journal of personality volume 35 issue 2 1967 [doi 10.1111%2fj.1467-6494.1967.tb01433.x] may e....

16
Personality impression change as a function of the favorableness of first impressions May E. Briscoe, Howard D. Woodyard, and Marvin E. Shaw, University of Flonda Smce the classical study of impression formation by Asch (1946), there has been much concem wath the vanables lnEuenc- mg this process The importance of one variable that seems to have been neglected, the quahty of the lmtial impression, is sug- gested by an mcidental findmg reported by Freedman and Stein- bmner (1964) They gave Ss mformation about an individual who was identified as a candidate for admission to graduate study at a large university Half of the Ss received highly favorable and half highly imfavorable mitial mformation about the candidate He was dien rated for desirability for admission to graduate study Next, Ss listened to a transcnpt of a purported mterview between the candidate and a member of the graduate faculty of the umversity The information revealed by the interview was highly favorable for Ss mitially receivmg unfavorable in- formation and highly unfavorable for those initially receivmg favorable information about the candidate All Ss again rated the candidate on desirabihty for admission to graduate study Significandy more change occurred m the group that had mitially favorable mformation than m the group that had mitially unfavor- able information This findmg suggests that an unfavorable first impression is more diflBcult to change than a favorable one However, since Freedman and Stembruner were mterested m a different problem, certam desirable controls were missmg m their stiidy In the first place, the initial and final mformation differed m kind, smce the mitial mformation represented an objective record whereas the final mformation apparendy was based on the mterviewer's opmion. Secondly, although the mitial favorable and unfavorable mformation was shown to be equally favorable or unfavorable.

Upload: adina-ploesteanu

Post on 17-Sep-2015

216 views

Category:

Documents


2 download

DESCRIPTION

Study

TRANSCRIPT

  • Personality impression change as afunction of the favorableness offirst impressionsMay E. Briscoe, Howard D. Woodyard, and Marvin E. Shaw,University of Flonda

    Smce the classical study of impression formation by Asch(1946), there has been much concem wath the vanables lnEuenc-mg this process The importance of one variable that seems tohave been neglected, the quahty of the lmtial impression, is sug-gested by an mcidental findmg reported by Freedman and Stein-bmner (1964) They gave Ss mformation about an individualwho was identified as a candidate for admission to graduate studyat a large university Half of the Ss received highly favorable andhalf highly imfavorable mitial mformation about the candidateHe was dien rated for desirability for admission to graduatestudy Next, Ss listened to a transcnpt of a purported mterviewbetween the candidate and a member of the graduate facultyof the umversity The information revealed by the interviewwas highly favorable for Ss mitially receivmg unfavorable in-formation and highly unfavorable for those initially receivmgfavorable information about the candidate All Ss again ratedthe candidate on desirabihty for admission to graduate studySignificandy more change occurred m the group that had mitiallyfavorable mformation than m the group that had mitially unfavor-able information

    This findmg suggests that an unfavorable first impression ismore diflBcult to change than a favorable one However, sinceFreedman and Stembruner were mterested m a different problem,certam desirable controls were missmg m their stiidy In thefirst place, the initial and final mformation differed m kind, smcethe mitial mformation represented an objective record whereasthe final mformation apparendy was based on the mterviewer'sopmion. Secondly, although the mitial favorable and unfavorablemformation was shown to be equally favorable or unfavorable.

  • 344 May E Briscoe, Howard D Woodyard, and Marvin E. Shawthere apparently was no attempt to equate the relative favor-ability of the final mformation Thus, there is the possibihty, aspomted out by Freedman and Steinbruner, that the result is dueto the particular communications employed (It may also havebeen due to the particular kmd of evaluation that Ss were askedto make ) Because of these questions, it seems desirable toexamine the effects of the favorability of the first impression undermore highly controlled conditions The major purpose of thepresent studies, therefore, was to provide a more general testof the hypothesis that a negative mitial impression of anotherperson is more difficult to change than a positive one

    There are two other variables that should be considered. Thefirst of these is the order of presentation of mformation about theperson, smce of course order is mextncably mvolved m impres-sion change Many studies have shovra. the importance of thisvariable, but the direction of its effect has not been consistentUnder some conditions a pnmacy effect is found (Anderson,1965, Anderson & Barrios, 1961, Anderson & Hubert, 1963, An-derson & Norman, 1964, Asch, 1946), whereas under others arecency effect is reported (Anderson & Hubert, 1963, Luchms,1957, 1958, Rosenkrantz & Crockett, 1965, Stewart, 1965)

    The second vanable that should be considered is the modeof respondmg Information about the person may be presentedsequentially and the impression measured at the end of thepresentation (Anderson, 1965, Anderson & Bamos, 1961, Asch,1946), or alternatively, Ss may be asked to give their impressionafter part of the information has been presented and agam atthe end of the presentation (Luchms, 1958, Stewart, 1965) Thefinal mode of respondmg usually results m a pnmacy effect,whereas the mterpolated mode (called "contmuous mode" byStewart, 1965) produces a recency effect

    A secondary purpose of the study, therefore, was tx) mvesti-gate further the effects of mode of respondmg on impressionchange If both differential difficulty of changmg unfavorableand favorable first impressions and mode of respondmg arevariables in determmmg the recency-pnmacy effect, it wouldbe worthwhile to determme the jomt effects of these vanablesDo the effects merely summate, or could one vanable be sopowerful that it negates the other? Such a test is difficult, smce

  • Favorableness of first impressions 345a direct measure of impression change is impossible with thefinal mode of respondmg However, an mdirect measure ispossible by means of control groups who rate the mdividualafter either initially favorable or mitially unfavorable informa-tion has been presented If we assume that randomly selectedSs are mitially eqmvalent, the control group scores may be usedas a base lme for estimatmg the relative difficulty of changmgmitially favorable or unfavorable impressions under final re-sponse mode conditions Assummg that the two variables areapproximately equal m strength and that their effects are addi-tive, both the mterpolated response with favorable mitial m-formation and the final response with unfavorable mitial m-formation should produce an unfavorable final impression. Simi-larly, the mterpolated response with unfavorable mitial mforma-tion and the final response with favorable mitial informationshould produce neutral final impressions.

    The pnmary purpose of the first experiment was to determmewhether a favorable first impression is easier to change than anunfavorable one, and secondarily, whether the favorability vari-able mteracts with the mode of respondmg vanable The secondexpenment was designed to test further the generality of thefindmgs from Experiment I

    EXPERIMENT I

    METHOD

    SubjectsThe Ss for this expenment were 37 male and 37 female volunteers

    (age 18-25 years) from mtroductory psychology courses at the Uni-versity of Flonda Two female Ss were discarded due to incompletedata Twelve additional Ss (seven males and five females) wererandomly discarded to provide for equal numbers m each cell, re-sultmg m a final sample of 30 males and 30 females Five Ss of eachsex were randomly assigned to each of six expenmental groups.

    Expenmental DesignSix expenmental conditions were employed, as outhned m Table

    1 Ss m die Interpolated UF condition read unfavorable informationabout an mdividual, rated the mdividual by means of semantic dif-ferential scales, then read the favorable information about the sameindividual, and rated him agam usmg the same scales Ss in theInterpolated FU condition were treated m the same way, except for

  • 346 May E Briscoe, Howard D. Woodyard, and Marvin ETable i Summary of experimental treatments (Experiment I)

    Condition

    nterpolated UFnterpolated FU

    Final UF

    Control U

    Sequence of event.

    Unfavorable information-Rahng-Favorable information-ReratingFavorable mformati on-Ratmg-Unfavorableinformation-ReratingUnfavorable informahon-Cancellation-Favorable information-Rating

    Unfavorable informFavorable informat

    ation-Rating-Cancellation-Reratingon-Roting-Cancellation-Rerating

    the order m which the information was presented The Fmal UFand FU conditions were the same as the correspondmg mterpolatedconditions except that a cancellation task was substituted for themitial ratmg The control groups received only favorable or un-favorable information, with a cancellation task mtervenmg betweenthe mitial and final ratmgMaterials

    The matenals used m this experiment consisted of a favorabledescnption of a person named John, an unfavorable descnption ofthe same person, two response sheets contammg identical sets ofrating scales, a cancellation task, and a questionnaire The vanousmatenals were prmted on separate sheets of paper and were assembledm booklet form Six different forms were prepared, correspondmg tothe SIX expenmental conditions oudmed m Table l

    Descrtpttons The two descriptions of John were selected from anongmal set of 14 descnptions, on the basis of ratmgs by 19 judges(11 males and eight females) drawn from the target populationRatmgs were made on the seven-pomt semantic differential scalesused m the mam part of the study The two descnptions were se-lected for use m the expenment accordmg to the followmg cntena(a) their respective scale values diflEered significandy from ^e neutralpomt (400) but m opposite directions from it, (b) the two descnp-tions did not differ significandy m terms of absolute distance fromthe neutral pomt, (c) both of the descnptions conceivably couldhave referred to the same mdividual

    The mean scale value for the favorable descnption was 613(SD = 062), while that of the unfavorable descnption was 201(SD = 0 61) The two descnptions follow

    Favorable DescnptionJohn IS a real extrovert He's a very fnendly mdividual

    who nnpresses acquamtances as bemg very personable andwell-adjusted At times he is qmte witty and entertaimng,and he nearly always displays a great deal of self-confidence

  • Favorableness of first impressions 347He IS prone to be enthusiastic, active, and generous with hispossessions When John's plans go awry, as they sometimesdo, he meets the setback with renewed stnvmg John isalmost always cheerful, and he refuses to let thmgs "gethim down"

    Unfavorable DescnptionJohn has always rehed on his family to take care of him

    He has never worked a day m his hfe, even when his parentswere havmg financial problems due to the poor health ofhis mother He expects his parents to protect him and re-gards this as their obhgation Although his parents canhardly aflFord it, he has let them pay for his schoolmgAlthough he demands so much of his parents, he is unwill-ing to give them even the shghtest adffection or gratitudeHe won't even try to succeed m school so that they might atleast take pnde m his accomphshments

    Ratmg Scales The response sheets contamed nme seven-pomtscales of the semantic diFerentiaI (Osgood, Suci, & Tannenbaum,1(957) Four of the scales (good-bad, positive-negative, valuable-worthless, and fair-unfair) were loaded on the evaluative factor andthey were used as measures of the Ss' impressions The other fivesemantic differential scales were mcluded for masking purposes

    Cancellation Task The cancellation task merely required Ss tocancel out the letter "e" m a passage taken from a philosophicaltreatise

    ProcedureThe data were collected m three separate sessions with the

    number of Ss participatmg m any one session rangmg from seven to25 At the begmnmg of the session, Ss were given booklets contammgidentical sets of mimeographed mstructions The first paragraphmformed the Ss that they were participating m an expenment con-cemed with mterpersonal perception and that the Es were mterestedm the sorts of impressions people form of others on the basis ofhmited mformation The second paragraph descnbed the generalcomposition of the booklets and the procedure to be followed musmg the booklets The thu-d paragraph provided detailed mstmc-tions for usmg the ratmg scales, emphasizing that Ss should attemptto mdicate their immediate feehngs about the person described andthat lookmg back and forth through the pages of the booklet wasnot permissible In the concludmg paragraph Ss were mformedthat they were taking a tuned test and that they should not openthe booklets or tum to a new page until the E gave the signal The

  • 348 May E Briscoe, Howard D Woodyard, and Marvin E ShawTable z Mean rating scores and impression change scores (Expen-ment I ) .

    Condition

    Interpolated UFInterpolated FUFinal UFRnal FU

    Control UControl F

    Time 1

    rating

    2 256 22

    2 446 20

    Time 2

    rating

    4 052 853 003 902 426 1 0

    Changescore

    1 803 38

    (0 56)*(2 30)*

    - 0 020 1 0

    Difference between Final Time 2 rating and the corresponding Control Time 1 rating

    order m which the vanous forms of the booklets were handed outwas randomized so that Ss were randomly assigned to the six ex-penmental conditions All Ss were given two mmutes to performeach of the tasks presented m the booklets Immediately after havmgmade their final ratings of John, all Ss responded to the four-itemquestiojmaire The results from the questionnaire were not enhghten-mg and will not be referred to agam

    REStTLTSThe mean rating scores and change scores for the six expen-

    mental conditions are given in Table 2 The first pomt of mterestIS the fact that the control group scores for mitial and finalratmgs are almost identical. This demonstrates that ratmgs arestable over time and that the cancellation task apparendy didnot mfiuence final ratings Secondly, the absolute deviation ofmitial and final ratmgs from the neutral pomt for Control Fand Control U conditions were not sigmficandy different, thusvenfymg the vahdity of the selection procedure Fmally, therewere no sex difffferences, either as mam effects or m mteractionwith other vanables

    We are now ready to consider the data relevant to our hy-potheses The impression change scores are of greatest im-portance for testmg the major hypothesis.

    The major hypothesis stated that an unfavorable first im-pression IS more resistant to change than a favorable first im-pression of another person. The chief test of this hypothesismvolves an analysis of the change scores for the mterpolatedand control groups Change scores were computed for each S

  • Favorableness of first impressions 349Table 3 Summary of analysis of variance of mean impression changescores for mterpolated response and control conditions (Experiment I)

    source

    Inihal information (1)Sex (S)Communication (C)I x SI x CS x CI x S x CError

    df MS

    6

    ' 2 2

    3 005 02)765 26

    )02}66

    F

    0 0098 49***

    1 147 96*0 050 00

    *revlsed p < 025**revised p < 005

    ***revi$ed p < 001

    m the Interpolated FU and Control F groups by subtractmgTune 2 scores from Time l scores Change scores for Ss m theInterpolated UF and Control U groups were computed by sub-tractmg Time 1 scores from Time 2 scores The mean changescore for each condition is given m column 4 of Table 2, and asummary of the analysis of vanance of these scores is given mTable 3 ^ Amount of impression change varied significantly withInitial Information (favorable vs unfavorable). Communication(experimental vs controls), and the Initial Information X Com-munication Interaction Commumcation accounted for the great-er part of the vanance, but the significant mteraction is of greaterrelevance for our hypothesis The Duncan range test (Duncan,1955) was used to make compansons among mteracbon means,showmg that the Interpolated FU condition mduced significantlygreater change than did the Interpolated UF condition (p