Two Experiments on Helper's Reactions to Rejection of Their Help.

Rosen，sidney(University of Georgia) Cheuk，Wai Hing(University of Macau) Wong，Kwok Sai(University of Macau)

助人者被他人拒絕時的感受及反應


　　我的研究是集中於助人者的感受及反應。假設有一位明顯需要他人幫助的人，助人者向他/她提供幫助。但需要幫助的人則不接受幫助。在這時助人者將產生不安感受。產生不安的原因是那拒絕令助人者覺得自我能力或自尊心受損。他繼而作出思維上、情緒上及評價上反應，圖令自己相信自我仍是一個有能力的人。兩項採用學生相互輔導方法的實驗，得出的結果是支持上述論點：相對於幫助被他人接受者，被拒絕的助人者認爲他的被拒絕更可能是於對方過份的防禦；以及對方對助人者的評價比對其自己的評價還低等等。文章亦提及兩項涉及澳門教師的研究，內容是有關教師的幫助不被學生接受時的反應及感受。

Introduction


　　We first reported a model relevant to　understanding how lay and professional helpers react when their offer of help is rejected by the recipientd of the offer. This　would be followed by two esperiments that　esamined major aspects of the model. In the　last section we described two studies currently in progress that extended the model　to encompassing long-term reactions of professional help-givers, including teachers.
　　Our model proposes that rejection of an　offer of help by a recipient who appears to　need help is experienced as a violation of the　helper's expectations of a favorable outcome, i.e., of acceptance of the offer. The　model suggests that this disconfirmed expectancy has unfavorable implications for　the would-be helper's self-image. Namely, the disconfirmation casts doubt on the helper's beliefs in his or her own interpersonal　efficacy and/or prosocial concerns, depending in part on how much importance this　helper attaches to the outcome of the offer. At this stage of theory development, these　self-image implications comprise an inferred　cluster of intervening variables.
　　The model further proposes that the helper tries to cope with these self-doubts　through various reaction modalities, affective, evaluative, cognitive, and behavioral. Finally, the model proposes that there are　both personal, i.e., individual difference factors, and situational factors that serve as　moderator variables that can influence the　perceived importance of the outcome and　the magnitude of expectancy violation.
　　To date, our basic laboraory paradigm　has used the cover story that we are exploring whether peer tutoring can successfully　supplement the alledged College Office ofRemedial Instruction's tutoring program. Each prospective helper is recruited from　the introductory psychology research participant pool in exchange for participation　credit. The helper is asked to administer a　practice task or two on word construction to　a so-called Learner. The Learner, the Prospective recipient (R), who is in reality a confederate, always fails to complete them on　time. Then the helper is coaxed to offer the　Learner some 3x5 index cards containing　helpful rules for word construcrion. and is　even invited to add rules if the helper wishes　to write some.As prearranged, the Learner　either rejects or accepts the offer that nearly every helper makes. After this, the helper　completes a series of questionnaires that address the dependent variables, in the mistaken belief that the Learner's test tesk will　come next. The following presents the gist　of two experiments that examined major aspects of the model.

Experiment 1


　　In one experiment we varied the outcome, i.e., rejection or acceptance of the offer, by way of replication. We also introduced as a possible situational moderator　the recipient's need level, by representing　the Learner eirher as a remedial student referral, or as a typical or control student picked at random from the undergraduate　student body. Need level could be thought　of as contributing to outcome importance. About a week before any individual experimental session was run, nine individual difference measures were group-administered　to our prospective helpers.
　　Expectancy violation. The results presented in Table 1 show that rejection of the offfer is experienced as an expectancy violation.
　　Insert Table 1 here
Specifically, the helpers were far more likely　to report that they were surprised by the　Learner's response to their offer if the offer　had been rejected than if it had been accepted. They were also much less likely to say that the surprise was pleasant, if the offer had been rejected.
　　Inset Table 2 here
　　Outcome main effects. The means in Table 2 are based on standardized scores. The　means are all ordered in the appropriate direction. As we expected, the overall reactions to the outcome of the offer took on a　negative cast if the offer was rejected: Rejected helpers, in contrast to their accepted　counterparts, expressed more negative affect, more negative evaluations of R's competence and sociability, and tended to be relatively unattracted to the Learner. They　also came up with a relatively uncomplimentary portrait of the Learner in explaining why the offer was rejected. One cluster　of these causal attrbutions we labelled defensiveness: it included such characterizations as stubborness, being too proud, and　being concerned about appearing inferior. We speculated that these rejected helpers　were projecting their own defensiveness　onto the Learner. Another cluster we called　perverse counter-dependency, to suggest　that, in the helper's judgment, the Learner　wanted to appear independent, even though　it was obvious that the Learner needed help　(a sort of attributed self-deception). Rejected　helpers, compared to accepted helpers, were　also more likely to say that the Learner's　poor task performance was due to the Learner's personality and general outlook, i.e., it　was due to stable-internal attributes. They　also gave a relatively gloomy prognosis concerning the Learner's future performance: they maintained that the Learner had learned relatively little from the present experience: and they maintained that they thmselves has contributed very little to that learning, as though to disclaim responsibility　for the consequences that would befall the　rejecter.
　　Hindsight bias. Helpers were asked toward toward the end of the session to what　extent they has thought it likely that the　Learner would accept their offer at the time　when they were making the offer. We werelooking for evidence of hindsight bias as a　cognitive coping mechanism in the present　context. One way of coping with the threat　that rejection posed to their self-image is　for rejected rather than accepted helpers to　assert after the fact, “I knew all along that　the offer would be rejected.” Such a main　effect is precisely what happened. Specifically, accepted helpers stated afterwards　that they had thought the likelihood　of acceptance was hihg. More important, however, rejected helpers stated that they　had considered the likelihood of acceptance　low.
　　Two rival rationales have been suggested in the literature as explanations for hindsight bias, one cognitive, the other　motivational (Leary, 1981, 1982). The cognitive version would go something like this: Because the causal structure leading up to the　critical event is not readily available, the person uses the critical event to construct or　reconstruct that causal structure so as to　make the event appear inevitable. Leary　(1981, 1982), in two conceptual replications, tried to test the motivational alternative, by　investigating, whether high egoinvolvement　or high eslf-presentation concerns, or both　could account for hindsight bias. Neither　one did. Leary concluded that may be a　cognitive explanation was the only viable　alternative. We feel that that conclusion　may have been premature. The following　shows where motivation might operate in　our model in bringing on the postdicted outcomes.
　　A plausible locus for relevant motivational factors is the set of factors we have called “Personal Moderators.” The rationale is　that would-be helpers who are chronically　disposed to see themselves as highly efficacious persons and/or as having hihg prosocial　concern for others, have more self-investment in the outcome of the offer than　would their low counterparts. That is, the　outcome of their offer is more important to　them, and a rejection would therefore pose　a greater threat to their selfimage, as evidenced from the effects of the dispositional　measures that had been collected earlier on　these helpers.
　　Insert Table 3 here
　　The first four measures in Table 3 could　be considered facets of a chronic self-as　efficacious outlook; they are Perceived Own　Competence, Masculinity, Personal Efficacy, and Interpersonal Control. The next four　measures could be considered facets of a　chronic self-as-prosocial outlook; they are　Perceived Own Sociability, Femininity, Perspective Taking, and Empathic Concern. Competence and Sociability are indices that　we have developed and used successfully before but as dependent variables. Masculinity　and Femininity are form the PAQ battery　(Spence, Helmreich, & Stapp,.1974). Personal　Efficacy and Interpersonal Control are from　Paulhus' Spheres of Control battery (1983). Perspective Taking and Empathic Concern　are from Davis' battery of empathy scales　(1983). The last two rows represent composites based on the first four and the next　four dispositional measures, and that we've　labelled self-as-efficacious and self-as-prosocial, respectively.
　　The means for those who are high on　these measures show more polarized postdiction than do those who are low. The Highs　claim that they knew all along that the needy R would accept help when in fact if was　accepted. That claim is quite logical, and　need nor necessarily constitute hindsight　bias. For instance, acceptance of aid from　helpers who generally harbor high expectations of interpersonal control, could simply　be just another confirmation which may be　their long history of interpersonal successes;such helpers would almost certainly have　predicted acceptance, and probably have　been correct. In contrast, however, for them　to postdict that they knew all along that the　needy recipient would reject their offer, is　hardly logical. That they did so, provides support for the motivational explanation of hindsighr bias. Notice that this pattern occurs　on all 8 dispositional measures, and of cou-rse on the two composites.
　　Recall now, that we manipulated recipient need level as a possible situational moderator of reactions to the outcome. Its performance as a moderator was disappointing　for the most part. Postdiction, however, exhibited interesting changes when need level　together with the personal moderators are　brought to bear. Suppose our reasoning is　correct that those who are high on these　dispositional factors have relatively more of　a stake in the outcome. If a peer whose need　for help was relatively small were to accept　their help, this would be most supportive of, if not flattering to, their self-image. On the　other hand, if a person whose need. for help　was great were to reject their help, this　would be most threatening to their self-image. Therefore, these dispositional Highs　should show greatest postdicted acceptance　after their offer is accepted by the typical　learner, but greatest postdicted rejection　after their offer is rejected by the remedial　Learner.
　　Insert Table 4 here
　　As shown in Table 4, with one exception, those who are high on these dispositional　moderatos were likely than their low counterparts to postdict acceptance from their　typical Learner, if the offer had been accepted. By contrast, on all 8 dispositional　moderatores those who are high were more　likely than their low counterparts to posdict　rejection from their remedial Learner if the　offer had been rejected. For such dispositional highs, to say that they knew all along　that this remedial person would reject their　offer is quite illogical. It is consistent, however, with the motivational interpretation of　hindsight bias. The form that hindsight bias　took in the present context might also be regarded as a manifestation of self-serving　bias.

Experiment 2


　　In a later experiment we modified our　peer-tutoring paradigm by presenting prospective helpers with a table of bogus performance norms allegedly based on a large sample of freshmen. The table indicated that　the average freshman was able to complete　six words on time on two successive tasks. Helpers were told that if their learner failed　both practice tasks and showed little improvement from the first to the second, theri　the learner probably needed help. It was also　intimated that due to random assignment　within each pair of research participants, they themselves were selected to be the tutor whereas their partner was selected to be　the learner. The learner, of course, did not　improve, and failed both practice tasks.
　　Expectancy vioation as a mediator. According to tne model, if expectancy violation is　truly a mediator in this context, then controlling for its influence should substantially　reduce the effects of the outcome on he lper's reactions. As one test of this, we carried out a series of Analyses of Covariance　(ANCOVAS) on each of our major dependent　variables. Our measure of expectancy violation, i.e., perceived surprise, served as the covariate in each analysis.
　　Insert Table 5 here
　　For each reaction modality, the least　square means and associated values in the　first row are based on an Analysis of Variance. They show the main effect of the　outcome of the offer on the reaction modality in question. The second row shows the　relevant adjusted means and associated va lues after the ANCOVE is performed. In the　case of outcome attributions, we came up　with two new factors besides Defensiveness, as helpers' explanations for the rejection. We　labelled the first new factor Chronic Reticence. It consisted of chatacterizations of　the Learner as shy and easily embarrassed. The other one, which we labelled Situational　Insensitivity, consisted of descriptions of the　Learner as lacking awareness of needing　help and of not wanting to be sociable.
　　Notice that on all five measures, if the　influence of perceived expectancy violation　is controlled, the impace of the outcome is　either eliminated or is substantially reduced.
　　Effects of outcome on self-other evaluations. Another objective of this experiment　was to demonstrate empirically something　that had previously only been assumed to　occur. Namely, our aim was to show that rejection abversely affected the helper's selfevaluation, but that helpers would cope with　the underlying threat by evaluating the recipient even more negatively after being rejected than they evaluated themselves.
　　Insert Table 6 here
　　These predictions were supported; larger　means in Table 6 denote more negative　evaluations of the target person. The interaction effect take the following form:
　　Insert Figure 1 here
　　Although rejection makes for more　negative evaluations of self and of recipient　that does acceptance, negative evaluation of　the recipuient increases much more that　does negative self-evaluation. Thus, one way　in which the helper's self-image is maintained following an unfavorable outcome is　via self-other comparisons that still favor　the self.
Other Work in Progress
　　The role and performance of teacher in　the process of education are generally regarded as extremely important. Yet how teachers in Macau perceive their role, what factors affect their performences, and how　satisfied they are with their work and what contribute to such have yet to be systematically explored. We believe that one of the factors that may contribute to teacher satisfaction is the extent to which students accept the advice and help from the teachers.
　　We are in the process of collecting relevant data from teachers. in Macau. The major　objecive is to determine to what extent students' refusal of their help, or students, resistance to being helped by them contribute　to burn-out, an indication of job satisfaction, in these professionals. Our model sees coping as a result of expectancy violation. We　are trying in another study with teachers　freshly graduated from the School of Education, University of Macau, to determine to　what extent the professional training that　these teachers had received functions as a　buffer that enables them to cope actively　with perceived student resistance. It seem　plausible, for instance, that the training they　had received would prepare them to anticipate resistance of rejection as a common　occurrence, thus implying that the fault may　not lie in these helpers. If so, their morivation for posdicting rejection is reduced.
　　References
　　Leary, M.R. (1981). The distorted nature　of hindsight. Journal of Social Psychology, 115, 25-29.
Leary, M.R. (1982). Hindsight distortion and　the 1980 presidential election. Personality　and Social Psychology Bulletin, 5, 257-263.


　　Note. N=148. Means are based on standardized scores.
Magnitude was measured by a 7-point scale on which 7 indicated that the rater was very surprised at the　recipient's response to the offer. Valence was measured by a 7-point scale on which 7 rneant that the surprise (however little), was very pleasant and 1 meant very unpleasant. Both differences are significant at P <. 0001.


　　Note. R = recipient. Except as noted, all outcome differences are significant at or beyond the .05 level. N = 148. Means are least square. Scores on attributed defensiveness, counterdependency, and causes of R's perforemance are negatively keyed such that low (negative) scores indicate high defensiveness, high counterdependency, and perforemance due to internal-stable (chatacterological) causes.



　　Note. High mean scores denote high postdicted acceptance. Low mean scores denote low postdicted acceptance (or high postdicted rejection). The self-as-efficacious composite index is based on the mean standardized scores of individuals in the top scores of individuals on the bottom four personal moderators.



　　Note. High means denote high postdicted acceptance. Low means denote low postdicted acceptance (or high　postdicted rejection). The self-as-efficactious and the self-as-prosocial composites are based on the mean　standardized scores of individuals on the top four and bottom four personal moderators, respectively.



　　Note. Analyses on first pair of (least square) means for each reaction modality are based on ANOVAs. Those　on the second pair of means are based on ANCOVAS with
expectancy violation as the covariate. The latter means are adjusted means. N = 67.