Calum M. Delaney
School of Health and Social Sciences, University of Wales Institute Cardiff , Western Avenue, Cardiff, CF5 2YB, United Kingdom
A study was carried out to investigate differences between a superordinate stereotype (relating to a hypothesised “typical” stutterer) and subtype stereotypes elicited after exposure to 4 stuttering individuals viewed on videotape. 22 participants rated the stutterers on a 6-item semantic differential scale. The results suggested that individuals may hold a variety of stereotypes towards people who stutter, varying in degree rather than the characteristics making up the stereotype, and these were not always negative. The results did not support a subtyping model used to explain stereotype change resulting from exposure to labelled individuals.
Explanations for the genesis of stereotypes as well as their modification have considered the role of exposure to individuals who stutter and to information about stuttering. In relation to stereotypes more generally Weber and Crocker (1983) described three possible models of stereotype change that might occur as a Consequence of Contact with a stereotyped individual or group of individuals, and how aspects of that Contact might influence that change. In the book-keeping model instances of disconfirming information challenging the stereotype gradually add to and modify the existing stereotype, resulting in a more complex and accura:e representa:ion of the stereotyped group. Fiske and Neuberg (1990) referred to this as piecemeal integration, and Bieri (1955) suggested that this might result in less stereotyping and more accurate predictions of behaviour. The results obtained by Weber and Crocker supported this model when disconfirming information was dispersed across many individuals. The conversion model suggests that change is dramatic rather than accumulative, and is the result of exposure to significant instances of disconfirming information that deviate strongly from the expectations aroused by the stereotype. The third model is the subtyping model, where stereotypes are viewed as being organised hierarchically (Brewer et al., 1981). In this model disconfirming information does not alter the superordinate stereotype of general characteristics that might be applied to an individual in the absence of any other information. Instead the information results in discriminations being made within the group, generating subtypes within the superordinate category. Johnson and Hewstone (1992) suggested that subtyping is most likely when the disconfirming information is concentrated in a few individuals. Subtyping thus allows an individual to maintain a stereotype of a group in spite of disconfirming information (Hewstone et al., 1994), especially if the subtype can be justified (Kunda & Oleson, 1995). Delaney (2002) offered some suggestions for how this subtyping model might help to explain some of the conflicting results of studies that aimed to alter stereotypes, and how it may be possible for an individual to continue to hold a (negative) superordinate stereotype in spite of exposure to specific individuals who stutter. In addition to suggesting that the nature of exposure to stuttering may influence the type and manner of stereotype acquisition, several studies have discussed aspects of the measurement of stereotypes and the role this might play in the results of previous studies. Hulit and Wirtz (1994) provided a useful critique of issues relating to the measurement of attitudes in relation to the disorder of stuttering. Many of the studies that have examined the negative stutterer stereotype have employed a variant of the semantic differential technique or an “attitudes toward stuttering” scale, and Hulit and Wirtz have commented on the validity of these measures and their dependence on interpretations based on personal perspectives. With regard to the use of the semantic differential scale there has been little discussion or investigation of the meaning attached by individuals to the pairs of bipolar adjectives, and whether the pole categorized as negative is seen as such by study participants. As an example, it might be argued that neither adjective of the pair “aggressive-passive” could be seen as positive. This will have implications for what a negative stereotype might mean. Additionally many of the studies on stutterer stereotypes note changes for some items and not others. Stereotypes may be made up of different components or groups of characteristics that might be differentially affected by particular experiences. The choice of measurement items, and the positive or negative values assigned to them, may thus have an important effect on the results of studies using these measurements.
There were two aims for the present study. The first was to investigate whether individuals might hold a variety of stereotypes (as measured using a semantic differential scale) towards people who stutter, and whether there might be evidence to support the notion of superordinate and subtype stereotypes as suggested by Weber and Crocker (1983). This aim was thus to compare a possible superordinate stereotype (or rating of a “typical adult male stutterer”) with ratings of 4 specific adult male stuttering individuals (possible subtype stereotypes) observed on a videotape recording. It was hypothesised that the superordinate stereotype would differ from the subtype stereotypes, and that these would differ from each other. Further, the superordinate stereotype would not alter as a consequence of exposure to specific individuals who stuttered.
The secondary aim was to examine the nature of the semantic differential scale measurement, and the meaning attaching to the measurement that is then termed a stereotype and how this might compare to other stereotypes. As a part of this aspect of the study two additional ratings were elicited of hypothetical “ideal” and “typical adult male” individuals.
The participants were required to carry out ratings on a 7-point semantic differential scale consisting of 6 bipolar adjectives drawn from the abridged Woods and Williams (1976) scale. Rather than rating each stutterer in terms of a single point on the 7-point scale between each pair of bipolar adjectives, the participants were required to indicate the range on the scale over which they felt each individual being rated was likely to fall. This provided a single range midpoint score and a range score for each of the 6 scale items. The range midpoint score provided a rating score on the positive- negative continuum of each item. The range score provided an indication of how specifically the participants felt able to locate their perceptions of the individuals being rated on each of the scale items. Consistent with the subtyping model it was hypothesised that this range would be greater for items describing the superordinate stereotype than for the subtypes.
The participants were required to rate eight individuals. Four of these were hypothetical individuals and the remainder were 4 adult males who stuttered and who were observed on videotape recordings. The 4 videotape recordings were obtained from programmes broadcast by public television stations, and the individuals recorded presented with moderate stuttering behaviours. Two of these demonstrated relatively positive attitudes towards themselves, their communication and their stuttering, while two presented relatively poorer attitudes.
The participants were initially required to provide a rating (as a single point on each scale item) of what they considered to be the optimum balance for each item, that is of an “ideal person”. The purpose of this was to provide an indication of a rating score that might be considered positive for each item, and against which other ratings might be compared. They then provided a rating of a hypothetical “typical adult male” followed by a rating of a “typical adult male stutterer”. This was followed by observing and rating in turn each of the individuals presented by videotape recording, one with positive attitudes, followed by two with negative attitudes, followed by the last with positive attitudes. They then completed a second rating of a hypothetical “typical adult male stutterer”.
Mean ratings across participants were calculated for each of the 6 scale items for descriptive purposes. Mean rating and range scores across the 6 items were calculated for each participant, and these were compared for differences. The Wilcoxon Signed Ranks test (2-tailed) was used to test the significance of these differences, applying Bonferroni adjustments to the desired P-level of 0.05.
Figure 1. Mean Ratings Across Participants
The ratings of the hypothetical and observed stutterers followed a similar pattern across the scale items, with the exception of the ratings converging near the ideal rating for the “passive- aggressive” item. The ratings of the hypothetical stutterer were more positive (closer to the Ideal) after the video observations than before, while the ratings of the observed stutterers also varied, the ratings of the individuals who expressed positive attitudes being more positive than those of the two who expressed negative attitudes. The ratings across the scale items for the Typical Male showed similar values, which meant they varied across the items in relation to the Ideal. For some of the scale items the ratings of some of the stutterers were more positive than those of the Typical Male, while for other items the ratings were more negative.
As the pattern of the ratings was similar across the hypothetical and observed stutterers, the mean of each participants’ midpoint ratings across the 6 scale items was obtained (a Rating Score) in order to compare the ratings across the different stutterers. A similar Range Score was obtained for each participant from the mean of the range of their ratings across the 6 scale items. The mean Rating Scores and Range Scores are presented in Table 1 and Table 2.
Table 1. Mean Rating Scores
The highest mean Rating Score was for the Ideal followed by the scores for the two observed stutterers who showed positive attitudes towards themselves, their communication and their stuttering (Observed Stutterers 1 and 4). The lowest score was for the participants’ rating of a hypothetical Typical Stutterer before observing the video-recorded stutterers followed by the scores for the two Observed Stutterers who expressed less positive attitudes. The mean Rating Scores for the hypothetical Typical Male and the hypothetical Typical Stutterer after observing the video- recorded stutterers were similar. The low standard deviations for all of the mean Rating Scores suggested that the participants rated the various individuals similarly.
With the exception of Observed Stutterers 2 and 3 (who expressed less positive attitudes) there was a significant difference (P=0.006 or better) between all the Rating Scores for the four Observed Stutterers (Wilcoxon Signed Ranks test, 2-tailed). There was also a significant difference (P=0.000) between the ratings of the hypothetical Typical Stutterer before and after observing the video-recorded stutterers. The Rating Scores for Observed Stutterers 1, 2 and 4 were significantly different to those for the Hypothetical Stutterer before the observation of the video-recordings (P=0.001 or better), and only the Rating Scores for Observed Stutterers 1 and 4 were significantly different (P=0.000 and P=0.007 respectively) to the Rating Scores for the Hypothetical Stutterer after the observation of the video-recordings. Thus there was a similarity in the Rating Scores for the Hypothetical Stutterer after observation and those of the two Observed Stutterers who expressed less positive attitudes.
Table 2. Mean Range Scores
Table 2 shows the mean Range Scores across participants for the hypothetical and observed individuals. These scores provide an indication of how specifically the participants felt able to locate their perceptions of the individuals being rated on each of the scale items. The means for the hypothetical individuals were greater than those for the individuals who were observed on the video-recordings. These differences were all significant at P=0.00l or better. In addition, the standard deviations of the Range Scores for the hypothetical individuals were also greater than those for the observed individuals, suggesting that the participants were able to locate their ratings more narrowly, and with less variation, for the specific individuals they observed than for more generally specified hypothetical individuals.
The ratings of an “ideal” individual suggested that not all scale items have the same value on a positive-negative continuum. Of particular interest was the “passive-aggressive” item. In previous studies the “aggressive” pole is typically taken to be the positive pole, which would explain why a “typical adult male” might be viewed more positively than a “typical adult male stutterer”, rather than the reverse asuwas found in this study. The use of the adjective “assertive” rather than “aggressive” also illustrates how the choice of item might alter the ratings obtained by the scale. Given the different patterns of stereotypes that might exist for different types of individuals, the potential effects of item choice on the stereotype obtained and the potential difficulties inherent in comparing the stereotypes in terms of how positive or negative they may be, it may not be meaningful to evaluate them in terms of whether they are positive or negative. Rather, stereotype research may provide us with useful information on how individuals who stutter are perceived, and some understanding of how those perceptions might function in interactions with people who stutter.
The significant differences between the mean Rating Scores of Observed Stutterers 1 and 4, and between them and Observed Stutterers 2 and 3, provide some support for the notion that individuals might hold a variety of stereotypes towards people who stutter, although the differences were more in terms of how positively these individuals were perceived than in terms of the constituent components of their stereotypes. In terms of the weighting of the stereotype measurement items in relation to each other, both the hypothesized and observed stutterers were rated similarly. This suggests that observing different individuals who stutter might not result in subtype stereotypes in the manner described by Weber and Crocker (1983), but that individual stutterers will not necessarily all be perceived in the same Way. Observers can discriminate between individual stutterers on the basis of relatively limited information about them, which might suggest a capacity for responding differentially to them as well. This might alleviate the concerns of some authors who have suggested that the existence of a “negative” stereotype held towards people who stutter, including by speech and language therapists, might compromise the management of this client group (Hulit & Wirtz, 1994; Lass et al., 1989).
Contrary to the expectation that the rating of the hypothetical “adult male stutterer” (being a measure of a superordinate stereotype) would not alter after exposure to specific individuals who stuttered, there was a significant difference between these ratings before and after the participants had observed the four video-recorded stutterers. Again, this finding does not support the notion that exposure to specific members of a stereotyped group does not affect a superordinate stereotype as suggested by Weber and Crocker (1983). Instead the results of this study support the book-keeping model where instances of information about specific individuals gradually modify the stereotype, although this alteration relates to the positiveness rather than the pattern of the stereotype. Contrary to some other findings (e. g. Doody et al., 1993; Leahy, 1994), these findings suggest that stereotypes can be altered by relatively limited exposure to individuals who stutter, albeit in terms of the positiveness rather than the pattern of the stereotype, and possibly only temporarily. In terms of this group of participants the exposure to the video-recorded stutterers improved the positiveness of their stereotype from a level below that of the worst rated Observed Stutterers 2 and 3, to a level approximately midway between their ratings of the four observed stutterers. The ratings of the two Observed Stutterers with relatively positive attitudes (l and 4) were significantly different to the rating of the Typical Stutterer after observing the video-recordings, whereas the ratings of Observed Stutterers 2 and 3 were not, suggesting that the rating of a Typical Stutterer after exposure was closer to that of the less favourably perceived Observed Stutterers. Again these results support the book- keeping model, where the exposure to specific instances of stutterers resulted in a general stereotype that more closely resembled the perceptions of the individual Observed Stutterers.
With respect to the Range Scores, the means and standard deviations for the hypothetical individuals were similar to and greater than those for the Observed Stutterers, which were also similar to each other. This suggests that the participants were able to locate their ratings more narrowly for the specific individuals they observed than they were for the more generally specified individuals, regardless of whether these were a hypothetical typical stutterer or typical male. Further the lack of a significant difference between the Range Scores for the Typical Stutterer before and after exposure to the Observed Stutterers suggests that the exposure did not increase the certainty of their perceptions about more generally specified stuttering individuals. While these findings might provide some support for a superordinate stutterer stereotype (elicited in relation to a “typical adult male stutterer”) and subtype stutterer stereotypes (elicited in relation to specific observed individuals), they may more accurately be a reflection of the amount of information available to the participants about the type of individual being rated.
Allport, G. W. (1958). The Nature of prejudice (Abridged). New York: Doubleday.
Bieri, J. (1955). Cognitive comp1eXity-simp1icity and predictive behaviour. The Journal of Abnormal and Social Psychology, 51, 263-268.
Brewer, M.B., Dull, V. & Lui, L. (1981). Perceptions of the elderly: Stereotypes as prototypes. Journal of Personality and Social Psychology, 4] , 656-670.
Delaney, C. M. (2000). Modification of the negative stutterer stereotype. In H-G. Bosshardt, J. S. Yaruss & H. F. M. Peters, (Eds.), Fluency Disorders: Theory, Research, Treatment and Self- Help. Proceedings of the Third World Congress on Fluency Disorders, Nyborg, Denmark (pp. 614-618). Nijmegen: Nijmegen University Press.
Delaney, C.M. (2002). Stereotypes and stereotyping of people who stutter: A new direction and an alternative rationale. In C. Williams & S. Leitiao, (Eds.), Journey from the Centre: Proceedings of the 2002 Speech Pathology Australia National Conference (pp. 105-111), Melbourne: Speech Pathology Australia.
Doody, I., Kalinowski, J ., Armson, J ., & Stuart, A. (1993). Stereotypes of stutterers and nonstutterers in three rural communities in Newfoundland. Journal of Fluency Disorders, 18, 363-373.
Dorsey, M., & Guenther, R. K. (2000). Attitudes of professors and students toward college students who stutter. Journal of Fluency Disorders, 25, 77-83.
Fiske, S.T. and Neuberg, S.L. (1990). A continuum of impression formation, from category- based to individuating processes: influences of information and motivation on attention and interpretation. In M.P. Zanna, (Ed.), Advances in Experimental Social Psychology, Vol. 23. London: Academic Press.
Franck, A.L., Jackson, R.A., Pimentel, J .T. & Greenwood, G.S. (2003). School-age chi1dren’s perceptions of a person who stutters. Journal of Fluency Disorders, 28, 1-15.
Hewstone, M., Macrae, C.N., Griffiths, R. & Milne, A.B. (1994). Cognitive models of stereotype change: (5). Measurement, development and consequences of subtyping. Journal of Experimental Social psychology, 30, 505-526.
Hulit, L. M., & Wirtz, L. (1994). The association of attitudes toward stuttering with selected variables. Journal of Fluency Disorders, 19, 247-267.
Johnson, L. & Hewstone, M. (1992). Cognitive models of stereotype change: 3. Subtyping and the perceived typicality of disconfirming group members. Journal of Experimental Social Psychology, 28, 360-386.
Kunda, Z & Oleson, K.C. (1995). Maintaining stereotypes in the face of disconfirmation: Constructing grounds for Subtyping deviants. Journal of Personality and Social Psychology, 68, 565-579.
Lass, N., Ruscello, D.M., Pannbacker, M.D., Schrnitt, J .F. & Everly-Myers, D.S. (1989). Speech- language pathologists’ perceptions of child and adult female and male stutterers. Journal of Fluency Disorders, 14, 127-134.
Leahy, M. M. (1994). Attempting to ameliorate student therapists’ negative stereotype of the stutterer. European Journal of Disorders of Communication, 29, 39-49.
Leyens, J -P., Yzerbyt, V., & Schadron, G. (1994). Stereotypes and Social Cognition. London: Sage Publications.
Snyder, G. J. (2001). Exploratory research in the measurement and modification of attitudes toward stuttering. Journal of Fluency Disorders, 26, 149-160. Section 8. Attitudes, Personality and Emotional Factors in Stuttering 45 9
Tajfel, H. (1981). Human Groups and Social Categories. Cambridge: Cambridge University Press.
Weber, R., & Crocker, J. (1983). Cognitive processes in the revision of stereotypic beliefs. Journal of Personality and Social Psychology, 45, 961-977.
Weisel, A., & Spektor, G. (1998). Attitudes toward own communication and toward stutterers. Journal ofFluency Disorders, 23, 157-172.
Woods, C.L. & Williams, D.E. (1976). Traits attributed to stuttering and normally fluent males. Journal of Speech and Hearing Research, I 9, 267-278.