SCthUlingeeeoeoeeeeeeeeee hl Preparation of Interview schedules....................... h2 Procedure in Conducting Interviews......................} hh Reliability and validity Tests of the Interview Schedules h6 Coding of the Responses and Classification Data.......... 50 IV ANALESIS OF ACTION EFFECTIVENESS OF LETTERS................ 55 ACtiQn.AnalyaiseeeeeeeeeeoeeeoeeeeeOooooeooeeeeoeeeeeeeee SS Criterion Group Analysis................................. 57 Percentage Analysis of Criterion Groups.................. 58 Analysis or NO Replies....o....................o......... 61 Summary....o...oo........o...........o....o...........o.. 63 ‘ V ANALESIS OF REACTION RESPONSES TO LETTER TYPES AND MIT-CACHE BY mm CmSmICATIONS.OOOCOOOOOCOOCOOOOOOO 65 Reconciliation of the Use of the Chiquuare Statistic.... 65 Letter 613331fication Analysis............a.............. 68 Percentage Analysis of Reaction Responses to.Appeals and ApproaChBSeeeeeeeeeeeeeeeeeeeeeeeoeeeeoeoeeeeeeeeeeees 68 viii TABLE OF CONTENTS 4 Continued CHAPTER Page Use of Noanest Group to Check Validity of Reaction Responses to TeSt Lettemeeeeoeeeeeeoeeeeoeeeeeooeeeee 79 summal'yeoeeeeeeeeeoeeeeeeeeeeeeeeeeeeeeeeooeeeoeeeeoeoee 83 v1 ANALYSIS OF REACTION RESPONSES OF CRITERION CLASSIFICATIONS GROUPSOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 85 methOdS 0f Analy'SiSeeoeeesoeeooooeeeoeeeoeeoooeooeooeeee 85 Significant Differences of Reactions by Interviewees.... 87 Percentage Analysis of Cross Tabulations of Respondent‘ ReaCtionseeoe‘eeeeoeeeeeeeooeeeeeeeceeeeeeeeeeoeooeeoee 88 Percentage Analysis of Cross Tabulations of Telephone Interviewee Reactions................................. 96 811mm.eeeeeeeooeeeoeeeeeeeooeoeeeoeeeeeeoooeeoeeooeeee 97 VII SUMMARY, CONCLUSIONS, SUGGESTIONS FOR FURTHER RESEARCH AND IMPLICATIONS FOR THE TEACHING OF BUSINESS IETTER WITmGOO0.0.0.000...O...0..00.0.0...OOOOOOOOOOCOOOOOOOOOO 1m Smumary........ooo.u..."Hue”...unuuu.”sue... 100 PrinCipal Findingseoeeeeoeeeeeeeeeeoeeeeeeoooeeeeeeeeeee 102 conClUSionSeeeeeoeeeeoeeeeeeeeeeeoeeeoeeeeeeeeeeeeeeeeee 108 Suggestions for Further Research and Implications for Teaching Business Letter Writing...................... 110 BIBLIOGRAPHY..................................................... 112 APPENDIX Ar-Test Letters and Card Enclosure Used with Letters.... 117 APPENDIX B--Respondent and Telephone Interviewee Schedule Forms.. 12h APPENDII C-FMaster Code Sheets Used in Coding Data............... 135 Forms Used to Check Coding Of Responses to Open- Ended queStionS 0f SChedUleSeeeee'eeooseoeeeooooeoee 1h]. Coded Responses to Opeannded Questions.............. 1h? LIST OF TABLES TABLE Page I Flesch Readability Index Scores of Letters................ 31 II Number and Percentage of Total Action Responses to utters and Letter Typesaccesses-econ.oesoeeeoeeeeeeseee 56 III Percentage of Negative Responses, in Various Tabulations, to Emotional Approach Letters by the Two Types of InterVieWGSSoeooeeeeooeeooeeeoeoeeeeeoeoeoeoeeeeeeeeooee 7S CHAPTER I INTRODUCTION Two issues which frequently confront the teacher of business letter writing in teaching the writing of persuasive letters of request are: l) the relative emphasis which should be given to the use of emotional and rational appeal, and 2) the stress that should be given to reader- benefit or “you-attitude“ viewpoint.' These two issues are the topics of this study. I Teachers of business letter writing talk about various persuasive teChniques and appeals without having any convincing evidence to support their recommendations. This was the consensus of a panel of prominent teachers of business letter writing, who said: “Mbst of the time we teach, as if we knew for sure, many principles which in reality we are only assuming to be true.“ (60, p. 7) This remark is followed by a comment made by Manning, whoconcluded: Many of these principles are the product of no more than thoughtful introspection on the part of the authors and close observations of what to all appearances marks the difference between success and failure of a letter as it actually performs in business . . . it appears that no material contributions may be expected_until actual testing conditions for the COpy of a letter can be set up and the tests carried out. (No, p. 127) A. 9;; Problem Statement Egan; problem. The purpose of this study is twofold: (l) to determine the action induced and reactions caused by the use of _ (a) emotional , (b) rational, and (c) combined emotional-rational approaches in persuasive letters of request to Michigan State University seniors; (2) to ascertain the effects of the use of (a) reader benefit appeal, and (b) group benefit appealin the letters of this study. ELMO 2!. £29. 9.9.24 £93. 3.132 aiLdIo During the past fifteen years. one of the most frequently mentioned areas of needed research in business letter writing is effect analysis. This need was emphasized by a Research Committee of the American Business Writing Association which indicated that the biggest opportunity for research in business letter writing was in effect analysis (6, p. 16). Anrner, (7, n.p), Lesser, (37, p. 25), and Wilkinson, (60, p. 7) three higuy regarded specialists in business comnications, i also give high priority to making effects studies. Despite the importance attached to effects studies, only a few studies, limited almost entirely to action studies, have been made by . people in business letter writing. Manning and Wilkinson found that most tests have been made on sales letters, and that some studies have been done on collection letters (141, p. 228). (he fairly comprehensive affects study of personalization in collection letters was done by Allgeier (h, p. 16). Another study of a specific type of letter is Boddy's (12, pp. 72-92) action study of letters of inquiry. However, soddy3s study, as well as Allgeier's, did not involve an analysis of effects of reaction on actiong(see-pp. ll-12 of this study). Review of many other studies indicates that results most frequently show letter A producing more action than letter B without any systematic account of the ways in which the two letters are the same and in which they are different. Twenty years ago, Boyd (16, p. 5) first called attention to the need for controlled experiments which include other factors besides action results. The past and present emphasis on action studies stems primarily from the accepted principle that action is the main criterion of success (6, p. 16). The investigator agrees that if a letter obtains an action response from a reader, the main objective of the reader has been achieved; however, action doesn't of itself explain reaction, nor does it present positive proof that the selected approach or appeal was the factor of success. (6, p. 16). Also, an action response usually provides no measure of the goodwill reactions of the readers. An approach or appeal could induce action which miglt be accompanied by strong feelings of favorable or unfavorable identification with the writer or his organization. The same could be true of inaction. Because of the writer's adherence to the hypothesis that an action response to a commication invariably involves some type of reaction” (51, pp. 1-5), this study has been designed to determine not only how many of a particular sample of readers responded to two appeals and three approaches of a persuasive request, but also, by directly contacting the readers, ascertain who, what, why, and how the, responses were made. No study involving the preceding procedures and objectives was found in the review of many research studies in business writing. There is virtually complete accord that reader benefits should be given main stress. (33, pp. 39-140) Most authorities in business letter writing agree that copy in a persuasive letter of request should be neither exclusively emotional nor entirely rational, but should be a combination of both (hl, p. 213), although one former president of the American Business Writing Association admits that there is no verifiable research evidence to support such views (9). Tucker similarly remarks: We have in our area a great deal of original observation, some critical analysis, a few hypotheses, and almost no verifi- cation of the value of adaptation and other such vaguely defined characteristics. (57, p. 12) Business writing teachers certainly should have more objective knowledge ofthe effectiveness of the approaches and appeals used in persuasive letters of request. m ind limitations 9;; the s_t_u_dy. This study deals only with persuasive letters of request sent to a random sample of 201; Michigan State University Seniors during the University's 1958 fall term and first half of the 1959 winter term. I The accessibility of this group overcame one of the major difficul- ties business letter writing researchers have had in making effect studies. The specific situation, contained in the letters, involved a request that the readers come to the investigator's office to discuss letters of application. The study is limited to the effects of six different letter types, each of which employed one of the two appeals (personal benefit or group benefit), combined with one of three approaches (emotional, rational, or emotional-rational) . One type letter was mailed, at specified intervals, to one of six groups of seniors in the sample. Each student received only one of the six letter types. Each of the six groups of students represents a random sample, and it is assumed that, the responses of a group to a particular letter is typical of that of another group receiving that same letter. As is discussed in Chapter III on "Methods," in order to limit the study to the variables being investigated, all of the six letters were set up in the same format, organized according to the same psychological sequence, and written at virtually the same readability and human interest level. The latter involved the use of the Flesch Readability Index (see pp.30-32 of this study). Business letter writing specialists axiomatically acknowledge that a rational approach need not be less readable than an emotional approach (141, p. M6). On the other hand, communications experts hold that a personal benefits. appeal is more interesting and consequently usually makes for greater readability than what is termed, in this study, a group appeal (52. pp- u-8)- The typical business letter writing expert would normally consider letters which stress the personal benefits appeal as more interesting and readable than letters which subordinate this appeal. However, in using the Flesch Index as a control device, it was necessary to make adjustments in the length of sentences, personal pronouns, and words referring to human relationships in order to obtain the same level of readability and human interest in all of the letters. Little difficulty was experienced in adapting length of sentences and number of syllables per word. However, some adjustments were necessary which might suggest an I'unnatural'i' use of personal words and sentences, and humanwrelations words in thegroup benefit letters. Only through these adjustments was it possible to retain the same reading and human interest levels in the letters and conform with requirements for control of all the variables except those which are being tested. The experienced business letter writer might question the length and necessity of using as many subsidiary appeals as are used in the test letters. .A lesser number of subsidiary appeals or merely an invi- tation to obtain professional help might be all.that would be necessary to get action. This, however, would limit, for experimental purposes, 'the potential range of reaction to only a few aspects of content stimuli. To encompass as wide a range as possible to fit the anticipated categories to be analyzed in the study, a variety of elements were included in the test letters. B. Definitions‘gf Terms Persuasive letter of request. As used in this investigation, the persuasive letter of request involved a request for action in a situation in which the benefit or merit in taking action was not directly or immediately apparent to the reader. Business writers believe that there is an initial tendency on the part of a reader to react negatively to this type of request (lb, p. 140). As a consequence, all the letters are psychologically structured so that the explicit request for action is not made until most of the reasons or benefits for the reader's compliance are established. (hl, pp. lO6-ll3). Appgal. Throughout this study the term “appeal“ should be interpreted to mean the main stimulus or incentive used to influence the reader to take action. The two main appeals used in the letters of this study are 1) personal benefit appeal and 2) group benefit appeal. Personal benefit appeal. This appeal emphasizes the benefit to the reader in taking the action suggested. An offer of help from the writer is the motivating incentive. This is the appeal used in the three Type I letters of this study. 9393p benefit appeal. is emphasized in the letters of this study, the group benefit appeal stresses the benefit to the groupjof which the reader is a member in taking the action suggested. The motivating incentive is the help the reader can give to others--particular1y his peer group at the University. This appeal was used in the three Type II letters of this study. Approach. This is the angle of attack used in presenting the appeals. Three approaches were used in setting up the letters: 1) emotional (used in letter types I-A and Ill-A; 2) rational (used in types I-B and II~B3 3) emotional-rational (used in types I—C and II-C). Emotional approach. The principal emphasis in this approach is directed at the reader's emotions rather than his intellect. Rationalapproach. This approach is used when main stress is given to an intellectual, logical, or rational plan of attack. Emotional-rational approach . This is a combined emotional-orational plan of attack. C. Basic Hypotheses The following are the investigator‘s hypotheses: l) the most effective approach uses a combined emotional-vrational approach; 2) the personal benefit appeal will produce more action and more favorable reactions than the group appeal; 3) the appeal factor will be more sign nificant in formulating reader decision and attitude than the approach factor. D. Plan pf Study This study is divided into seven chapters. Chapter I deals with the background of the problem, a statement of the problem, importance of the problem, the limitations and scope of the study, definitions of terns and basic hypotheses. Chapter II is a review of pertinent literature in business writing and communications research in related areas concerning the use of emotional, rational approaches and personal and group benefits appeals. Chapter III deals with the procedures and techniques used in the study, including composition of the letters, use of the Flesch Readability Index, critical review of letters by experts, pilot study of the letters, drawing the sample, mailing system used, interview schedules and techniques used, coding, and machine tabulation. Chapter IV is an analysis of action results induced by the appeals and approaches. Chapter V is an analysis of reaction responses to the appeals and approaches by letter classifications . Chapter VI is an analysis of reaction responses of the criterion classification groups . Chapter VII presents the summary, conclusions, and some implications for further research and the teaching of business letter writing. CHAPTER II REVIEW or LITERATURE A comprehensive review of available published and unpublished literature in business writing revealed that much has been written about the need for effects studies in business letter writing, but little has been done in making such studies. No published action-reaction effects studies of persuasive letters of request were located. while there are few studies which relate to effects analysis on letters, either in business writing or related areas of persuasive com- munications, there is a considerable amount of empirical data and a number of experimental tests and studies on the types of appeals and approaches employed in this study. This research material has been useful in provid» ing ideas, hypotheses, and suggestions in interpreting the results of this study. In consequence, most of this chapter is concentrated on a review of literature pertaining to appeals and approaches. I. Literature gr; Effect Studies m—m studies pyteachers if: business $93222 writing. Most studies by teachers of business letter writing in both published and un- published media deal mainly with the techniques of writing and the methods and materials of teaching. Often "testing" of effectiveness consists of making a selection of letters on the basis of professional lO standards of effectiveness or the judgment of a single man or panel rather than upon scientifically controlled experimentation (61, p. 105). An examination of writings in theses, monographs, and articles on business letter writing, written between l930—l9hl (59, pp. 10-18) reveals only one study which was related to an effects analysis of a particular letter type. This was a controlled action study of letters of inquiry by Boddy (12, pp. 72-92). The purpose of the investigation was to determine the action-inducing effectiveness of various factors in letters of request sent to a random sample of h32 business firms. The letters used were non-sales, unsolicited, personal inquiries request- ing information. One of the letters, termed the base-line letter, was deliberately designed to include violations of all the accepted rules of writing a direct inquiry; a second letter employed a direct request in the lead, and another letter used an indirect or inductive arrange- ment in making its request (this is the structure used in the present study). Boddy's findings were: 1) There was no significant difference in the replies of the three letters. Actually, the letter which violated all the rules and was supposed to be virtually worthless as an answer- getter outpulled both the direct and indirect letters. The indirect letter did only slightly better than the direct letter. - 2) Little difference was found in favor of an individually typed letter as compared with a mimeographed one having the same content. 3) letters without enclosures did almost as well as those having an easybanswer device. 14) Letters mailed by an unidentified writer pulled almost as many replies as letters mailed from a prestige source. In brief, Boddy found considerable discrepancy in what textbooks said about letters of inquiry and what his findings indicated. In answering the question "Upon what does a successful inquiry depend for its effectiveness?“ (12, p. 81;) Boddy concluded: The vital point seems that of reasonableness . . . if the investigation [request] is justified and reasonable, the resu 1+5 lel be nearly the same whatever method is used in ga+hering the material and whatever may be the garnishings placed upon the letter itself (12, pp. 85-86). This year the American Business Writing Association published an index on articles written in The ABW____A_ Bulletin from December, 1938 to May, 1958 (25’) (Th3 ABWA 1_3_u1_l_._etin is regarded as the most authoritative source of research studies in the field of business writing). In this index, only one study, Allgeier's, (1;, pp. 10414) involved effects analysis. Allgeierls study is summarized here. In making an analysis of the effects of personalization in collection letters, Allgeier sent four collection letters to two groups of debtors. The first group was termed the test group, the second the control group. Allgeier prepared the (letters which were sent to the test group. These letters were individuallybtyped form letters written in a friendly, personal, and informal style. Four different collection appeals were used. The letters sent to the control group were standard form letters which had been used by 85 different retail stores. These letters were written in a stiff and stilted language, used stock expressions, and showed little or no interest in the reader. 13 The criterion of effectiveness was the difference in receiving payments from the two groups. Responses were classed as full payment, part payment, and explanation or promises . Allgeier found little significant difference between the groups in the way payments were made. He was surprised to find that for the first six months of delinquency there was pp significant difference in results of the first four letters sent. However, he did find that personali- zation did make a significant difference in results in the latter stage of collection (after six months' delinquency). An important difference between Allgeier's study and the present study is that he had no direct knowledge of his reader's reaction to the letters. No follow-up questionnaire or interview was used. Because of Allgeier's limited information about his readers, his study is essentially an action-effects study. kiwi-M studies and Leela in <_i_i_r_e_g_t_ fl. Numerous action studies have been made on sales letters and similar types by people in the direct mail field (55, pp. 3-1;). However, most of these studies and tests deal with methods and techniques in effectively using direct mail, rather than with content of letters. Test methods employed by direct mail people apparently have been unchanged during the past 20 years. In 1939, Boyd (16, pp. h-S) observed that testing for content was not done by direct mail prac- titioners. All the tests which were being made were tests of mechanical things: - color of stamp, color;of copy, alternate offer or singleness of aim, name fill-in orfiheadline, and so forth. Because of the tendency of mail users to keep their most successful results confidential, such findings, when they were published, were presented in such generalized form as to make them valueless for other test applications. The dis- cussion which follows indicates that the situation as Boyd found it in 1939 has remained essentially the same. In recent literature, many actionreffects tests were found, but, surprisingly, none was found which was limited solely to a test of the two basic appeals or the three approaches used in the present study. The test findings or methods which are regarded as having functional significance in direct mail may not have the same significant value in other applications or writing situations. Boyd mentioned this in 1939: Almost axiomatic among these men (direct mail), however, is that one must make his own tests, because almost identical mailings have produced widely different results (16, p. 5). Such suggestion has been made by Henry Hoke, editor and publisher of mammmmw. who said: There has been altogether too much bunk handed out in class- rooms and in books about so-called "tested methods“ . . . without defining exactly what uses or types are affected by the "methods" suggested (29, p. 9). For certain types of letters Hoke believes testing may not be applicable; however, he continues: . . . with persuasive direct mail, testing may be very important, but . . . your results with persuasive direct mail depend to a great extent on what kind of action you want your persuasion to induce (30, pp. 17-18). 15 Thus, the action one could expect in inducing people to come to a writer's office to obtain cost-sfree help for themselves or others would be different from inducing readers to buy paint brushes in a direct mail situation. Also, a test situation which employs follow~up interviews of readers would not be comparable to direct mail testing because of the absence of interviews in direct mail procedure. Only one study in direct mail was found which involved interviewing people who had received mail. This study by Starch (Sh, pp. 28-s29) was designed to measure sales produced by a mailed circular (no letter was included). Interviews were made of a random sample of customers in retail sales . No questions pertaining to reader reaction to the circular were asked. The principal finding was that circular-receiving customers who came to a store (without ordering by mail) bought more than the non-circular customers. Because no comparable methods or conditions in direct mail testing could be found, research data from direct mail was excluded from consideration in the planning of this study. This is in accord with Hoke's advice: -If you quote statistics or case histories from results of others, be sure the cases quoted correspond as closely as possible to your own overall problems,--your own type of direct mail used, and your own functional use of these types (30, p. 17). Newspaper readership effects studies. Newspaper and advertising people have for a long time been conducting readership studies. to test the effectiveness of their media. A number of such readership studies were reviewed. Particularly helpful in suggesting the interview technique used in the present study was the lid-Study Summary, (56, p. 58) 16 an extensive readership study, conducted by the Bureau of Advertising of the American Newspaper Publishers Association. In newspaper reader— ship studies, trained interviewers call upon a sampling of readers a short time after a newspaper has appeared, and ask the reader what he had read or noticed in the newspaper. Such tests are sometimes extended by'further questions about the depths of the impressions or attitudes which may have resulted from reading an article or advertisement (35, p. 567). In the present investigation, the interview method used also involved personal contact with readers to ascertain their impressions and attitudes; however, here the readers were requested to call upon the interviewer; calls were made to the readers only when there was no response. 11. Literature pp Appeals (Personal benefits appea . Business letter writers almost unanimously agree that the one ingredient which should be contained in practially all types of letters is the personal.benefits appeal, or as it is more commonly termed, the “you“ attitude. Saunders well reflects this acceptance in stating: - The correspondent . . . seeks techniques of persuasion which are to be found in such cardinal qualities of business letters as the uyoua attitude, adaptation, personalization, courtesy, character, the human touch, the positive aspect of ideas, and vividness of their presentation . . . The first and foremost of these techniques we call the “you“ attitude (’49, p. 168). Further evidence of the prime rank which the personal.benefit appeal has among business letter writers is revealed in Keithley‘s (33, pp. 39—h0) 17 review of issues of The ABWA Bulletin published from 1952 to 1958. This review indicates that the idea most frequently mentioned or stressed in the articles of the issues was the personal benefit appeal. The emphasis given to this appeal has long pervaded the teaching and writing of business letters. Boyd, one of the first members of the American Business writing Association, wrote: For thirty years now we have based our material and ways on an expanding conception of the "you“ viewpoint (15, p. h). The unquestioned acceptance of this-appeal by practitioners of business letter writing is predicated on human selfsinterest as the core of motivation. Smart and McKelvey manifest this acceptance in writing: Each reader is primarily interested in the effect that an action will have on himself. ‘Will he benefit from it? Is it ' the best thing for m? (53, p. 213‘ Some evidence from social research suggests support for the strength of the personal benefits appeal. Derived from the'welludemonstrated proposition that people tend to respond in the direction of reducing drives, Schramm (50, p. 210) reports that a suggestion of action is more likely to be accepted if it meets the wants and needs of individuals. 'Qgppp.app§§l. Much less often mentioned than personal.benefit in the literature of business letter writing is the use of the group benefit appeal as a technique of persuasion. When it is recommended, it is usually couched in terms of the personal satisfaction, prestige, or selfw enhancement that one can derive from doing things for others. This is typified in Fellows and'Koenig's (2h, pp. 13-1h, 15-16) text, Hg! £2.§§$§2 Funds‘py'Mail, which frequently makes the point that giving the reader 18 a chance to help others makes him feel important because his help or advice is needed by someone, and, in consequence, his life has in some way been significant. _ Schramm, (50, p. 210), in quoting Krech and Crutchfield, indicates that a group appeal will be more readily accepted if the reader can be made to feel that in helping people he is supporting the norms of a group with which he identifies or wishes to identify with. A view not evident in the literature of business letter writing, but held by some social scientists is that the group appeal may be as basic an urge as the personal benefit appeal. This view is suggested somewhat jointly by Murray (15, pp. 3h2vs3h9) and Bogardus (13, p. 57) who feel that Thomas“ FourIWishes are incomplete. Thomas” Four wishes, namely, the desires for: 1) security, 2) new experience, 3) recognition, and 1;) response, Murray (15, p. 3142) asserts, are regarded by most sociologists as the most fundamental and universal motives or incentives. Hurray who holds that Thomas° classification of motives is largely individualistic, would add other urges, such as: l) the urge to aid others, 2) the urge for just treatment, 3) the urge for freedom, and 1'.) the urge to be creative (us, p. 58). Bogardus would only add a fifth basic urge, namely, to aid others. In support of his reasoning, he states: The urge to help, to aid‘ others in trouble, also seem at times to be generic and to operate even in the face of the other four self-building urges. It may lead to sacrificing one's life for the welfare of persons. It endures great hardships without thought of personal gain or praise (13, p. 58). 19 According to the writer of a monthly letter published by a Canadian bank, high responsiveness to group appeals ,which involve societal objectives, indicates a high degree of personal maturity and cultural attainment (l, p. 12). While many persuasive request letters are written with group appeals, most business letter writing experts express the belief of Menning and Wilkinson, who say: .You will write more successful favor-seeking letters if you select and emphasize reader-benefit talking points (hl, p. 108). £9.92 communications research. Among specialists in communications there is some doubt about the dominance of any one single appeal. Doob (23, p. 122) believes that because of an incomplete knowledge of human physiology, anthropology, and theory of human motivation that no definitive list of appeals can be produced. Martineau, (38, p. 122) an advertising expert who recommends that major stress be given to the emotional approach in advertising, holds that: There is no such things as a single, pure motive for anything. There are many dominant motives and subordinate motives which may be involved in people's acts . . . actually there is no completely stable aspect to motives. What I may feel impelled to do today may be. changed tomorrow by circumstances, or my mood may be different; however, in a general sense, there is a definite consistency in motives (38, p. 32). Martineau's contention about the affect of circumstances and disposition on the effectiveness of an appeal are supported by Katz and Lazarsfeld, (32, p. 216) who maintain that different types of people might react 20 differently to various appeals; but they claim there is no consistent data in communications research to determine what the effect of an appeal will be. Other research data obtained by Hovland, Janis and Kelley (31, pp. llr12) indicates that an individualis predisposition, his thinking habits about the rationale of the arguments presented to him, and his antici» pation of reward and punishment are factors which form the basis of acceptance or rejection of a given appeal. To provide some practical applications of social science research to business prOblems, Abelson made a survey of various studies of per- suasion in communications. A summary of his survey findings which are pertinent to this study are: 1. There is nothing definite on whether he opening or closing parts of the communication should contain the more important material. 2. A strong threat is generally less effective than a mild threat. 3. A.person“s opinions and attitudes are strongly influenced by the groups to which he belongs or wants to belong. h. The level of intelligence of an audience determines the effective~ ness of some kinds of appeals. S. The individual's personality traits affect his susceptibility to persuasion. ' 6. There will be more opinion change in the desired direction if the communicator has high-credibility than if he has 10w credibility. 7. Suspicion of the motives of the communicator may not work against desired opinion change (2, pp. 7-66). 21 A number of other studies indicate that the appeal itself might not be the prime mover. The appeal, these studies claim, affects some desire or want which was already in latent existence (for example, a desire for employment), the appeal suggests a means of obtaining employment. In this case the appeal is 'not the basic thing that would cause action but is a means to an and (employment). This view is sustained by Katz and Lazarsfeld: Appeals might play a more effective role in maintaining and reinforcing decisions, once made, than in initiating them (32, p. 217). This latter point is not at variance with accepted business letter writing practice is suggested by Williams: The "you“ attitude means the letter writer must not only be will» ing to see from the other personis point of view, but be able to visualize what will bring satisfaction, or joy to the other person, and get that into his letters (61, p. 93). III. Literature 93 Approaches In the literature of both business letter writing and mass communi- cations, the terms emotional and rational are generally used as names of types of appeals. In this study, the terms have a different meaning. Here the terms represent a plan of attack, or the setting within which the'two appeals are developed. The approaches thus can be considered the way in which the appeals are presented. Frequently in communication situations, |'how" (or the way something is said) has more impact on action or reaction than "what" is said. This is a commonly accepted principle in business letter writing and other types of communications . 22 hcept fbr the emotional-rational approach, not much pertinent data was found which dealt with the emotional or rational factors as approaches, as defined in this study. Emotional am oach. This approach attempts to provide a setting in which the tone and coloration of the content arouses feelings or an emotional state which induces the reader to be receptive to the appeal. Emmasis of the emotional approach is recommended by someconimunication specialists who believe that action and reaction stem mainly from feeling rather than intellect . This viewpoint can be briefly cited in summary statements by three business communication writers, who fairly well represent the position of the advocates'of this approach. Martineau, an advertising man and one of the strongest proponents of the emotional approach, says: Communications is an interaction not of logic, but of feeling (38: P° 122): In another part of his book, he adds: Human communication is essentially an exchange of feeling, not information (38, p. 197). Bedell, another advertising man, maintains that because people spend most of their time living in a subjective world they are essentially emotional: . . . an emotional people respond to the here and now, near at hand things which have made them 'feel' . A [hilosomer has said that to be alive is 'to be in relation to' . We are in closer relation to things-that make us laugh and cry and feel than to things that make us think . . . emotional reactions provide short circuits to instant response (10, pp. 66-68). 23 Dobbins, a teacher of business letter writing, holds nearly the same idea as Martineau and Bedell, in saying: Usually we feel before we think . . . our automatic reactions are usually influenced more by the spirit behind things than by the things themselves (22, p. 27). No specific studies made under controlled conditions were found which would substantiate the effectiveness in giving primary emphasis to the emotional approach in business letters. Rational approach. Here the major emliiasis in the setting of the appeal is intellectual, logical, or rational. In this approach facts, figures, reasons, contrast and comparison are stressed to induce the reader to be receptive to the appeal. The rational approach in business letter writing, as Boyd (11;, pp. 214-29) points out, is generally stressed in persuasive letters in which the request for action involves deliberation on the part of a discriminate reader. Because the population of this study is college students, who are presumed to possess above average discrimination, of special interest is a report by Schramm (50, pp. 211-212) in which he indicates that there is no evidence that intelligence correlates with susceptibility to logical argument. The exception to these findings is the special case more more intelligent audiences seem to be mpr_e_ influenced by logical arguments, and less influenced by slogans, unsupported generalities and similar devices . 2h In persuasive letters, Manning and Wilkinson (hl, pp. lO6—llll) show that the course of action suggested should be a logical outcome of the reasons given. That is, the reader should be able to deduce with a minimum of effort the conclusion that taking action will help him, or. others in whom he has an interest. Related to Manning and Wilkinson's recommendation is Klapper's (314, p. 296) report on the findings of Hovland, Imnsdaine, and Sheffield, whose research findings indicate that in any successful situation which relies primarin on “letting the facts speak for themselves" the facts presented must be sufficiently specific to require the absolute minimum of inference on the part of the audience. Further, as cited by Klapper, (314, pp. 1064.134) Hovland, Lumsdaine, and'Sheffield's conclusions indicate that sole reliance on rational presentation may prove effective for only a small number of people who respond primarily to rational considerations. Most other people they say, can be motivated mainly through non-rational channels and may be highly resistant to rational considerations. In the literature of business letter writing and other areas of persuasive communications, a number‘of writers suggest that the rational approach cannot be adapted well for use with a group appeal. Brewster and others (17, p. 130) state that appeals employing this approach can promote action if the copy is interesting, informative, educational, and filled with facts that emphasize benefits to the reader. 25 In the literature which was examined, no experimental studies could be found that tested the effectiveness of the rational approach. Emotional-rational approach. Most writers of persuasive communi- cations hold that it is difficult to classify any one approach as emotional or rational (uh, p. 82), since it is commonly accepted that individuals respond differently to the same stimuli, a given approach could evoke an emotional response in one individual and an intellectualized response in the case of another individual. Actually, the literature indicates it is unlikely that either approach could be used without eliciting a mixed emotionalurational reaction in some degree. Ogle affirms this with regard to attendant emotional responses to any type of communications, by saying: It is probably almost impossible to convey any meaning whatsoever by use of language without evoking emotional side-responses (’46: Po 69) ' Emotional reaction is likewise present in communications of a rational nature which involve needs and wants. 'Hepner makes thisobservation in writing: men a need or want is keenly felt, it becomes an emotionally- tinted reaction (28, p. 670). An hypothesis for the presence of different degrees of mixed re- actions to a communication is provided by Albig: The ways of thinking characterized as reason and emotion are not distinct entities motivating particular instances of behavior, but exist in varying proportions in the different situations. Man is never exclusively, usually not even essentially, a reason- ing being. Feelings, emotions, likes, dislikes, in varying degree are component parts of every human situation. It is only for descriptive purposes that one may use the terns 'reason' and 'emotion' (3, pp. 89-90). . 26 The belief that emotion is as important as reason is reflected in many of the articles and books written by people in business communi- cations. Representative of this belief is Aurner's following comment: We must bring within the scope of our teachings both the rational and emotional, both fact and feeling. And one of our greater teaching difficulties is to interpret clearly to the student the complex way in which facts may have an impact upon human feeling and the way in which human feelings may, in turn, seriously affect the interpretation of fact (8, p. 13). Few business letter writers recommend the exclusive use of either the emotional or the rational approach. Most of these writers would use a combined emotional-rational approach. To create attention and interest in a letter, a subjectiVe or emotional treatment would be used first, followed by an objective or factual treatment which would provide reason. for taking action. This technique follows the approach suggested by Manning and Wilkinson, who say: Man is both rational and emotional. He needs a rational reason why to support an emotional desire for something (bl, p. 213). -Expanding Menning and Wilkinson's recommendation, Bedell indicates the same line of reasoning for use of this approach in advertising: People decide emotionally that they want things but consciously or unconsciously they hang back from announcing their decisions until they "have rallied some “good“ reasons acceptable to their intellects. writers should supply-those "good“ reasons . . . recognize that ”people are always seeking Qreasons' for doing what they want tp do or for believing what they want to believe-- the writer must provide a basis for rationalization . . . make the reader want something and then provide what the reader will consider a good excuse for buying (10, pp. 2hl-2h2). Again, as in the case of the rational and emotional appeal, no tests or studies of the emotional-rational approach conducted under scientifically controlled conditions were found. CHAPTER III PROCEDURE AND TECHNIQUE or THE STUDY I. Initiation 2;; Study . “The planning phase of this study was begun by discussing with various. associates the value of and need for making an action~reaction effects study of a particular type of business letter. Dr. Clyde W. Wilkinson, Professor in charge of the Business Writing courses at Michigan State University, encouraged the study. A comprehensive search was made for studies in business writing, mass communications, and social science research. Materials reviewed included unpublished theses, textbooks, bulletins, pamphlets, periodicals, and speeches. I II. greparation pf 3h_e_ Letters Classification 5; $9321.: 1512.9...“ The two appeals and three different approaches employed in the six letters used in this study were divided into two groups of three letters each. The first group, termed Type I, is characterized by the use of the personal-benefits appeal in all of the letters. Only one of the three approaches is used in each of the letters of this group. Individual letters are designated I-A, I-B, and LC. The second group, termed Type II, is characterized by the use of the group-benefits appeal; and , like the Type I letters, each employs 27 28 only one of the three approaches. The individual letters are identi~ fied as IImA, II~B, and II—C. Structural organization pf letters. As pointed out by Hovland and others in communications research, (31, p. 99) the effectiveness of persuasive communications depends not only on the choice of motivating appeals, but also upon the organization of the arguments used in support of the position advocated. Business letter writing experts, likewise, are aware of the need for a specific type of organizational pattern in the writing of persuasive requests. As in the case of the present study, the technique of the persuasive request is used when there is an anticipation of resistance to take action on a request the reader is probably not already willing to take. Taking action on the part of the readers of this study involves direct personal contact with the writer in order to act on the proposition of the letter. Such action might be viewed as an inconvenience by the readers. When the request appears difficult to comply with business letter writing experts maintain that the reader's first reaction is to refuse the request if it is presented in the early part of the letter, (hl, p. 107). Thus as Boyd (lh, p. 140) suggests, because a situation involving requests may constitute a physical or mental burden great enough to tend to deter action, the inductive technique should be used. The inductive technique defers the request for action until the benefits of complying with the request have been presented. 29 Each of the six letters of this study is organized according to the pattern recommended by Boyd and by Manning and Wilkinson. Because the reader is under no obligation to act, arousal of intention and stimulation of interest are particularly urgent in a persuasive request. Cartier and Harwood say: Attention has been viewed by experts in human communi- cations as the sine quo non of reception and has been assumed by them to be important in formulation and transmission (19, pp. 106‘ll0). In the lead of each of the six letters an effort is made to elicit attention by employing some point which suggests a reader benefit or assumed interest on the part of the reader. In the Type I letters the emphasis is on the reader's personal interests; in the Type II letters, the stress is on the assumed interest of the reader in the welfare of others with whom he is associated. To sustain interest in the rest of the letter, continued emphasis is placed on what the reader can get or give. No direct request for action is made until the final paragraph The beginnings and endings of letters are recognized by business writers as occupying positions of h1g1 physical emphasis and visibility (bl, p. 148). However, some communications specialists indicate that there is no evidence to prove that beginnings or endings convey the greatest persuasive argument (31, pp. 118-119). Newspaper readership studies definitely indicate that beginnings command the highest reader attention and influence (ll, pp. 175-177). In the preliminary test of the letters of this study, the endings of four of the six letters commanded considerably higher attention than 30 the lead. Because of a seemingly disproportionate attention to the closes which were written in an emotionalized or partly emotionalized approach, the close of all the letters was written the same way. The close of the letters emphasized what to do, how todo it, and tried to make action easy. So that many details of how to take action in the close of the letters could be avoided, a printed card was enclosed with the letter. The card provided information about the specific hours the reader could telephone to make an appointment, directions concerning the location of the writer9s office, and the telephone number to call (see Appendix A). Layout _a_r_1d_ format pf letters. To give the letters a personalized appearance, individuallyhtyped form letters were prepared. The surname and title (Mr., Mrs. or Miss) of the reader were used in the salutation of the letter and on the envelope. ..F_‘_I_L_e_s_c_h_Readability Scoring 3f Letters. In Chapter I (page 5), reference is made to the Flesch Readability Index as a control device to minimize the effects of the language and style of the letters as factors in action and reaction. Use of this index made it possible to retain essentiale the same reading and human interest level in all of the letters. Thus, at least theoretically, there was no difference in the readability or human interest scores of the letters. No statistical measure of readability or human interest provides an absolute test of the characteristics measured; it is not possible to reduce language to a number. The experience in this study was that a 31 strict application of a readability formula could produce a mechanistic arrangement of sentences, unnatural usage of personal references, and a forced substitution of undesirable shorter or longer words to comply with a requisite syllable count. Despite these limitations, the Flesch Readability Index.is widely used in many areas of communications, and it is generally regarded as one of the most reliable measures of readability (5, p. lh2). The Flesch Readability Formula used in this study is the one described in Flesch's book, _Tpg _Ar__t 2f Readable Writing (26, pp. 213-216). This formula.has two parts. The first measures reading ease on.the‘basis of sentence length and word length (Reading Ease Scores). The second part measures reading interest on the basis of “personal? words and sentences (Human Interest Scores). Table I of this study contains the human interest scores and reading ease scores of the six letters used in this study. TABLE I FLESCH READABILITY INDEX SCORES OF LETTERS 1 ‘ Reading Ease Human Interest . letter Types Score Score I‘A 7605 58 .2 1-8 7600 56 .0 I-0 75 .3 56 .5 II" 76 05 58 .0 11"3 7600 S7 .0 11‘0” 76 oh S7 05 Averages 76.h 57.2 32 All of the reading ease scores of the six letters are within the l-'fairly easy“ reading level. This reading level, according to Flesch,“ is comparable to the general reading level used in.Saturday Evening Post. Here the human interest score of the six letters is on the high side of the uhighly interesting" level. The human interest scores of the rationaland the emotionalwrational letters of both Type I and Type II groups are higher than the level that would normally be used in a nonrexperimental situation--ordinarily, not so many personal references and sentences would be employed. Faculty evaluation'pf preliminary test letters. At various times during the preparation of the letters, the writer consulted with four members of the business letter writing faculty to obtain their critical evaluation of the proposed content of the letters. After the firstdraft of the letters had been completed and the Flesch Readability Scores had been computed, the letters were submitted to the faculty. All four of the faculty agreed that the letters substantially conveyed the intended theme of the appeals and approaches. Three drafts of the letters incorporating the suggestions of the business letter writing faculty were prepared. Upon the final approval of the faculty, the letters were readied for preliminary testing with student readers. 33 III. Pre-Test pf Letters An attempt was made to make the pre-test as similar as possible to the planned procedure for the regular study (described on pages 36-h6 of this study). The object was to provide assurance that the contemplated larger study was sound in concept and as complete as possible in every detail. Selection pf premtest sample. Preliminary test letters were mailed to twelve randomly selected students, whose names were Obtained from the enrollment card files of the Registrar's office. A “goldfish bowl" method of random selection was used. Two letters of each individual type were mailed to the pilot test sample. Care was exercised so that one of the six letter types was received by both men and women. Pre-test results pf respondent schedules. The pilot study results caused eight changes in the interview schedules: 1. In the interviews with the respondents, it was necessary that each question be so designed that a parallel could be drawn between it and some pertinent point of contrast in the letter of request. 2. To avoid structured responses or suggested reactions, a general form of interrogation was used in constructing most of the questions. For example, instead of saying, “What did you first think about when you got a letter from an instructor offering 3h you help in writing a letter of application?" the following unstructured question was used: "What did you first think about the letter I sent to you?“ Several phrasings of each question were prepared to determine which seemed best adapted to the level of the students. To make the questions appear more applicable to the re- spondents who received the personal benefits letters, all the questions in that schedule were phrased in a second person, personalized form. On the other hand, to avoid making the respondents of the group benefits appeal feel that they were the primary subjects of the interview, most of the questions in that schedule were phrased in the third person, impersonalized form. The phrasing of the non-respondent schedule was second person, personalized style for the recipients of both types of appeals. Here, because of the brevity of the schedule, it was necessary to obtain direct information about the informants. Suspicion seemed to be aroused when a question was too deeply probed. This appeared to be particularly evident in the first part of the interview. In consequence, the depth aspect was minimized. Because the students regarded their problem as vocational, the depth interview did not seem appropriate in this situation. 35 h. A."warm~up“ period of casual discussion of the respondent’s employment plans or other interests was used to establish rapport. 5. Care had to be exercised not to suggest in any way the inter- viewer's real intentions in obtaining information. 6. Placing the microphone of the tape recorder out of the studentis view, and making only incidental reference to the use of it . during the interview seemed to make the student less conscious of it and put him more at ease. 7. To obtain a free exchange of conversation, the interview had to assume the characteristics of a natural, personal conversation. The interviews seemed less mechanistic and seemingly more natural to the student when the interviewer stated his questions from memory rather than read from the typed schedules. 8. Reference to a letter of request at the beginning of the interview was acceptable if only a limited number of questions were asked about it. Pre-test results of telephone interviewee schedules. In interviewing the telephone respondents, the following observations and problems were noted: 1. In these interviews there did not seem to be much difference in reaction or cooperation in presenting the questions about the letters of request first or last in the interview. 36 2. Because of the need to limit the number of questions and narrow the depth, specificity, and range of the questions, the prewtest indicated that it would not be possible to make direct or complete comparisons between some of the replies of the respondents. The results of the pre—test were not included in the larger study. IV. Sampling Procedure Size 2f sampl . At the outset of the study, the writer realized that a restricted sample size was imperative because of the conditions under which the study had to be made. These conditions were: 1. Only one person would conduct the study. 2. The sample group, which consisted of college seniors, would be available for a limited period. 3. The personal interview technique employed would limit the number Of interviews that could be handled by one person during the time available for the interviewing. Computation'gf sample size. The procedure used followed the method described by Parten (h8, pp. 290-330). As a first step, an estimate was made of the percentage of the replies which could be expected from the entire sample. The pre-test produced a 50 percent return which might be regarded as suggestive of what the action response to all six letters might be. However, since the pre-test sample was quite small there is some question as to the 37 accuracy of the yield of the pre-test. No data was found in the literature of business letter writing which dealt with the percentage returns which could be expected of non-sales, persuasive letters of this type. In the case of new product selling by mail, the results are usually 2 percent (18, p. 295). 'Hhen in complete doubt about an expected percentage return, Parten (h8, p. 327) recommends a preliminary estimate of 50 percent. This figure was used as the initial estimate of returns. The next step was to decide how much the percentage to be secured from the entire sample could vary from the true (unknown) value for the population and still be acceptable. Again, no specific published data was found which would indicate an acceptable limit of tolerance on returns of persuasive letters of request. However, in assisting many individuals engaged in educational research and administration, who used persuasive letters, it has been the experience of the writer that a range of 15 to 20 percent error on a 50 percent estimate return of the first request was acceptable. However, to comply with the minimum statistical requisites in selection of a sample size for this study, a 7 percent limit of error was chosen. ' To determine the necessary sample size, a formula provided by Parten (h8, p. 312) was used. Computation of the formula indicated that a random sample of 20h cases should yield an estimate with an assurance of 19 out of 20 that it is within the 7 percent chosen limit. Another factor considered in the selection of the sample size was its adequacy to provide sufficient cases to take care of the number of 38 categories and classes into which the findings of the study were to be grouped and analyzed. .A procedure recommended by Parten (h8, p. 298) was employed which provided assurance that the sample size selected should be adequate to handle the various classes of the study. However, since the plan of the study had called for the use of two or more categories to analyze each letter class, some difficulty was anticipated in the use of the Chiwsquare statistic (see pp. 51, 65-67, and 87--88 of this study). Drawing the sample. As mentioned previously, the population of the study consisted of a random sample of 20h undergraduate senior students who were registered in the fall term, 1958, at Michigan State University. The students were enrolled in.every college of the University except the Basic College in which, generally, only freshmen and sophomores are enrolled. To obtain a reliable, up—to-date listing from which a sample could bedrawn, the Michigan State University'Directoqugf Students, £22§122E2: was used. 8 Since a convenient listing was available, sample cases of seniors were selected at regular numerical intervals in the Directory. The numerical interval used for the sample selection was obtained by the following computation: 3,133 (Total number of seniors) , 20b, (Sample number) a 16.82813 (sample interval) '. 39 To insure that each name would have an equal chance for inclusion in the sample, this number 16.828h3 was used in the whole range of the sample. This was accomplished by first using a table of random numbers (27, p. 171) to Obtain the first name and number of the sample. Succeeding sample names were obtained by accumulative addition of the sample interval starting with the first sample number. Because the sample of 20h had to be split up into six groups to whom six letters were to be sent, a further refinement of the sample had to be made. In consequence, 20h was used as a subwuniverse to yield a sub» random sample for each of the six groupsu~in effect, a random sample (20h) was used to obtain six subwrandom samples containing 3h cases each. A die was used to determine both the starting point of the sub“ samples and the order in which the letter assignments to each sample group was made. Reliability of sample. To determine how well the base random sample was proportionately representative of its universe, two comparisons were made. The comparisons had to be made of two types of information which were available about both the sample and the universe. In this study, knowledge of the sex of the students and their college registration was available. Information about the sex and college enrollment of the sample was obtained from the main entries in the Directogy.9§ Students. The Registrar's office-provided information about the actual college and male~and~female enrollment of the seniors who were graduated in the 1958~1959 period. hO Because these statistics were not available until July 3, 1959, the 4% 3.3.2933. 9_f_ the Registrar for 1957-1958 was used to obtain an initial estimate of the proportionate representativeness of the sample. The Report provided data on the total number of undergraduate degrees conferred by all the colleges of the University and the percentage of men and women graduates. The percentage of women graduates in the 1957- 1958 Report was 27.5h percent. The sample percentage was 29.hl percent, a difference of 1.87 percent; however, the difference between the college enrollment percentage of the sample and the Registrar's figure was an average of approximately 6 percent. Since the comparisons appeared reasonably close, allowing for possible differences in the 1958~l959 ' figures, the assumption was made that the sample was random. The 1958~l959 Report from the Registrar confirmed the assumption of the random list, at least on the basis of sex and college enrollment. A 0.15 percent difference was found between the nunber of male- female graduates and the number in the sample population. .A Chissquare computation between the college enrollment of the graduates and the number of seniors in the sample indicated a high association.between the population universe and the sample; what differences there are in this comparison could be attributed to factors of chance. The percentage difference in the proportions between the graduates of the colleges and the sample is 1.66 percent. hl V. Mailing System and Appointment Scheduling Mailing procedure. After the sample processing was completed, a mailing list was prepared from the Student Directory. To allow for sufficient and convenient time for reader action, one of the six letter types was sent to one group each week over a period of six weeks. The letters were mailed on Saturday afternoons from the East Lansing Post Office. Thus, the letters arrived at the students1 local residences on Monday~~usually in the morning. The first letters were sent to those students who were to be graduated in the fall term. This involved the sending of all six letter types in one mailing, a practice not followed in the other mailings. Handling appointments and calls. The students were given one week in which to make an appointment (during the week they received the letters) because a longer period might affect the retention of their initial impression and reaction to the letters. Students who, on their own accord, called after one week at the office or by phone were given appointments for interviews, but they were not classified as respondents. Telephone calls were made to students who did not respond one week after a set of letters had been mailed. Classification 2; unavailable students. No other means of communi- cation than the telephone calls was used to reach non-respondents who had no listed local telephone numbers or local addresses. Follow-up h2 letters or questionnaires were not used because such procedures would inject variables difficult to evaluate. However, an attempt was made to identify the non~respondent students who could not be personally reached by phone. Calls were made to former households of the students, their college departments, the Registrar‘s Office, and associates of the non-respondents. On the basis of the information that could be obtained about this group of non~respondents, they were classified as follows: 1. Restrained. This group included individuals who did not have equal opportunity to act on the proposition of the letters. They included: (a) practice teachers, (b) those who lived out» side a twelve~mile radius from the campus, (c) students serving internships (other than teaching), (d) dropuouts, and (h) those who changed addresses without leaving forwarding addresses. 2. Indifferent. No information could be obtained from or about this group. VI. Preparation pf Interview Schedules ‘Q g gf‘g disguised interview for respondents. The two principal considerations in designing the interview schedule was to construct a schedule which would cause a minimum of suspicion about the interviewer‘s actual purpose in (1) requesting action from the student, and (2) using— an interview to obtain primarily the kind of information desired. The essential characteristic of the disguised interview used in this study is that it conceals from the student the intent of the request for action and the interview method used and thus allows him to respond h3 freely without feeling he is the subject of an experiment designed to measure his feelings and attitudes about the letter he received. ‘Qgsigg.2£ interviewee schedules. The respondent interview schedule was prepared to cover questions about: 1. Factual data about respondent. 2. Reaction to content and appeal of letters. 3. Reaction to writer. h. Employment plans and job qualifications of interviewee. 5. Comparisons of letter of request and letter of application. 6. Organization of letter of application. Of the 28 questions in this schedule, 18 pertained to information which was not directly needed for analysis purposes; they were included to provide a setting for the disguised interview. The telephone-interviewee schedule questions were similar to those used in the respondent schedule except that no questions were included which directly referred to 1) reaction to writer, 2) comparison of letter of request and letter of application, and 3) organization of letter of application. Such questions could not be used in the interviewee schedule because they involved more detailed questioning than could be expediently handled in a:telephone interview. Each question in the respondent and telephone interviewee schedules was carefully designed and selected with the purpose of obtaining the specific kinds of pertinent information necessary to accomplish the Objectives of the study» The type of analysis made of each question is illustrated in the following example: m. Question: "Had you given any consideration to the possible use of a letter of application before coming here?" Objective: This question was designed to gain information about the predisposition to action on the part of the studentwwwas he in a state of readiness to act? VII. Procedure in Conducting Interviews Since the interviews were conducted entirely by one individual, under no direct supervision, various checks were made to be sure that accepted procedure was being followed. To supplement the training and experience which the investigator has had in conducting interviews, selected references were reviewed. The interview procedure most closely followed was an adaptation of the procedure described by Merton, Fiske, and Kendal, in ‘Thg Focused Interview, (h2). A chapter by Barley, in Brayfield's text, Readings 3g Modern Methods 2: Counseling (21, pp. 265-272), on the conduct of the interview, also proved quite helpful as a guide. Conduct 2; respondent interview. The first discussion with the respondent dealt with the objectives of the interview; this was followed by an informal reference to the respondent's training or jdb plans. Reference to the use of the tape recorder (a small desk-size recorder) was carefully made in terms of being an aid to the interview procedure. The machine was turned off during the discussion periods of the interview (listed as A, B, C, D, and F in the respondent schedule) and whenever any interruptions occurred. hS Because the interviewer had memorized the questions, they could be asked in a conversational manner which helped to create an informal atmosphere rather than a test situation. Care was taken to suggest little motivation for either suppression, or distortion, or put the student on the defensive. To provide control of the interview and avoid conversational digressions, the schedule order of presentation was followed as closely as practically possible (copy of schedule is in Appendix B). Sufficient flexibility was permitted to give the student an opportunity to ask related questions about various points; thus an opening was provided for obtaining, sometimes, valuable "unanticipated“ responses or views. The interview was.terminated with an expression of the interviewerls appreciation for the respondentis participation. The average interview time was one hour, sixteen minutes, with a range of fortyheight minutes to two and onemhalf hours. Conduct 2; telephone interview. Because the writer did not have the necessary equipment to tape-record telephone interviews, the record of the interviews was made on the schedule forms. Unlike the personal interviews, not much time was available after the initial introduction to establish rapport and explain the purpose of the call. After the objectives of the requested interview had been introduced and discussed, the questions of the schedule were asked in the sequence indicated in the telephone interview schedule (see Appendix B). he During the telephone interviews a request for a personal interview was often made. In many instances, the informants stated that they had planned to arrange for an interview, but for various reasons they were unable to call for an appointment. Such requests for appointments were granted for two reasons: (1) to avoid suspicion on the part of the informant and (2) to employ such personal interviews as another means of checking the reliability of the schedules (see pages h8~h9). All office interviews of this type were classified as nonnrespondent interviews; however, the schedule used for analysis purposes was the one used in the regular telephone interviews. No difficulty was experienced in obtaining information from the informants, once they were contacted by phone; and to the surprise of the investigator, the informants appeared to remember considerably more about the content of the letter than had been expected. Immediately following the interview, the investigator read through the schedule form and edited the schedule for omissions, incomplete statements, etc. The time of the interviews ranged from 17 minutes to one hour and 1h minutes. The average telephone interview time was 21 minutes. VIII. Reliability and Validity Tests 2; The Interview Schedules Three methods were employed to determine the reliability of the interview schedules. The first was a modified application of the "splitmballot" technique (h8, p. D95) Which was employed during the pre- test of the schedules. The second was a repetition of the telephone ' I p h? interview schedule. The third method (described in the section on coding, page Sh of this study) was a cbmparison of the coding of assigned categories of the open-ended responses in the schedules by three graduate assistants. The first two methods are described in the following two sections. Split~ballot test gf reliability. Because of the small number of interviewees used in the pre-test, this method was employed on a very restricted basis. To test the effect of using personal as opposed to impersonal phras— ing of the questions, each question dealing with the same point of inquiry was phrased in the two opposite ways. Sequence was tested by using in the first half of one schedule questions relating to the student‘s personal job qualifications and a letter of application, and questions about the letter of request in the second half; in the second schedule, the arrangement was reversed. Phrasing and sequence were combined to form two schedule types. Type A consisted of personal phrasing and the vocational qualifications sequence, and Type B was the opposite. The two schedule types were split between the six pre—test respondents and four non-respondents. While it is a somewhat questionable practice to generalize on the basis of only ten cases, the results suggested a marked difference in the reactions to the two schedule types. All the students who were interviewed with the Type A schedules had strong favorable reactions to the letter of request, while the other five students all expressed some negative h8 reactions to questions concerning the letter of request and seemed more suspicious of the writer9s motives. Since there appeared to be two extremes here, the sequence was modified as arranged in the completed schedules. The 50ucalled splitmballot technique used might not properly be regarded as a test of reliability in the usual sense in.which this technique is employed (reliability would be measured in terms of how closely the results of different schedules approximate one another). However, the modifications which were made as a consequence of studying the schedules suggested that the schedules would be reliable. Reliability testfigf telephone interview schedule. As mentioned previously, telephone interviewees who had asked for an office interview were granted such requests. This provided an opportunity to check the reliability of the telephone interview schedule since the same schedule was used in the office interviews. The only difference in the office interviews and the telephone interviews was the elimination of points A, B, C, and question 1-1 of the schedule used with the informants who indicated that they had received the letter of request and that they had carefully read the letter. An average of 93.5h percent of the questions of the office inter- views yielded replies which were in agreement with those of the telephone interviews. The percentage of agreement indicates a fairly high consist- ency in responses to the two interviews; this fact suggests that the telephone schedule can be regarded as reliable. 10 Since most of the questions in the telephone schedule are similar or parallel to the respondent schedule, results of the comparison of agreement between the telephone interviews and the office interviews suggest that the respondent schedule is likewise reliable. Validity £331; 9; schedules. The complex nature of this study and its "pioneer" status made it very difficult to obtain some objective measure which could be used to test the validity of the answers to the questions contained in the various schedules. A measure which was assumed to be indicative of validity was a comparison of the overall concept which most of the interviewees had of the six test letters with the major impressions which a selected non-u test group had of the letters in terms of perceived differences and similarities of the letters. The nonutest group was composed of ten faculty members and graduate assistants in the writer°s college. Each individual was asked to read three paired sets of thesix test letters. The pair consisted of a Type I .appeal'with one of the three approaches paired with a Type II appeal and the same approach. For example, a Type I-A letter was paired with a Type II-A letter, I-B with II-B, and I-C with II~C. Each time after the individual had read one set of letters, he was asked the following questions: 1. Do you see anything in these two letters which they might have in common? SO 2. Do you see any difference between these letters? 3. In considering these letters, do you think there are more differences or more similarities? .After the three sets had been read, the next series of questions was asked: 1. Do you see any similarities in the six letters? 2. Any differences? 3. What was most meaningful to you, the differences or the s imilarit ies‘? As reported on pages 80-81 of Chapter V, there was (1) a considerw able uniformity in the replies of this group and (2) a close parallel to the general concept which most of the interviewees had of the letters of request. II. Eggigg 9f the Responses and Classification.Data In the preparation of the data for machine tabulation and analysis, a coding system employing numerical symbols for each type of data was used. Three types of data were coded: 1. Identification data (sample number and type of interviewee). 2. Classification data of two main types. The first was the letter types, and the second type of classification data were the criterion groups (college, sex, marital status, grade point average, socio-economic status, training in business letter writing, extra-curricular activities). 51 3. Responses to questions of the schedules. The codes for all three types of data were entered on code sheets. The code sheets were used as master sheets by the IBM machine operators to enter the code on the punch cards. The code sheets also provided summary totals of the code numbers in each column. A separate code sheet was prepared for the respondent and telephone interviewees; thus the coding for the two groups was entered on two different punch cards. In the first assignment of code symbols for the various criterion groups and responses to the questions, broader classes were used than appear in the code sheet which was used for mechanical tabulation. Because of the small sample size and the use of the Chiwsquare statistic, it was necessary to reduce the classes and categories to the smallest feasible combinations. Thus, the codes which were entered on the punch cards represent the combined classes and categories. ‘Letter types. The six letter types (two basic appeals and three approaches) were used as classification data to ascertain how the inter- viewees as a whole, that is, without regard to college enrollment, sex, marital status, etc., responded to the two appeals and three approaches. The letters were first coded and recorded on the punch cards as six classes (OuS codes); however, in order to use the Chiesquare statistic to determine significance, the six classes had to be broken down into the two classes of the appeals and, again, set up as three classes of the approaches. S2 College enrollment. Eight of the ten colleges which comprise the University were included as a criterion group. The colleges which were excluded were the Basic College and the Honors College. The first classification table of colleges was broken down into two college groups. One group included those colleges whose main function involved technical laboratory studies. This combination of colleges is. identified as the “Science" group. They included the following colleges: Agriculture ‘ Engineering Home Economics Science and Arts* Veterinary Medicine The uApplied Arts" college group is the second division. This group consisted of the colleges of: Business and Public Service Education Communication Arts Science and Arts* Sex. This particular classification presented no special problem. Marital status. Only two classes~~married and singleuwwere used. The Student Directogy did not provide information about those students who were separated, divorced, etc. Grade_point average. A rather broad range was used here to avoid the use of more classes than the data could handle. Three grade *Students in the Science and Arts College who were in the non- technical departments of this College (Art and Foreign Languages, Foreign Studies, History, etc.) were classified with the second group of colleges, and those who were in the technical departments were classified with the I first group. 53 classes were used: 2 .002 .19 2 .5042 .99 3 .OO-b, .OO flrawcurricular activities. Three code classes were used: Very active: Participant. in three or more organizations; or elected officer in an organization. Active: Affiliated with fewer than three organizations; not an elected officer; active in minor committee work. Inactive: Not a member of any organization, or inactive member in one organization. Socio-«economic w. This classification, based on skill, status, and economicgposition, is a modification of Alba Edward's (I43, pp. 367m368) and W. 1.. Warner and others' (58, p. 66) occupational classification according to socio—economic status. I The three classifications are: Upper upper-Lower upper . Upper middle-Lower middle Upper lower-Lower lower Training in business letter‘writing'. To determine whether there is any relationship between the responses of the interviewees and their experience with business letter writing training, three classes were employed . 5h Had course or was taking course from writer. Had course or was being trained by other Business Letter 'Writing instructors. Had not had course and was not enrolled. Reliability Test gf‘ggding. Three assistant‘instructors from the ‘writer°s college were employed to check the reliability of the coding of the various categories of the responses to the cpenwended questions of the schedules (one criterion question~~socio~economic status~ewas included in the reliability check). To assist the code verifiers, the edited verbatim responses to the openwended questions were typed on separate pages. A typed page was provided which included the coded categories and typical responses subsumed by the categories. The checkers were in~ structed to code the verbatim responses to each question in accordance with the code which they believed was appropriate. The checkers worked independently. Their coding was recorded on separate forms which were provided (see.Appendix C). After the code check was completed by the verifiers, their coding of each question was compared with one another's to determine agreement in the coding of the questions. A table for each question was prepared. The average percent of agreement in the coding of the answers of the respondent schedule was 87.23 percent, the average percent of agree~ ment in the case of the telephone responses was 89.23 percent. CHAPTER IV ANALXSIS OF ACTION EFFECTIVENESS OF LETTERS This chapter deals with the first objective of the study: determination of which letter type and approach will produce the most action from the sample. An action analysis employing Chiesquares and percentages is made of the letters, letter types, and approaches in terms of: 1) action responses, 2) telephone responses, and 3) no re’ sponses. The same kind of response analysis is made of the criterion classification groups for the respondents and telephone interviewees. Because of a lack of information, the analysis of the criterion groups of the no~reply group is limited to criterion classifications of: 1) college enrollment, 2) sex, and 3) marital status. I. Action Analysis ‘Egtgl responses 32 letters. Table II presents a summary of the action responses to all six letters and two letter types (appeals). No significant difference at the .05 level of probability was found between letter types ofe~ (a) Respondents (b) Telephone interviewees (c) No reply group (d) Respondents and telephone interviewees. SS 56 madamm Ho mommo now do momma mmmpdmouonnlmflmmnvqoamm ma mwmpdoonmam .mHmpop menace do vmmmp amendmoumanumwmompnonmd ma mmmpdmonomm .pmppma some Ho mmwwo an no women owmpcmonmma sow a: Ao.smv mma no.0sv Na Am.mmv mm as.osv sense happen to mampoaa Noe mm Am.::v ow no.amv a: no.5mv mm Ao.~:v HH snap noppog maesoam an a m.em mm m.ms m m.mm Fa o.om . onaH an s a.sa mm m.ms ma a.ss ma N.mm muHH am e o.om aw :.ms we m.mm m m.om «sHH moa am Aa.mmv ms as.®sv am Ao.msv as Ao.mmv H mass popped. newness :m 0H s.mm am 0.05 a 0.0m as 0.0m o-H an a ©.om am s.ms NH m.mm ma H.:s muH am as s.am em 0.05 Na m.mm NH m.mma «uH z z pcmonmm z vamopmm z pmoonom z pqmohom nausea moflaamm pmpompdoo mpompnoo popped asses -. oz Hmpoa meosamama mmmwa mmaamn n24 mmmaamn OB muconcommmm unoppou [II mmmzommmm 26H904.A