Stephen Davis, Peter Howell, Anna Killick, and Hannah Finch
Department of Psychology, University College London, Gower Street, London WC1E 6BT
The development of a computer-based instrument designed to assess communication attitudes in children up to 7 years is described. The instrument will be included in a battery of tasks to assess and evaluate factors close to stuttering onset to determine whether stuttering will persist or recover. Preliminary analysis indicated that the instrument was reliable and valid. Results replicated previous studies (De Nil & Brutten, 1986, 1991; Boutsen & Brutten, 1989) indicating the communication attitudes of children who stutter were more negative than those of their peers.
Few attempts have been made to assess the attitudes of children who stutter (CWS) (Costello, 1984). Two reasons suggest that measuring attitude in pre-school children is important: i) Stuttering usually starts between three and five years (Dalton & Hardcastle, 1977) ii) Attitudes about many things change dramatically between ages 3 and 5 (Perry, Bussey & Fischer, 1980). This is especially likely to be the case when the child is affected by a problem with the behavior in question. There is also evidence that attitudes are more easily changed during or close to formation (Niven, 1994). The problem is, that designing procedures for measuring attitude at these ages 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 children who do not stutter. 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 children who do not stutter 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, indicating that their speech-related attitudes were more negative. Similar between group differences were found with a group of American children by Boutsen & 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. Subsequently Vanryckeghem & Brutten (2001, 2002) have reported the development of KiddyCAT, an instrument that is administered verbally and is designed to measure the communication attitude of children younger than six years.
The work described in this current paper concerns the development of a computerized instrument designed to assess the communication attitudes of young children. An instrument capable of determining the communication attitudes of children close to stuttering onset would be useful in a number of areas. There have been several proposals that the onset of stuttering is a result of the belief that speech is difficult (Bloodstein, 1987; Brutten & Dunham, 1989; Guitar, 1976; Peters & Guitar, 1991). However, because there is currently no objective method of ascertaining the communication attitudes of young children it has not been possible to establish whether the negative beliefs that people who stutter have about speech is a product of, rather than a cause of their dysfluency. 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).
Design and construction of the assessment instrument
The age of the children under investigation renders paper-and-pencil assessments impractical and parental-report methods of assessing children’s communication attitudes have been shown to be unreliable (Vanryckeghem, 1995). This study uses a computer-based instrument - the computerised communication attitude assessment instrument (COMCAS). Lap top computers, touch screens and computer-generated sound have been used successfully to assess the language development of pre- school children. Au-Yeung, Howell, Davis, Sackin, & Cunniffe (2001) demonstrated that children as young as 2 years were successfully able to complete a language assessment that contained up to 70 test items when the test was controlled by a lap top computer. In Au-Yeung et al. (2001), children were required to respond to a computer-generated audible statement by touching the relevant picture on the computer screen. This is a natural response for a child and aids in keeping his or her attention. Audio presentation is standardized and response collection and statistical analysis are automated (ruling out the possibility of error in obtaining results and allowing the person administering the test to concentrate on monitoring the child). The rationale and methods employed in the design and administration of the instrument to measure the communication attitudes of young children are detailed next.
Content of the COMCAS
Previous studies investigating communication attitude use several different response methods. The Erickson S-24 (Andrews & Cutler, 1974) used Thurstone scaling and required the respondent to answer “true” or “false” to a series of statements. Brutten et al (1985) added an additional “don’t know” response for their Communication Attitude Test. More recently, studies of communication attitudes have used Likert scales. For example Watson (1987) used a seven-point Likert scale in her communication attitude test. Pilot work with young children indicated that they were unable to respond accurately to four response options. The COMCAS used a dichotomous like/dislike response scale using computer- generated animated response panels showing happyl sad human faces. Discussions with researchers, clinicians, teachers and parents elicited approximately fifty statements relating to communication situations. Pilot work and further discussion reduced this to the twenty statements in the final assessment. These are listed at the end of this chapter. The statements were presented in a neutral manner, e. g. Talking to older children and the child was required to indicate if they liked or disliked talking in this situation.
A total of 181 children (88 male and 93 female) aged between 3.67 years and 6.75 years (mean age 5.27 years, sd .65) took part in this study. The control group contained monolingual English speakers with no reported hearing, speech or language difficulties and comprised 143 children (68 male, 75 female, mean age 5.20 years, sd.59). The two experimental groups were (i) 22 children (8 male, 14 female, mean age 5.32 years, sd.58) with English as an (language), and (ii) 16 EAL alternative CWS (12 male, 4 female, mean age 5.93 years, sd.82).
The laptop computer containing the communication attitude test was set up in a quiet room. In the case of the controls the assessment was carried out while they were at school. For CWS it was carried out at the clinic and either incorporated into their Visit with the speech and language therapist or in their home. The computer presented the twenty statements individually. The order in which the questions were presented was fixed. In addition to the statements being presented on-screen they were read aloud by the assessor. Presenting COMCAS on a computer allowed for the response panels to be presented in an animated format that would gain and hold the attention of the children. As each statement was presented two response panels were displayed on-screen. Both displayed video images on a continuous loop - one with a person nodding and smiling to represent ‘like’, and the other shaking her head and frowning to represent ‘dislike’. This made it easy to represent like and dislike in a distinct and clear way for the children to understand. The children respond to each statement by touching the appropriate image on the screen to show if they like or dislike talking in that given situation - pointing and touching is a natural response for young children. The assessment took approximately 5 minutes.
Test-retest. Fifteen randomly selected children from the control group were re-assessed one week after initial assessment. Analysis indicated a significant positive correlation between original test score and retest score r: .750, p<0.01.
Item Facility Index. This scale indicates that a statement or item is reliable if it has a facility index value of between approximately 0.10 and 0.25. Any values below 0.10 are deemed to be unreliable. The results of this test suggested that there were three statements (1, 2 & 17) that were not reliable.
Cronbachs alpha coefficient. Analysis indicated a reliability coefficient for the assessment instrument of .7696. The standard figure for showing reliability through Cronbach’s Alpha Scale is .8, as the figure here is very nearly .8, the instrument is considered reliable. Analysis of the reliability scale however suggests that if two of the same statements as identified by the Facility Index (l&2) were deleted from the scale the alpha coefficient would be higher (.7876), and the scale more reliable. Subsequent analysis and results are based on the most reliable version of the instrument.
Comparison of COMCAS Score across the different language groups
The mean COMCAS score for each experimental group is shown in Table 1. A lower score indicates a more positive attitude.
Table l. Mean Communication Attitude Score and Standard Deviation for each Language Group
A one way ANOVA indicated a significant effect of experimental group F(2,177)=7.22, Sig, p=.001. Post hoc tests_ indicated that the difference between the mean scores of the monolingual English children and the CWS was significant (p=.O01) as was the difference between the EAL children and the CWS (p=.031). This indicates that the communication attitudes (as measured by the COMCAS) of CWS in this age group are more negative than those of the children in either of the other two groups.
The relationship between age and communication attitude of CWS was explored. The correlation between age and total score on COMCAS was not significant (r(l6) = -.O95, ns), but there is a tendency for communication attitude of CWS to become more positive as they get older.
Exploratory factor analysis
Principal components analysis with orthogonal (varimax) rotation revealed 4 factors with eigenvalues greater than 1.0. These factors accounted for 43% of the variance. Subjective interpretation of the factors indicated that items loading onto Factor 1 were those concerned with classroom communication situations, those items loading onto Factor 2 were relating to communication situations within the home. Factor 3 contained items that concerned speaking to known older children and adults and Factor 4 contained items relating to speaking to children and adults with whom the child has had little contact.
Inspection of the factor scores showed that CWS scored lower (indicating a more negative attitude) than children in the other two groups on all four factors. A one way AN OVA indicated a significant effect of experimental group for each of the four factors. Post hoc tests indicated that the factor scores of the CWS differed significantly from those of monolingual English children only on factors 1 (p=.002) and 3 (p=.039).
The instrument is judged to be valid through face Validity. The children responding to the test understood the statements being used and responded appropriately. The use of touch screen technology allowed the children to interact with the computer, maintaining their interest throughout the assessment. The fact that significant differences were found in the communication attitudes of the CWS and their fluent peers also suggests discriminative validity for the instrument. This supports previous findings that some CWS have more negative attitudes than children who do not stutter e.g. De Nil et al. (1991).
Comparison of the communication attitudes of fluent children and CWS
The results of this study showed that the young CWS scored significantly higher on COMCAS than did their fluent peers, indicating a more negative attitude toward speaking situations. This demonstrates that CWS close to the onset of the disorder are already more wary of certain speaking situations. This could be taken as evidence to support the hypothesis that stuttering is caused by a belief that speaking is difficult. (Bloodstein, 1987; Brutten & Dunham, 1989). However all of the CWS in this study were receiving or had received therapy and so had been made aware of the nature and severity of their dysfluency. Such awareness has the potential to cause CWS to re- evaluate the problems that may be caused by certain speaking situations. The cause/effect question could be resolved by assessing the communication attitudes of CWS prior to the start of therapeutic intervention.
COMCAS failed to find any indication that the communication attitude of CWS becomes more negative with age (De Nil et al, 1991). De Nil and colleagues found changes in attitude between 7-10 years, although COMCAS was designed to assess the communication attitudes of children younger than 7 years (as in this study), because of the content and mode of presentation it would also be suitable for use with older children.
What is COMCAS actually measuring?
Factor analysis identified four factors accounting for 43% of the variance. Examination of the statements in each factor, although subjective, highlights the underlying variables being measured by each factor. This enables a more fine grain analysis of the attitudes being measured. For example, the statements with high loadings on Factorl all relate to speaking in classroom situations those whilst those items loading onto Factor 2 were related to communication situations within the home. Examination of the responses contained within these factors would enable clinicians establish specific areas of concern. The results of this study indicated young CWS have negative attitudes toward communication in the classroom and with people that are not well known to them.
- Talking to my mother or father
- Talking to my friend at playtime
- Talking to my teacher in class
- Talking aloud to lots of people (whole class, assembly etc)
- Talking to a child who is older than me
- Asking my teacher a question
- Talking to my friends at lunchtime (at lunch, not in the playground)
- Talking to other adults in class (not the teacher)
- Talking to children I don’t know very well at playtime
- Talking to adults out of school (friend’s mum or dad etc)
- Talking to a new boy or girl in school
- Talking to adults at playtime (playground supervisor etc)
- Talking to my friends in class
- Talking to adults at lunch (dinner lady, supervisor etc)
- Asking an adult (no: a parent or teacher) a question
- Asking a group of children if I can join their game
- Talking to my friencs in class
- Talking to a child who is younger than me
- Explaining to another child how to do something (how to play a game etc)
- Talking on the telephone
Andrews, G. & Cutler, J. (1974). Stuttering therapy: The relation between changes in symptom level and attitudes. Journal of Speech and Hearing Disorders,_3__9_, 312-319.
Au-Yeung, J ., Howell, P., Davis, S., Sackin, S., & Cunniffe, P. (2000). Introducing the Preschoolers Reception of Syntax Test (PROST). Proceedings of conference on Cognitive Development, Besancon France.
Bloodstein, O. (1987). A handbook on stuttering. The National Easter Seal Society :Chicago
Boutsen, F. & Brutten, G. (1989). Stutterers and nonstutterers: A normative investigation of of children ‘s speech associated attitudes. Unpublished ‘manuscript, Southern Illinois University.
Brown, S.F. & Hull, H.C. (1942) A study of some social attitudes of a group of 59 stutterers. Journal of Speech Disorders, 1, 153-159. '
Brutten, G.J. (1985). Communication Attitude Test. Unpublished manuscript, Southern Illinois University, Department of Communication Disorders and Sciences, Carbondale.
Brutten, G.J. & Dunham, S. (1989) The communication attitude test. A normative study of grade school children. Journal of Fluency Disorders, 14, 371-377.
Costello, J,M. (1984) Treatment of the young chronic stutterer in, R,F. Curles & W, H. Perkins (Eds.) Nature and Treatment of Stuttering: College-Hill Press: San Diego
Dalton, P. & Hardcastle, W.J. (1977) Disorders of fluency and their afiects on communication. London: Edward Arnold.
De Nil, L. & Brutten, G. (1986). Stutterers and nonstutterers: A preliminary investigation of children’s speech-associated attitudes. Tijdschrift voor Logopedie en Audiologie, 16, 85-92.
De Nil, L. & Brutten, G. (1991) Speech-associated attitudes of stuttering and nonstuttering children. Journal of Speech and Hearing Research, 34, 60-66.
Erikson, R. L. (1969) Assessing communication attitudes among stutterers. Journal of Speech and Hearing Research,12_, 711-724.
Guitar, B. (1976) Pretreatment factors associated with the outcome of stuttering therapy. Journal of Speech and Hearing Research, 19, 590-600.
Johnson, W., Brown, S.F., Curtis, J .F., Edney, C.W., & Keaster, J. (1956). Speech handicapped school children (revised edition). New York: Harper Row. -
Niven, N. (1994) Health Psychology. Churchill Livingstone: London
Onslow, M. (1994). The Lidcombe Programme for early stuttering intervention: the hazards of compromise. Australian Communication Quarterly, Supplementary Issue, 11-14.
Perry, D.G, Bussey, K. and Fischer, J. (1980). Effects of rewarding children for resisting temptation on attitude change in the forbidden toy paradigm. Australian Journal of Psychology, 32, 225- 234.
Peters, T, J. and Guitar, B. (1991) Stuttering: An Integrated Approach to Its Nature and Treatment. Williams & Wilkins: Maryland.
Shearer W.M. (1961). A theoretical consideration of the self-concept and body image in stuttering therapy. American Speech- Hearing Association, 3, 115-116.
Sheehan, J.G. (1970). Stuttering: Research and therapy. New York: Harper and Row. Travis, L.E. (1957). Handbook of speech pathology. East Norwalk, CT,: Appleton-Century-Crofts.
Van Riper, C. (1948). Stuttering. Chicago, IL., National Society for Crippled Children and Adults, Inc.
Vanryckeghem, M. (1995) The Communication Attitude Test: A concordancy investigation of stuttering and nonstuttering children and their parents. Journal of Fluency Disorders, 20, 191-
Vanryckeghem, M. & Brutten, G, J. (1992) The Communication Attitude Test: A test-retest reliability investigation. Journal of Fluency Disorders, 17, 109-118.
Vanryckeghem, M. & Brutten, G, J. (1996) The relationship between communication attitude and fluency failure of stuttering and nonstuttering children. Journal of Fluency Disorders, 21, 109- 118.
Vanryckeghem, M., Hernandez, L. & Brutten, G. (2001). The KiddyCAT: A measure of speech- associated attitude of preschoolers. ASHA, 6, 136
Vanryckeghern M. & Brutten, G. (2002). KiddyCAT: A measure of stuttering and nonstuttering preschoolers’ attitude. ASHA, 7, 104
Vanryckeghem, M., Hylebos, C., Brutten, G.J., & Perelman, M. (2001) The relationship ‘between communication attitude and emotion of CWS. Journal of Fluency Disorders, 26, 1-15.
Watson, J, B. (1987) Profiles of stutterers’ and nonstutterers’ affective, cognitive and behavioral communication attitudes. Journal of Fluency Disorders, 12, 389-405.
Yairi, E., & Ambrose, N. (1992). Onset of stuttering in pre-school children: selected factors. Journal of Speech and Hearing Research, 35, 782-788