|Home | About | Journals | Submit | Contact Us | Français|
Most models that explain the onset and development of stuttering include a social and emotional component. This paper has two intentions. One is to review the methods and findings of previous research that investigated the role of affective and social factors in stuttering. The second intention is to alert readers to various methods and issues being applied in social psychology to investigate these phenomena and to indicate where these methods could be useful in assessing the role of social and affective components in stuttering.
Stuttering affects approximately 5% of the population at some time in their life according to work in the United States (Conture, 1996). It is also reported that the disorder affects children disproportionately. The usual age at which the problem starts (onset) is between three and five years (Dalton & Hardcastle, 1977). Eighty per cent of young children who are diagnosed as stuttering recover to normal fluency during school years (Starkweather, 1985; Yairi & Ambrose, 1999). Around one in a hundred of the adult population persist in their stuttering (Andrews & Harris, 1964; Bloodstein, 1987). It is important to know what role social factors, and their changing role over development, play in stuttering. This is because speech is a social phenomenon as people speak to one another about various topics in a variety of situations. Also, stuttering has been shown to be topic- and situation-specific and the disorder is governed in part by affective factors that are socially moderated.
The disorder also affects language and motor performance. Historically, the evidence that showed deficits in these areas of performance led to accounts which maintained that stuttering arises either because of psychological, physiological, linguistic or learned behaviors. In contrast to such views, accounts have emerged over the last quarter of a century, that maintain stuttering is a multifaceted speech disorder that involves all the preceding factors and, in addition, social ones. There have been several models that approach stuttering from a multifactor perspective, most of which include a social or emotional component. One of the earliest multifactor models of stuttering (Zimmerman, 1980) emphasised the importance of the interaction between motor speech behavior and a range of emotional and environmental conditions in the development and maintenance of the disorder. Wall and Myers (1984) suggested that psycholinguistic factors, psychosocial factors (i.e. discourse loads and interactions with parents and/or peers) and physiological components interacted to cause and maintain stuttering. The demands and capacities model (Starkweather, Gottwald & Halfond, 1990) views the onset and development of stuttering as related to a mismatch between the child’s capacities (motor, linguistic, cognitive and emotional) and self-imposed or external speech demands. The models proposed by Smith (1999) and de Nil (1999) concentrate on the importance of disrupted speech processes and their relationship with social, emotional and learned factors. All these models suggest that cognitive, linguistic and affective factors influence speech motor functions. Riley and Riley (2000) revised their 1979 stuttering assessment instrument and maintained that speaker temperament factors and listeners reactions to people who stutter were two of the three main factors that contributed to the onset of stuttering.
Multifactor models describe how stuttering might start and/or be maintained and, by implication, indicate that components identified in the models should be included in the identification and treatment of the disorder. However, with the possible exception of the Riley and Riley work, the models lack the detail necessary to make them useful in the collection and interpretation of assessment and treatment outcome data (Healy, Trautman & Susca, 2004). The influence of social and emotional factors in the onset, development and treatment of stuttering is ubiquitous in the models. The case for inclusion of social aspects in stuttering is aided further by confirmation of influences of social variables on stuttering (in both onset and treatment) using the most rigorous techniques applied in this area.
This paper has two intentions. One is to review the methods and findings of previous research that has investigated the role of affective and social factors in stuttering. This will constitute an examination of the research that investigated differences between people who stutter, PWS (children and adults) and people who do not stutter and will also look at research investigating how others perceive PWS. This will include studies of how PWS (children and adults) are perceived by fluent children and adults (including clinicians and parents). A common finding of past research is that PWS are stereotyped as being more guarded, nervous, self-conscious, tense, sensitive, hesitant, introverted, and insecure than speakers who do not stutter (Klassen, 2001). In cases where there is an influence, they can be included in multifactor models that address the onset, development and treatment of the disorder. The second intention is to alert readers to various methods and issues being applied in social psychology to investigate these phenomena and to indicate where these methods could be useful in assessing the role of social and affective components in stuttering.
Social factors are examined around the start of school, at adolescence and in adulthood. These age groups are looked at separately as different social factors kick in as speakers get older. Thus the factors that will be considered with pre-school and start of school children are intelligence, personality, attitudes and temperament, bullying can occur once a child is established in a school and progress through adolescence and probably into adulthood and occupational matters specifically affect adults. Factors affecting one of the early age groups can have an influence through life and such factors have sometimes been investigated in later age groups. Thus, though intelligence exerts an effect on young CWS, the question can also be raised whether it also affects adolescents or adults. Factors are considered in the age groups where they may first exert an influence, but evidence on the same factor in older age groups is also considered so developmental trends can be established. The target group at the first two ages is referred to as children who stutter (CWS) and the older age group as persons who stutter (PWS). All PWS is used to refer to all ages. Children who do not stutter (CWNS) and adults who do not stutter (PWNS) represent controls. CWS have been evaluated by peers (CWNS) and by adults (PWS or PWNS). Adult PWS have also been quizzed retrospectively about how social influences affected their stuttering in their childhood. An attempt has been made to distinguish these different options in the studies reviewed.
Differential psychologists distinguish between two major variables: ‘Abilities’ measured by tests of statistical power (how robust the statistical findings are) and ‘personality’ measured by tests of preference. The latter is usually further sub-divided into beliefs (attitudes, attributions, values) traits (types, disorders) and coping styles. To get a full understanding of the role of individual differences in stuttering, it is important to understand all aspects of individual differences that may act as moderator variables on stuttering.
In the intelligence researcher community the world is divided into ‘lumpers’ and ‘splitters’. The former believes in general, the latter in multiple, intelligence. All intelligence tests, particularly those of crystallized intelligence, have verbal tests (usually vocabulary tests). Indeed, vocabulary is very highly loaded on ‘g’ that measures general intelligence (REF). It is, therefore, a valid hypothesis that a person with low IQ manifests in part by low vocabulary may be particularly frustrated or anxious in social settings when they are required to articulate.
Empirical evidence suggests that CWS score significantly lower on intelligence tests than do fluent controls. Studies with school-age CWS (Andrews & Harris, 1964, Okasha, Bishry, Kamel & Hassan, 1974, Schindler, 1955) indicated that this deficit is evident in both verbal and non-verbal intelligence tests. As non-verbal intelligence is affected, it appears unlikely that these performance deficits could be explained by difficulties in communication because of stuttering. In contrast, the intelligence and social class of CWS who are receiving treatment is found to be above average (Andrews & Harris, 1964, Cox, 1982). Though a developmental change could be responsible, the difference between younger and older CWS could also be explained by the influence of intelligence and social class on access to health care.
The intelligence of all PWS as perceived by others appears to change across age groups. Franck, Jackson, Pimentel and Greenwood (2003) used bi-polar adjective pairs taken from Freeby and Madison (1989) and Wencker et al (1996). Examples of adjective pairs used were intelligent-stupid and competent-incompetent. Fluent children aged 9 - 11 years were asked to view videotapes of an adult who stuttered and a fluent control. They found that the CWNS used adjectives that indicated they judged the intelligence of the PWS more negatively than the PWNS. However Craig, Tran and Craig (2003) reported that telephone interviews with 502 adults indicated that a large number of the respondents (none of whom stuttered or had any interaction with CWS or PWS) believed that PWS had average or above average intelligence. Taking the results together, there seems to be a difference between the way children perceive PWS (less intelligent) and the way adults perceive PWS (more intelligent).
Personality traits have been shown to relate to second language learning and work productivity (REFS). Indeed they have been shown to account for a tenth to a third of the variance on a wide range of factors like academic success, work productivity and health. It seems quite plausible that traits, particularly extraversion and neuroticism, both as main effects and in interaction, relate to speech fluency and stuttering. Indeed fluency versus disfluency may affect, in turn, the personality of young people. What, however, is important is a description of the process by which stable traits relate to problems in language production.
Early descriptions of stuttering regarded it as a manifestation of emotional disorder in childhood. Considerable effort was invested in collecting data on the personality attributes of PWS and their tendency to show anxiety or neurotic symptoms that would support such descriptions. The work used a variety of measurement instruments, and indicated that there were no differences between school-age children who stutter and controls in personality factors related to neuroticism or anxiety. This conclusion is based on tests with the Sarson General Anxiety Scale for Children and the Structured Psychiatric Interview (Andrews & Harris, 1964); Eysenck Personality Inventory (Hegde, 1972); California Test of Personality (Prins, 1972; Minnesota Multiphasic Personality Inventory (Horlick & Miller, 1960, Lanyon, Goldsworty and Lanyon, 1978, Pizzat, 1951); and the Speolberger Anxiety Scales (Molt & Gifford, 1979).
Manning, Dailey and Wallace (1984) found that self-perceived personality characteristics of older PWS (29 participants, mean age 62 years) were not significantly different from those of nonstuttering controls (13 participants, mean age 65 years). They used a bi-polar adjective scale (Wood & Williams, 1976) containing pairs such as anxious-composed and introverted-extroverted. The PWS had a tendency to see themselves as more inflexible, withdrawn, self conscious, anxious and introverted than the control group, however no significant differences between the groups were reported.
Personality alone does not appear to be a predictor of the developmental pathway of stuttering. Guitar (1976) showed that neither neuroticism or extraversion as measured by the Eysenck Personality Inventory (Eysenck & Eysenck, 1963) were, by themselves, significant predictors of recovery or persistence. However Guitar did conclude that a combination of pre-treatment factors (e.g. personality traits, % stuttered syllables, attitudes) was useful in predicting the outcome of speech therapy. His study indicated that a combination of measure of several pre-treatment factors taken on 20 adults who stutter was highly correlated with post-treatment speech measures (% stuttered syllables, % change in frequency of stuttering). Guitar (1976) concluded that, of the personality traits, only neuroticism was strongly related to outcome measures. However, in this study neuroticism was also found to significantly relate to attitude measures.
Other studies have looked at the way the personality of PWS is perceived (as opposed to measuring what it is actually like). For instance, PWS are generally stereotyped as more nervous, shy, withdrawn, tense and anxious (Horsley & FitzGibbon, 1987). The Franck et al (2003) study mentioned above used the adjective pairs from Freeby & Madison (1989) and Wencker et al (1996) and included pairs such as outgoing-shy and relaxed-tense. They found that fluent school-age children rated adults who stutter more negatively than controls on a series of personality traits. However, unlike the rating of the intelligence of PWS, the negative perception of the personality of PWS continues into adulthood. Such stereotypes have been found to be held by a wide range of groups, from members of the general public, including college professors and teachers and even clinicians (Dorsey & Guenther 2000; St Louis & Lass, 1981; Yeakle & Cooper, 1986) as well as the parents of CWS (Fowlie & Cooper, 1978).
Wood and Williams (1976) found evidence for a strong negative stereotype of CWS. Using an analysis of responses to bi-polar scales derived from words previously judged by speech clinicians to be descriptive of CWS (e.g. nervous-calm, afraid-confident) they found the stereotype to be predominantly unfavorable. Wood and Williams (1971) found that even people with professional experience of working with disfluent patients attributed undesirable characteristics to CWS. When asked to list adjectives to describe CWS, approximately 75% of clinicians used words that were grouped within the category of “nervous and fearful” and 64% listed words that were included in the category “shy and insecure”.
Looked at from the broader social psychology perspective, De Waele and Furnham (1999), in a meta-analysis of the role of trait-extroversion, found that linguists were very out-of-date in their understanding and measurement of personality traits. Also they were seriously misinformed about the role of personality in language. De Waele and Furnham’s own analysis showed clear and consistent evidence of the role of extraversion in language learning and production. Another important limitation in many studies is that many researchers examining stuttering have not looked at interactions. The interaction of extraversion, neuroticism and intelligence may well show powerful effects where looking at these factors alone does not.
The best way to determine the role of personality variables would be incrementally through block stepwise regression. Thus, assuming one has a sensitive and robust criterion variable like degree of stuttering or some measure of recovery, one could regress demographic variables first (age, sex, class), then intelligence, and finally personality measures using the big five personality dimensions (as main and interaction effects). This would then demonstrate if and whether personality traits accounted for any incremental variance in explaining stuttering-related variables.
The attitudes and reactions of all PWS to interpersonal verbal communication (communication attitudes) have been regarded as constituting a basic component of stuttering for many years (van Riper, 1948; Johnson, Brown, Curtis, Edney & Keaster, 1956; Travis, 1957; Sheehan, 1970; De Nil & Brutten, 1991; Vanryckegham & Brutten, 1996). Several studies have produced evidence that the communication attitudes of adult PWS are more negative than those of adult PWNS (Brown & Hull, 1942; Erikson, 1969, Andrews & Cutler, 1974).
According to Costello (1984), few attempts have been made to assess the attitudes of CWS. Two reasons suggest that measuring attitude in pre-school children is important: i) The most important point that applies to speakers of all ages is that attitudes have been shown (in particular circumstances) to be causally-related to behavior. ii) Stuttering usually starts between three and five years (Dalton & Hardcastle, 1977) Attitudes about many things change dramatically between ages 3 and 5 (Perry, Bussey & Fischer, 1980). This is especially likely to be the case when a child has a problem that affects overt behavior. An observation which may be relevant to effective treatment is that attitudes are more easily changed during or close to their formation (Niven, 1994). Though all these points underline the importance of having procedures and measured of attitude in young people who stutter, methodologically the development of appropriate instruments is not straightforward.
One attempt was made by Brutten (1985) who developed the Communication Attitude Test (CAT) in order to determine if the speech-related attitudes of CWS differed from those of CWNS. A Dutch version of the CAT (CAT-D) was developed in a series of studies that aimed to establish if the communication attitudes of CWS were significantly different to those of CWNS for speakers of this language (De Nil & Brutten, 1986; 1991). These studies revealed that CWS scored significantly higher on the CAT-D than their peers who did not stutter (i.e. CWNS), indicating that their speech-related attitudes were more negative. Similar between group differences were found with a group of American children by Boutsen and Brutten, (1989). The internal, and test-retest, reliability of the CAT and CAT-D has been demonstrated in several studies (Brutten & Dunham, 1989; Vanryckeghem & Brutten, 1992; Vanryckeghem & Brutten, 1992).
Vanryckeghem (1995) proposed that the CAT and CAT-D are useful clinical and research tools for evaluating between group differences when investigating communication attitudes. However she considered that the scope of the CAT is limited in that it requires a child to have the ability to read and understand the concepts covered by the test items and consequently it is not generally accurate when used with children younger than 7 years of age.
An instrument capable of determining the communication attitudes of children close to stuttering onset would be useful in several areas. In research it would assist in pinpointing the exact role of communication attitudes in the onset and development of stuttering. There have been proposals that the onset of stuttering is a result of the belief that speech is difficult (Bloodstein, 1987; Brutten & Dunham, 1989). Diametrically opposed to this viewpoint, are theories that have proposed that the negative beliefs that PWS have about speech are a product of, rather than a cause of their dysfluency (Guitar, 1976; Peters & Guitar, 1991). In terms of clinical efficacy Shearer (1961) and Erikson (1969) indicated that changes in the self-concept of someone who stutters is an important aspect of success both during and following treatment. An instrument that offers an indication of attitude in young children might also be useful for early diagnosis and intervention (Yairi & Ambrose, 1992: Onslow, 1994).
Attitudes to stuttering are comparatively rare in older speakers but there are at least four areas of research into attitudes that merit further work in older (as well as younger populations). First, there are the attitudes of speakers who stutter to themselves and to other speakers who stutter in social situations in which they stutter. Second, the attitudee of speakers who stutter to social communication in general need examining as they might yield patterns of situational phobia or anxiety. Third, there are the attitudes of various groups - parents, teachers, peers and the public - to those with different patterns and degrees of stuttering. Fourth, and perhaps most importantly, models like the theory of planned action need to be employed to determine how, when and why attitude in observers relate to their behavior towards them. Ideally this would involve costly, but ever important, longitudinal research.
Many authors, such as Sermas and Cox (1982), hold that a child’s temperament is a contributing factor to both the development and maintenance of stuttering. Temperament is distinguished from traits being much more physiologically based. Theorists in the last decade or so have suggested that CWS exhibit a more vulnerable or sensitive temperament, which could possibly be a contributing factor in the development, maintenance or chances in recovery of stuttering (e.g., Conture, 2001; Guitar, 1998; Zebrowski & Conture, 1998). As well as temperament as a whole, it has been suggested that particular aspects or dimensions encompassed in temperament play a role in stuttering, for instance attending problems, which refer to distractibility, perseveration, inability to concentrate on tasks and low frustration tolerance (Riley & Riley, 2000). In the few studies that have addressed stuttering and temperament directly, a number of dimensions have been highlighted as differing in CWS relative to CWNS. In general, CWS tend to be more responsive or reactive to stimuli in their environment (Wakaba, 1998), and are more sensitive, anxious, withdrawn and introverted (Fowlie & Cooper, 1978). These findings support the speculations of Bloodstein (1995) and Guitar (1998) that CWS have a more sensitive temperament leading to greater reactivity to unfamiliar, challenging or threatening situations, thus supporting the notion of multiple dimensions of temperament.
Embrechts, Ebben, Franke and van de Poel (2000) assessed temperamental dimensions using parental reports on the Children’s Behaviour Questionnaire (Rothbart & Bates, 1998), and found that parents rated their child who stutters as displaying reduced attention span and less success in adapting to new environments. More recently, Anderson, Pellowski, Conture & Kelly (2003) found, using the Children’s Behaviour Questionnaire, that CWS were significantly less likely to adapt to change, were less distractible and displayed greater irregularity with biological functions.
However, there are discrepancies in findings in terms of what dimensions CWS differ from CWNS. In the Embrechts et al. (2000) and Anderson et al. (2003) papers, two dimensions were found to be significantly different between CWS and CWNS: First, CWS were found to be significantly less adaptable that their controls. A second difference is in terms of distractibility. Here, however, though this dimension was found to be significantly different in both the studies cited, the direction was not the same; in Embrechts et al., CWS were more distractible than the controls, but in Anderson et al., they were found to be less distractible. In addition to these two dimensions, Anderson et al. found that CWS were significantly less rhythmic than CWNS in terms of the rhythmicity dimension, whereas Embrechts et al. did not. In a paper by Howell, Davis, Patel, Cuniffe, Downing-Wilson, Au-Yeung and Williams (in press), four dimensions were found to differ significantly. As was the case for both Anderson et al. and Embrechts et al. studies, CWS were found to be non-adaptable. However, none of the remaining three dimensions matched in direction across this and the two cited studies; Howell et al. (in press) found, in particular, that CWS were significantly more active, more negative in mood, and less persistent than CWNS.
Future work should examine Eastern European models of temperament, it is possible to test both cross-sectionally and longitudinally whether temperament variables relate to stuttering. Temperament variables, like ‘strength of the nervous system’ might lead to problems in speech production, particularly in stressful situations. Temperament factors might link, then, to physiological predispositions to stuttering. Furthermore, temperamental factors might indicate the likely success or otherwise of effective treatments (assuming disorders with physiological underpinnings are less likely to be treated effectively). Surprisingly no research as yet has looked at whether temperament predicts the onset and likelihood of recovery in children who stutter.
Two fundamental questions are 1) whether speakers who stutter are bullied and 2) whether bullying exacerbates the problem of stuttering? There only appears to be evidence on the former.
Parker and Asher (1987) reviewed work on bullying and concluded peer rejection and bullying can have severe and long-lasting effects such as low peer acceptance or peer rejection and their influence on later personal adjustment problems such as depression and early school dropout. Hodges and Parry (1996) identified three peer-related factors that increased the risk of a child being bullied - few friends, low-status friends and rejection by peers. O’Moore and Hillery (1989), Martlew and Hodson (1991), Nabuzoka and Smith (1993) and Whitney, Smith and Thompson (1994) have all reported that children with special educational needs are more susceptible to bullying than their peers, and are more likely to have few friends and be rejected.
The reason for thinking social acceptance might affect children who stutter is that they are often reluctant or unable to participate verbally in school activities (or social groups in general). In turn, this may lead them to be seen as shy or withdrawn and possibly, because of these perceived characteristics, to have difficulties in peer relationships making them targets of bullying. There is some previous research investigating sociodynamic factors and their relationship to stuttering and this is arranged for review under the techniques they have used (ratings, retrospective, sociometric). It is apparent that the majority of this research was carried out over a quarter of a century ago and, although useful from an historical perspective, may or may not be relevant to present practice (thus, highlighting the need for work on this topic in current schools).
Perrin (1954) found that CWS were not readily accepted as members of their classroom, and suggested that stuttering students who were more able to adjust to interpersonal situations responded better to therapy. Though some studies present a coherent picture of problems a CWS has in different social groups, other studies have failed to find any influence of social factors on the status of CWS. Brissey and Trotter (1955) examined the social relationships within a group of speech-impaired children enrolled in a six-week summer residential clinic. They found no indication that social status was correlated with the severity of speech impediment. They concluded that the composition of the study group may have led the members to be more tolerant of speech dysfluencies compared with when one stuttering child had been in a group of fluent peers. Woods (1974) reported evidence that showed the social relationships of a stuttering child were no better or worse than those of fluent classmates. These studies using early clinical investigative tools, besides being dated, have not resulted consistent findings.
Many other groups can do observational data: parents, peers, teachers being obvious examples. Observational data have the advantage of rater-reliability in the sense that different raters can be compared as this can serve as an index of reliability. However, raters see their ‘target’ in different contexts that may indeed affect the rating. Thus teachers see classroom behavior, parents dining table behavior and peers contact with strangers. Observers have qualitatively and quantitatively different data. Thus low reliability (poor alphas) may not be an index of poor ratings.
Retrospective reports of problems in social groups by PWS have also been obtained. Mooney and Smith (1995) used a questionnaire to obtain information regarding their time at school from adults who stutter. They found 11% of adults who stutter said they had been bullied at school and that this had a negative effect on the fluency of their speech. Comparison of this with estimates about how many fluent school children are bullied, on the other hand, indicates that children who stutter are no more at risk of bullying than their peers. Haynie, Nansel, Eitel, Crump, Saylor, Yu and Simons-Morton, (2001) reported that over 30% of schoolchildren stated that they had been bullied within the last school year (much higher than the 11% of adults that stutter who reported having been bullied by Mooney and Smith, 1994).
The most recent retrospective report was by Hugh-Jones and Smith (1999). In this study 74% of 276 adults who stutter that took part in the survey reported that they had been bullied during their time at school. Of the 205 respondents that indicated they were bullied at school, 6% reported that the bullying had a long-term effect on their fluency. However, the study lacked a control group who does not stutter to establish whether fluent speakers were bullied less often. The authors also note other limitations in the project, common to all retrospective studies: Respondent’s recollections may be distorted and there is no way of validating the responses. The authors also concede that the sample may be limited by the fact that the respondents were a volunteer sample from the British Stammering Association that may have resulted in a cohort that was particularly aware of the issues surrounding their dysfluency and its effects.
The usefulness of data from retrospective studies is limited by the contradictory findings. Also, retrospective self-reports are filled with methodological problems. Besides the problems already mentioned, studies that require retrospective reports (such as on parenting styles) are regularly rejected by journals as authors want to assert causality in that these styles affect adult behavior patterns. The reason for rejection is that these reports are compromised by the variable being investigated. Thus, for instance, it is possible that those who successfully ‘recover’ from stuttering offer a very different set of descriptives/ explanations than those who do not. These attributions and selective memory problems have led many to seriously question such data.
Sociometric methods have also been used to assess the dynamics of groups containing CWS. Marge (1966) reported a study that used these procedures to assess intellectual and social status, physical ability and speech skills of CWS. The study examined 197 third grade (8-9 years) public school students, of whom 36 had been diagnosed with moderate or severe speech dysfluency. Sociograms were obtained on each of the four components investigated based on Moreno (1960). The study required children to rate other children in the class by responding to statements such as “I would like to work with this child” or “I would like to play with this child”. Marge reported that dysfluent children held a lower social position than fluent ones. Sociograms were also obtained from teachers on the same four component skills for each child. The data from the study indicated that, with regard to intellectual skills in school and social activity outside school, the child who stuttered held a significantly lower position than that of his or her fluent peers. In the other areas of playground activity and speech skills, no significant differences between the groups were found. The results from the teachers corroborated the findings of the peers. More recently Davis, Howell and Cook (2002) used a forced choice sociometric scale to assess the peer relationships of 16 CWS and their 403 classmates. They reported that CWS were rejected significantly more often than their fluent peers, were categorized as less popular and were less likely to be named as leaders. The CWS were three times more likely to be identified as victims of bullying than their fluent peers and generally demonstrated low social acceptance among peers.
A general problem to note that cuts across assessment methods used in studies of bullying, is that the majority of research into the social status of children who stutter uses data from respondents who were in the educational system more than two decades ago (either because the publications are dated or adult respondents provided retrospective reports). It is possible that the attitude of children toward their peers with disabilities (including those of speech) has changed in the intervening period. A secons general point is that researchers have always wisely called for the triangulation of methodologies to help reduce biases associated with each. There is no reason to believe that research in stuttering should be any different. Finally, it is possible to use other indexes of peer relationships. Diaries can prove useful as they can indicate the social world of the person who stutters. Phones can be logged to access number and length of calls.
There is consensus among clinicians and researchers working with PWS that speech disorders can have detrimental effects on self-perception and, specifically, on self-esteem (Bajina; Luper; Shames; Starkweather & Van Riper). As a result, therapeutic interventions often include either implicit or explicit goals to improve an individual’s concept of self-worth (Bloodstein; Cooper; LaBlance; Luper &). Yet, there are minimal empirical data that indicate a need for the implementation of regular clinical attention to self esteem for PWS in general (Yovetich, Leschied & Flicht, 2000).
Self-perception and self-concept have often been addressed in therapy for PWS (Sheehan; Silverman & Van Riper). Sheehan and Martyn (1966) suggest that individuals who have developed a concept of self as a PWS are less likely to recover spontaneously than those who have not. Beach and Fransella (1968), however, believe that for therapy to be successful, PWS must accept their speech disorder as part of their self-concept.
Pukacova (1973) used a projective technique (incomplete sentences) to estimate the self-esteem of 74 CWS; 94% of this sample evidenced low self-esteem. More recent work (Yovetich et al 2000; Blood, Blood, Tellis & Gabel, 2003) found no evidence for low self-esteem for school-age and adolescent stutterers, Yovetich et al (2000) used Battle’s (1992) Culture Free Self-Esteem Inventory. They reported no differences between the mean scores for CWS and the normative data. Eighty per cent of the participants (school age CWS) scored above the standardised mean on the Total Self-Esteem score. Blood et al (2003) found that 85% of the adolescents who stutter that participated in the study scored within 1 SD from the mean on a standardised measure of self-esteem - the Rosenberg Self-Esteem Scale (Rosenberg, 1965) meaning they were not significantly different to the norms.
Bardrick and Sheehan (1956) found that individuals with lower self-esteem showed higher rates of stuttering. Bajina (1995) noted a similar trend toward lower self-esteem in 28 PWS. Shames and Rubin (1986) report that the most common attitudes expressed by PWS are anxiety, helplessness, victimization, and low self-esteem.
To summarize the work on self esteem, how speakers who stutter perceive themselves may be decided in terms of their personality, their demographic background and also the severity of their stuttering. Though this topic appears to be ‘easy’ to research, it is hard to do well. The reason for this is threefold. First, a good control group is needed and a large representative sample for control speakers and speakers who stutter. Next, it is important to measure and analyze all the factors that relate to self esteem. This is because stuttering may be a moderator or mediator variable rather than a variable that affects stuttering directly. Most important, it is desirable to do longitudinal research to determine whether, why and when self esteem is a cause, consequence or both of stuttering.
Anxiety is widely believed to a causal factor in stuttering and plays a central role in many theories about the origin of stuttering (Miller & Watson, 1992). Stuttering usually worsens when a PWS speaks to strangers or addresses large audiences or those felt to be his or her superiors. This often leads to PWS avoiding social and public speaking situations and experiencing anxiety when in those situations (Van Riper, 1992). Yet PWS may not differ from those who do not stutter in baseline levels of anxiety (Blood & Miller), although this finding has been questioned (Craig, 1992).
Preliminary searches indicate that there has been no empirical work conducted examining the anxiety levels of children who stutter when compared to fluent controls. Perhaps this is an oversight and, if so, hopefully this will be highlighted in commentaries. The lack of data from CWS does not allow for causal implications to be considered. Is stuttering a consequence of anxiety or are adults who stutter more anxious because they stutter?
Several methods have shown consistently that anxiety levels in adult PWS are higher than those of adults who do not stutter. Trait anxiety questionnaire (Craig, Hancock, Tran & Craig, 2003); Inventory of Interpersonal Situations (IIS) (Kraaimaat, Vanryckeghem & Van Dam-Baggen, 2002); and the Cognitive Anxiety Scale (DiLollo, Manning & Neimeyer, 2003). Stein, Baird and Walker, (1996) found that seven out of 16 adults seeking treatment for stuttering could be classified as social phobics when using the DSM-IV criteria. Blood, Blood, Bennett, Simpson and Susman, (1994) found no differences between the anxiety levels of adult PWS who stutter and those of fluent controls using the State Anxiety Inventory, Trait Anxiety Inventory or Personal Report of Communication Apprehension. However, they did find that during high stress situations levels of salivary cortisol was significantly greater in the adult PWS than in the control group, indicating higher anxiety levels for the adults who stutter.
Miller and Watson (1992) examined self-perceptions of state and trait anxiety and refuted the assertion that PWS are more anxious than those who do not. Anxiety was not related to stuttering severity and their results indicated that high levels of anxiety in adult PWS were restricted to communications situations.
Anxiety can be measured by self-report measures of trait or state anxiety, observational reports of others or physiological measures like heart beat, galvanic skin response etc. In most research, there are surprisingly low correlations between them. For some it is subjective anxiety that is crucial and they advocate self report. Other researchers prefer physiological measures irrespective of whether speakers who stutter claim to feel stressed or merely aroused.
There is some evidence that negative stereotypes and perceptions can spread and lead to role entrapment, or limited career choices, for women, minorities, people with physical disabilities, and individuals with hearing impairment with some interesting high celebrity exceptions. In the light of this conclusion, it is important to explore whether the negative stereotypes of stuttering will spread and lead to limited career choices for PWS. Gabel, Blood, Tellis and Althouse (2004) concluded that to date, no research has explored whether this applies. However, two types of research studies have explored related issues.
The first type of research in this area involves the experiences of PWS at work. In a study exploring these issues, Rice and Kroll (1997) surveyed 568 National Stuttering Project members regarding their perceptions of past work experiences. Results indicated that stuttering directly affected these individuals’ perceptions of work experiences and career choices. In particular, 70% of the participants reported that they believed they could have had a better job if they did not stutter, and 56% reported choosing a career that required less speaking. 35% of the participants reported that they believed stuttering had affected their chances of being promoted, reported feeling discriminated against in the hiring process, and perceived that their supervisors had misjudged their performance because they stuttered. In a similar study, Opp, Hayden, and Cottrell (1997) surveyed 166 PWS. These participants answered questions about their job choices, number of years they were employed, and whether or not they reported experiencing discrimination in their careers. Results of the study suggested that 35% of the participants reported being in careers that required a low level of communication, and 39% believed they had experienced discrimination in the hiring process because of their stuttering.
To summarize, these studies have found that a significant number of people who stutter believed that: (a) they could have had a better job if they did not stutter; (b) they chose careers that required less communication; and (c) they felt discriminated against in the hiring process. Neither of these studies offered any explanations about why PWS reported these experiences. These perceptions may have occurred because of the participant’s own insecurities and beliefs related to stuttering or the negative perceptions of others in their environment.
It should also be noted that the Rice and Kroll (1997) and Opp, et al (1997) studies reviewed above suffer from the ubiquitous methodological problems allied to self-report data that have been addressed elsewhere in this paper. Briefly, there is no way of validating the respondent’s reports and those responses may also be distorted by the passage of time. These attributions and selective memory problems have led many to seriously question this type of data. Although the nature of both studies make it difficult to include a control group, the data may be further compromised by the fact that the respondents were a volunteer sample. This may have produced a cohort that was particularly aware of the issues surrounding stuttering and employment.
Craig and Calver (1991) xplored issues related to employment of PWS who had attended a treatment program. The first part of the study surveyed 34 employers regarding their perceptions of two groups of PWS who were employed at their companies. One group of individuals who stuttered received therapy to speak more fluently, while the other group did not. The employers perceived the speech of the individuals who had completed treatment to be more acceptable and the perceptions toward the individuals who had not received therapy did not change. Results indicated that PWS were perceived more positively when they are able to improve their fluency. In addition, Craig and Calver (1991) polled the 62 individuals who had completed therapy program regarding vocation and career changes following treatment. Nineteen of the respondents reported a promotion following the completion of therapy and 18 responded that a positive job change (an upgrade from their former position) followed treatment. Results of this study suggested that people who stuttered were not only perceived in a more positive manner by employers following treatment, but also experienced a positive change in career. Scloss, Espin, Smith and Suffolk (1987) also found that employers produced significantly more favorable ratings of PWS at employment interviews after they had received specialist assertiveness training relating to their stutter.
There is a third perspective - how employers and others view employment prospects for PWS. Gabel et al (2004) asked university students to report their perceptions of appropriate career choices for PWS. Results suggested that university students reported a perception that stuttering affected career opportunities and that 20 careers were judged to be unsuitable for PWS. Conversely, 23 careers were judged to be more appropriate for PWS than for fluent controls. In another study (Silverman & Paynter, 1990) students rated four scenarios (“a lawyer”, a lawyer who stutters”, a factory worker”, and “a factory worker who stutters”). Both the lawyer and factory worker who stuttered were rated as being less competent than others in these occupations. The negative impact on the “appearance of competence” was greater for the former than for latter. Craig and Calver (1991) compared employer perceptions of their employees’ speech between a group who had received treatment for stuttering and a nontreatment control. The employers’ perceptions of the treatment group were significantly enhanced, whereas no significant change occurred in employers’ perceptions for the control group. Hurst & Cooper (1983) used a questionnaire that required employers to indicate their strength of agreement to seven attitudinal statements concerning stuttering. Employers rejected the suggestion that stuttering has a negative effect on job performance but agreed that stuttering decreases employment prospects and interferes with promotion opportunities.
As a cautionary footnote, people in the business of recruitment and selection are sensitive to issues around discrimination. Unless the job specifically concerns/ necessitates clear communication (e.g. air traffic control) it is possible a PWS who is rejected at interview solely because of their stuttering may take legal action. There appears to be no studies in this area: namely the extent to which speech dysfluency on its own or together with other specific factors, influences occupational selection decisions.
Social and emotional factors appear to be crucial to the onset and maintenance of stuttering. This is evidenced by the fact that most models explaining the onset and development of stuttering include a social and emotional component. This paper has reviewed the methods and findings of previous research that investigated the role of affective and social factors in stuttering from pre-school to adulthood. It has also attempted to make readers aware of the various methods and issues in social psychology that could be used to investigate these phenomena and to indicate where these methods could be useful in assessing the role of social and affective components in stuttering.
This research was supported by the Wellcome Trust.