The relation between intrapersonal and interpersonal staff behaviour towards clients with ID and challenging behaviour: a validation study of the Staff–Client Interactive Behaviour Inventory
Abstract
Background Interpersonal staff behaviour is one of the instigating factors associated with challenging behaviour in clients with intellectual disabilities (ID). There are several studies focusing on the influence of intrapersonal staff characteristics – such as beliefs, attributions and emotional reactions – on staff behaviour. Little is known, however, about interpersonal staff behaviour itself. This study describes the development and validation of the Staff–Client Interactive Behaviour Inventory (SCIBI), measuring both intrapersonal and interpersonal staff behaviour in response to challenging behaviour in clients with ID.
Method A total of 292 staff members, employed in residential and community services, completed the SCIBI for 34 clients with ID and challenging behaviour.
Results Confirmatory factor analysis of a seven-factor model – with assertive control, hostile, friendly and support-seeking interpersonal behaviour; proactive thinking; self-reflection; and critical expressed emotion as reliable factors – showed an exact fit to the data, indicating construct validity and reliability of the SCIBI. A series of multilevel regression analyses showed higher age of the client to be negatively associated with assertive control. Job experience, level of education, type and sex of staff predicted interpersonal behaviour. Also, intrapersonal staff behaviour, including critical expressed emotion, proactive thinking and self-reflection, predicted interpersonal behaviour.
Conclusions The SCIBI can be used to identify staff intrapersonal and interpersonal behaviour towards clients with ID and challenging behaviour. Results obtained with the SCIBI can provide new directions for individual client treatment plans and staff training programmes.
Introduction
People with intellectual disabilities (ID) are at higher risk for behavioural problems and mental health problems than people without ID (Deb et al. 2001). It is often suggested that deficits in the behaviour of disabled individuals are due to a failure of the social environment to support the appropriate behaviours. Research has even shown that staff behaviour can be counterproductive and sometimes encourages and fosters, for example, challenging behaviours (Hastings 1996). As in multi-dimensional models (Griffiths et al. 1997; Willems 2007), the instigating conditions for the occurrence of behaviour problems should thus be understood in terms of not only personal and internal (i.e. medical, psychological and psychiatric) client conditions, but also interpersonal and external conditions.
Staff members may be considered as the key agents in the behavioural interventions for people with ID and challenging behaviour (Felce et al. 2000). In keeping with the preceding, the American Association on Intellectual and Developmental Disabilities introduced a guideline that calls for the provision of supportive counselling (e.g. friendly relationship, affirming, giving advice) in addition to applied behavioural analyses, management of the environment and client/family education for the treatment of clients with ID (Rush & Frances 2000). Examination of the impact interpersonal staff behaviour could have on the behaviour of clients is therefore critical for both assessment and therapeutic purposes.
Although interpersonal staff behaviour is considered a vital component in the provision of care, our knowledge about measuring interpersonal skills of staff working with clients with ID remains limited. Reliable and well-validated instruments to measure interpersonal behaviour of parents are available though. In models of parental care, there is widespread consensus regarding two underlying dimensions of parenting behaviour, namely control and discipline (i.e. high vs. low) and support and warmth (i.e. high vs. low) (Maccoby & Martin 1983; Reitz et al. 2006). More general, in several interpersonal models of personality, two orthogonal dimensions of personality have consistently been corroborated: (1) dominance/control/power and (2) affiliation. Leary (1957) was one of the first to introduce a complete interpersonal diagnostic system for the assessment of personality or the so-called Interpersonal Circle (ICL), which entails a control dimension (i.e. dominance-submission) and an affiliation dimension (i.e. love-hate). Wiggins later constructed a more carefully documented and clearly validated version of the ICL, the so-called Interpersonal Adjective Scale-Revised (Wiggins et al. 1988). Benjamin (1996) described in her Structural Analysis of Social Behavior three orthogonal dimensions of personality: (1) an interpersonal focus involving control versus emancipation or dominance versus autonomy giving; (2) the affiliation dimension involving love versus hate or friendliness versus hostility; and (3) the enmeshment-differentiation dimension involving submission versus separation/independence. Overall, control, submission, friendliness and hostility appear to be the four basic factors to describe interpersonal behaviour.
In addition to these critical factors of interpersonal staff behaviour, Hastings (2005) points at the relation between severe client behaviour problems and staff emotional reactions. In response to challenging behaviour, staff members in the field of ID can experience such emotional reactions as sadness, despair, anger or disgust (e.g. Bromley & Emerson 1995). These emotional responses tend to increase or decrease, depending on the nature of the attribution and staff willingness to help a client (Noone et al. 2006). Within the literature regarding staff behaviour, the term expressed emotions (EE) has become common to refer to the emotional climate of a relationship. Staff can be categorized as high or low on EE based upon measures of criticism, hostility and emotional overinvolvement. Staff members working in healthcare settings have been shown to display relatively high levels of EE (Moore et al. 1992). Investigators in the field of ID found evidence for undesirable effects of high levels of EE in staff on the quality of relationship (e.g. Van Humbeeck et al. 2003), but there are no studies focusing on the association of EE as an intrapersonal characteristic with interpersonal staff behaviour (Hastings & Brown 2002).
Next, when confronted with such intense emotions, the ability to reflect upon one's feelings and attributions can foster a stable sense of self, and thereby help staff to avoid entry into power struggles with clients (Morasky 2006) and countertransference (i.e. projection of emotions onto clients) (Norcross 2002). Also, Jackson et al. (2007) found that emotional insight and being more reflective as intrapersonal characteristics were important for enhancing personal resilience when confronted with adversities as bullying and violence.
Finally, exploring staff coping strategies for these emotional reactions, Mitchell & Hastings (2001) found that staff often used adaptive coping strategies as planning and active coping when confronted with challenging behaviour. Especially proactive thinking (Kirby et al. 2002) has been significantly associated with positive job performance, and it therefore should be considered an important intrapersonal characteristic of staff who have to deal with challenging behaviour in clients with ID.
The first aim of this study was to develop and evaluate an instrument to measure staff–client interactive behaviour, focusing on both interpersonal behaviour based on Leary and Benjamin in terms of (a) control, (b) submission, (c) friendliness and (d) hostility and on intrapersonal behaviour, including EE, self-reflection and proactive coping. The inventory was completed by staff working with individuals with ID and challenging behaviour living in a residential or community facility. Second, we examined relations between intrapersonal and interpersonal staff behaviours, accounting for several background variables of clients and staff (i.e. sex, age, ID level, diagnosis, setting, level of education, type of job and years of current job experience).
Method
Participants and procedure
A total of 292 staff employed in 12 facilities (34 teams) for individuals with ID participated in the present study, which was carried out in the Netherlands in 2004–08. Of the 292 staff members 78% was female and 22% was male. The mean age was 36 years with a range of 21–57 (SD = 9 years). In addition to high school, 81% of the staff had a 3-year professional training in the domain of nursing, social work or occupational therapy, which is standard in the Netherlands for direct care staff; 19% had a college-degree in nursing, teaching or social science. Three-fourth of the participants (74%) was employed as direct care staff, and one-fourth (26%) as occupational therapy staff. The mean length of experience with care in the current facility was 9 years, with a range of 1–34 years (SD = 7 years).
Staff data on the present Staff–Client Interactive Behaviour Inventory (SCIBI, a translation of a Dutch instrument) were analyzed with respect to 34 clients ranging from mild to profound ID and challenging behaviour, of which 16 clients with mild ID and 18 with lower ID levels (12 with moderate ID, five with severe ID and one with profound ID). In 30 cases intelligence was measured by Wechsler Intelligent Scale for Children or Wechsler Adult Intelligence Scale and in four cases with Dutch equivalents of the Vineland Adaptive Behaviour Scales or the Bayley Scales of Infant Development, second edition. In all these cases, the first author was consulted as a member of the Multi-Disciplinary Centre for Dual Disabilities, a specialized interdisciplinary team in the south of the Netherlands. Staff as well as their associated psychologists and physicians consult this team when there are serious concerns about the diagnosis and treatment of clients with severe behaviour or psychiatric problems. After interdisciplinary assessment, 18 clients were diagnosed with an autism spectrum disorder, six clients with personality disorders and 10 with other disorders (i.e. schizophrenia, mood disorder, reactive attachment disorder, and adjustment disorder). Twelve of the clients were male and 22 of the clients were female. The clients had a mean age of 36 years with a range of 14–70 years (SD = 15 years). A total of 26 clients were living in residential care and eight clients in community care. The SCIBI was completed by different numbers of staff members, ranging from 3 to 20 for a particular client.
Instrument
Staff members were asked to complete the pilot-version of the SCIBI, which was a 72 items self-report questionnaire using a five-point Likert Scale, ranging from completely inapplicable (1) to completely applicable (5). This pilot-version of the SCIBI was based on relevant literature and opinions of experts (i.e. 18 staff members and team managers). Staff behaviour addressed by the SCIBI includes randomly distributed questions on (a) control (n = 17) (b) hostility (n = 15) (c) friendliness (n = 14) and (d) submission (n = 9), as well as the following intrapersonal staff behaviours: (e) proactive thinking (n = 4), (f) self-reflection (n = 5) and (g) EE (n = 8).
Statistical analysis
The results are reported in three sections. In the first section, construct validity and internal consistency reliability of the SCIBI were examined by means of confirmatory factor analysis, using Mplus (Muthén & Muthén 1998), and the computation of Cronbach's alpha, respectively. A multifactor model was specified in which each indicator (item) loaded on only one factor, allowing items to correlate in case of similar wordings (e.g. ‘I impose strict demands upon this client’ and ‘I impose my will irrespective of what he may think’). Both fit-indices (CFI, TLI and RMSEA) and the model chi-square, also designated as the generalized likelihood ratio, were used to evaluate model fit (Kline 2005). The following fit index cut-off values are indicative of good model fit: CFI >0.95, TLI >0.95, and RMSEA <0.06, whereas a non-significant chi-square indicates exact model fit (Hu & Bentler 1999; Kline 2005). A modification index, giving the expected drop in chi-square if a parameter in question is freely estimated, was used to improve model fit. We thus identified parameters that could improve model fit by freeing those parameters. Examples of such parameters were items loading on more than one factor or the wrong factor. In stead of freeing those parameters, we removed them. Further improvement of model fit was achieved by removing items that did not load significantly on their respective factors.
The second section includes a preliminary analysis, where we examined associations among continuous client and staff background variables and the SCIBI scales by computing simple Pearson correlations coefficients.
In the third section, multi-level analyses were conducted, using MLwiN (Rasbash et al. 2000), in order to examine relations between intrapersonal staff behaviour and interpersonal staff behaviour, controlling for client and staff background variables. Traditional analyses, such as ordinary regression analysis, would only account for the individual staff member as the unit of analysis, thereby ignoring the fact that individual staff members (level 1) are nested within clients (level 2). It should be noted that ignoring the multi-level structure of the data would produce standard errors that are too small, which may generate spurious results (Hox 2002). Multilevel analysis, however, allows the simultaneous examination of how individual and group level variables are related to individual level outcomes, accounting for the non-independence of observations within groups (Goldstein 1995).
A stepwise procedure was followed in analyzing the data. In the first step, a null-model, which is a random intercept-only model containing an outcome variable and no explanatory variables, was fitted to the data as a baseline. In the next step, the explanatory variables were entered in order to test whether the explanatory model would make a significant improvement compared with the null model. Improvement of model fit was tested by the difference in deviance, which has a chi-squared distribution and can be used to test whether the more elaborate explanatory model fits significantly better than the null model. Finally, we examined whether random slope models, allowing the regression coefficients for staff and client explanatory variables to vary randomly across staff and clients, provided a better fit to the data, and tested for same and cross-level interactions between explanatory variables.
Results
Construct validity and internal consistency reliability
In order to establish construct validity, a confirmatory factor analysis was performed on all items of the SCIBI. After removing 42 items that did not fulfil the specified criteria, a seven-factor solution showed an exact fit to the data: χ2 (375) = 412.48, P = 0.09 (ns.). The fit-indexes were excellent: RMSEA = 0.02, CFI = 0.99 and TLI = 0.98 (Hu & Bentler 1999). All items loaded highly (over 0.50) and exclusively on their corresponding factors (see Table 1). In order to appropriately describe and interpret the seven-factor solution, we slightly renamed the seven factors found in the literature into assertive control (seven items), hostility (four items), friendliness (five items), support-seeking (three items), proactive thinking (three items), self-reflection (three items) and critical EE (five items). Table 1 presents the factor solution, with Cronbach's alpha for all scales, and an overview of all 30 remaining SCIBI items with their corresponding factor loadings and their random item numbers. The alpha values were satisfactory, ranging from 0.68 (support-seeking) to 0.89 (proactive thinking).
| Factor 1 | Assertive control interpersonal behaviour | Cronbach's α = 0.84 |
| Item no. | Factor loadings | |
| 1 | I handle my rules in a strict manner | 0.65 |
| 9 | I go my own way despite critique from this client | 0.53 |
| 11 | I impose strict demands upon this client | 0.61 |
| 13 | I impose my will irrespective of what he may think | 0.59 |
| 20 | I act correctively towards this client | 0.80 |
| 22 | I act prohibitively towards him | 0.76 |
| 25 | I take the lead when I am with this client | 0.54 |
| Factor 2 | Hostile interpersonal behaviour | Cronbach's α = 0.72 |
| Item no. | Factor loadings | |
| 8 | I protest with this client when I do not agree with him | 0.51 |
| 14 | I state my opinion directly to him | 0.54 |
| 23 | I let him see my anger | 0.74 |
| 26 | I grumble at this client | 0.78 |
| Factor 3 | Friendly interpersonal behaviour | Cronbach's α = 0.82 |
| Item no. | Factor loadings | |
| 2 | I value this client | 0.57 |
| 4 | I like to communicate with him | 0.70 |
| 7 | I like doing something with this client | 0.79 |
| 17 | I can work well with this client | 0.66 |
| 28 | I often feel nice with this client | 0.82 |
| Factor 4 | Support-seeking interpersonal behaviour | Cronbach's α = 0.68 |
| Item no. | Factor loadings | |
| 10 | I can handle everything better when this client supports me | 0.77 |
| 15 | I need encouragement from him | 0.57 |
| 19 | I like to be backed up by him | 0.63 |
| Factor 5 | Proactive thinking (intrapersonal behaviour) | Cronbach's α = 0.89 |
| Item no. | Factor loadings | |
| 21 | In working with this client, I think about WHAT I am going to do. | 0.77 |
| 27 | In working with this client, I think about HOW I am going to do things | 0.91 |
| 30 | In working with this client, I think about WHY I am going to do things in such a manner | 0.87 |
| Factor 6 | Self-reflection (intrapersonal behaviour) | Cronbach's α = 0.70 |
| Item no. | Factor loadings | |
| 3 | In working with this client, I think about what I myself want to attain | 0.60 |
| 24 | In working with this client, I think about what I would like to receive in return from him | 0.52 |
| 29 | In working with this client, I think about how I feel | 0.79 |
| Factor 7 | Critical expressed emotion (intrapersonal behaviour) | Cronbach's α = 0.75 |
| Item no. | Factor loadings | |
| 5 | In working with this client, I have the tendency to deliver a long ‘sermon’ | 0.61 |
| 6 | In working with this client, I have the tendency to work hard in order not to have to think about anything | 0.69 |
| 12 | In working with this client, I have the tendency to sometimes reject a reasonable proposal | 0.51 |
| 16 | In working with this client, I have the tendency to act directly without knowing where I really want to go | 0.71 |
| 18 | In working with this client, I have the tendency to approach him cynically | 0.58 |
- SCIBI, Staff–Client Interactive Behaviour Inventory.
Preliminary analysis
Associations among continuous background variables of clients (age, ID level), staff (age, job experience in current facility) and the seven SCIBI dimensions are presented in Table 2. Only effects at P < 0.001 were considered significant in order to avoid chance capitalization because of multiple testing. Only one significant association was found between continuous background variables and interpersonal staff behaviour and none with intrapersonal staff behaviour. Staff reported a higher level of assertive controlling behaviour towards clients with lower ID levels (r = 0.25). Only one of the correlations between the four interpersonal behaviours – namely, between hostility and assertive control – proved to be significant (r = 0.57).
| Variables | M | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Client and staff characteristics | |||||||||||||
| 1. Client age | 39.59 | 14.73 | 1.00 | ||||||||||
| 2. Level of intellectual disability | 1.76 | 0.83 | 0.10 | 1.00 | |||||||||
| 3. Age of staff | 36.33 | 8.79 | 0.27* | −0.18 | 1.00 | ||||||||
| 4. Years of job experience in current facility | 9.08 | 6.74 | 0.23* | −0.03 | 0.57* | 1.00 | |||||||
| Interpersonal behaviour (SCIBI) | |||||||||||||
| 5. Assertive Control | 3.16 | 0.80 | −0.15 | 0.25* | −0.16 | 0.05 | 1.00 | ||||||
| 6. Hostile | 2.83 | 0.88 | −0.01 | 0.04 | −0.13 | 0.07 | 0.57* | 1.00 | |||||
| 7. Friendly | 3.79 | 0.70 | 0.10 | −0.04 | 0.07 | 0.20 | 0.01 | 0.14 | 1.00 | ||||
| 8. Support-seeking | 1.46 | 0.64 | 0.06 | 0.05 | −0.10 | −0.02 | 0.06 | 0.17 | 0.03 | 1.00 | |||
| Intrapersonal behaviour (SCIBI) | |||||||||||||
| 9. Proactive thinking | 3.86 | 0.96 | 0.16 | 0.06 | −0.13 | −0.15 | 0.21* | 0.06 | 0.06 | 0.11 | 1.00 | ||
| 10. Self-reflection | 2.75 | 1.01 | 0.00 | 0.11 | −0.14 | −0.10 | 0.22* | 0.13 | −0.04 | 0.29* | 0.45* | 1.00 | |
| 11. Critical expressed emotion | 1.51 | 0.55 | 0.05 | −0.07 | −0.08 | 0.04 | 0.11 | 0.29* | −0.22* | 0.27* | −0.13 | 0.12 | 1.00 |
- n = 292 staff members; n = 34 clients; * P < 0.001 (two-tailed).
- SCIBI, Staff–Client Interactive Behaviour Inventory.
Half of the correlations among intrapersonal and interpersonal staff behaviour proved to be significant. More self-reflection was strongly related to higher levels of proactive thinking (r = .45), while proactive thinking proved to be positively and moderately associated with assertive control (r = 0.21). Increased self-reflection was moderately correlated with both more assertive control (r = 0.22) and support-seeking interpersonal behaviour (r = 0.29) towards clients. Finally, we found more critical EE to be moderately associated with more hostile (r = 0.29) and less friendly behaviour (r = −0.22) towards clients, and higher levels of support-seeking (r = 0.27).
Multilevel regression analyses
In order to account for the nested structure of the data, associations between intrapersonal staff behaviour (proactive thinking, self-reflection and critical EE) and interpersonal staff behaviour (assertive control, hostility, friendliness and support-seeking) were tested in four consecutive multilevel regression analyses. The variance components, the standardized regression coefficients (beta's) and chi-squared statistics for improvement of model fit of the explanatory models are presented in Table 3. As shown, all explanatory models resulted in a significant improvement of model fit compared with the null model. Random slope models, allowing the regression coefficients for staff and client explanatory variables to vary randomly across staff and clients, did not provide a better fit to the data, and no significant same or cross-level interactions were found, which means that we did not find any evidence of moderation.
| Predictors | Staff interpersonal behaviour | |||
|---|---|---|---|---|
| Assertive control | Hostile | Friendly | Support-seeking | |
| Staff level (beta's) | ||||
| Female | 0.08‡ | 0.10 | 0.04 | −0.14* |
| Age | −0.02 | −0.11 | −0.10 | −0.13 |
| Years of job experience in current facility | 0.09 | 0.07 | 0.26*** | 0.07 |
| Level of education (1 = 3-year study, 2 = 4-year study) | −0.07 | −0.14** | 0.00 | −0.08 |
| Care staff | 0.14* | 0.10 | 0.04 | −0.06 |
| Proactive thinking | 0.18** | 0.11 | 0.07 | 0.04 |
| Self-reflection | 0.08 | 0.07 | −0.04 | 0.23** |
| Critical expressed emotion | 0.10 | 0.23*** | −0.21*** | 0.23*** |
| Client level (beta's) | ||||
| Sex (1 = male; 2 = female) | 0.02 | 0.06 | 0.08 | −0.03 |
| Age | −0.26* | −0.05 | 0.08 | 0.06 |
| Level of intellectual disability (1 = mild ID to 4 = profound ID) | 0.21† | −0.03 | −0.15 | 0.05 |
| ASD (1 = no; 2 = yes) | 0.10 | 0.09 | 0.15 | −0.06 |
| Setting (1 = residential; 2 community) | 0.07 | −0.00 | 0.03 | 0.04 |
| Variance components intercept only or null model | ||||
| Staff level | 0.384 | 0.573 | 0.419 | 0.372 |
| Client level | 0.278 | 0.210 | 0.080 | 0.037 |
| Variance components explained by predictors | ||||
| Staff level | 0.354 | 0.518 | 0.383 | 0.326 |
| Client level | 0.136 | 0.115 | 0.046 | 0.010 |
| Improvement in model fit because of predictors: | ||||
| χ2 (d.f. = 13) | 41.06*** | 41.12*** | 34.84*** | 51.16*** |
- n = 292 Staff, n = 34 Clients; † P < 0.10, * P < 0.05, ** P < 0.01, *** P < 0.001.
- ‡ Standardized regression coefficients.
- ASD, Autism Spectrum Disorder.
Assertive control
It can be derived from Table 3 that 58% [100 * (0.384)/(0.384 + 0.278)] of the variance in assertive control could be attributed to differences between staff members, and 42% [100 * (0.278)/(0.384 + 0.278)] to differences between clients. The explanatory model gave a significantly better fit than the null model –χ2 (13, n = 292) = 41.06, P < 0.001 – accounting for 26% of the variance in assertive control. The proportion of explained variance was 5% [100 * (0.384 − 0.354)/(0.384 + 0.278)] at the staff level and 21% [100 * (0.278 − 0.136)/(0.384 + 0.278)] at the client level. Staff who rated themselves higher in proactive thinking (b = 0.18) and care staff in stead of occupational therapy staff (b = 0.14) showed increased levels of assertive control, while higher age of the clients proved to be associated with less assertive control (b = −0.26). Finally, a positive association between more severe ID and higher levels of assertive control just failed to reach significance (b = 0.21). This relation may be regarded as a trend.
Hostile behaviour
A total of 73% [100 * (0.573)/(0.573 + 0.210)] of the variance in hostile behaviour could be attributed to differences between staff members, and 27% [100 * (0.210)/(0.573 + 0.210)] to differences between clients. The explanatory model resulted in a significant improvement of model fit –χ2 (13, n = 292) = 41.12, P < 0.001 – accounting for 19% of the variance in hostile behaviour, which was distributed as follows: 7% [100 * (0.573 − 0.518)/(0.573 + 0.210)] at the staff level, and 12% [100 * (0.210 − 0.115)/(0.573 + 0.210)] at the client level. Although a substantial part of the explained variance in hostile behaviour was distributed at the client level, none of the single client level explanatory variables proved to be significantly related to hostile behaviour. At the staff level, however, more critical EE (b = 0.23) was significantly related to more hostile behaviour, whereas higher-level vocational education (b = −0.14) was associated with less hostile behaviour.
Friendly behaviour
Table 3 shows that 84% [100 * (0.419)/(0.419 + 0.080)] of the variance in friendly behaviour could be attributed to differences between staff members, and 16% [100 * (0.080)/(0.419 + 0.080)] to differences between clients. The model fit of the explanatory model was significantly better than that of the null model: χ2 (13, n = 292) = 34.84, P < 0.001. The explanatory model accounted for 14% of the variance in friendly behaviour, which was distributed as follows: 7% [100 * (0.419 − 0.383)/(0.419 + 0.080)] at the staff level and 7% [100 * (0.080 − 0.046)/(0.419 + 0.080)] at the client level. Although the explained variance was equally distributed across the staff and client level, no single client characteristic proved to be significant. Staff with more job experience reported more friendly behaviour (b = 0.26), and staff with higher critical EE reported less friendly behaviour (b = −0.21).
Support-seeking behaviour
It can be derived from Table 3 that 91% [100 * (0.372)/(0.372 + 0.037)] of the variance in support-seeking behaviour could be attributed to differences between staff members, and 9% [100 * (0.073)/(0.372 + 0.037)] to differences between clients. Again, the explanatory model generated a highly significant improvement over the null model –χ2 (13, n = 292) = 51.16, P < 0.001 – accounting for 18% of the variance in support-seeking behaviour. The proportion of explained variance was 11% [100 * (0.372 − 0.326)/(0.372 + 0.037)] at the staff level and 7% [100 * (0.037 − 0.010)/(0.372 + 0.037)] at the client level. No single client explanatory variable contributed significantly to the prediction of support-seeking behaviour. At the staff level, however, critical EE (b = 0.23) and self-reflection (b = 0.23) were both positively associated with support-seeking behaviour. Finally, female staff showed less support-seeking behaviour than male staff did (b = −0.14).
Discussion
In the present study, the validity of the SCIBI was examined in a sample of clients with severe behaviour or psychiatric problems. Support for construct validity was found in a confirmatory factor analysis of seven reliable factors, including assertive control, hostile behaviour, friendly behaviour, support-seeking behaviour, proactive thinking, self-reflection and critical EE. Staff hostile behaviour proved to be strongly associated with assertive control. As Shechtman & Horowitz (2006) have demonstrated, people who are frustrated in their attempt to exert sufficient control over other people's behaviour can experience a disproportionate amount of hostility. This means that, when control motives are frustrated, as can be expected when staff are confronted with challenging behaviour, the negative pole of the affiliation dimension (i.e. hostility) can manifest itself. A second strong correlation was found between two intrapersonal staff characteristics, that is, self-reflection and proactive thinking. Self-reflection and proactive thinking can be intertwined, as both are aspects of emotional intelligence (Gerits et al. 2004) and concern adequate ways of coping with challenging behaviour, one more self- and emotion-focused and the other more client- and task-focused (Zeidner & Endler 1996).
Although most of the differences in interpersonal staff behaviour could be ascribed to staff characteristics, a considerable amount of variance was still attributable to client factors, ranging from 9% to 42%. First, our results show that staff members tend to behave in a more assertive controlling way towards clients who are younger. This may be explained by the fact that adolescents and young adults with challenging behaviour need relatively more control and support because of their lack of decision-making capacities. In line with this explanation, there was also a trend suggesting that staff members use more assertive controlling behaviour in response to clients with lower ID levels.
Differences in interpersonal staff behaviours were attributed to staff characteristics for a large 58–91%. First, more job experience was associated with increased friendly behaviour towards clients with challenging behaviour. This is in line with Knotter et al. (2008), who found that staff members with more job experience were more supportive, comforting and positive reinforcing towards clients. Second, the higher the staff educational level, the lower their hostile behaviour, which can be considered as an argument for implementing further training and coaching programmes directed at staff dealing with challenging behaviour in clients with ID. Third, care staff as opposed to occupational therapy staff used a more assertive controlling style towards clients in employment settings. A plausible explanation might be that clients tend to profit from the structuring features of the work itself in these settings. Finally, female staff showed significantly less support-seeking behaviour towards clients with challenging behaviour compared with male staff. The items that represent support-seeking in the SCIBI are focused on the support and encouragement the worker needs, thereby reflecting a strong focus on him- or herself. This finding is consistent with results from a study by Gerits et al. (2004), who found that male staff scored significantly higher on intrapersonal emotional intelligence (i.e. assertiveness and self-regard), which is in line with support-seeking, whereas female staff scored higher on interpersonal emotional intelligence (i.e. empathy, interpersonal relationship, social responsibility), which is comparable with support-giving.
An important finding of our study was that intrapersonal staff behaviour proved to be associated with interpersonal staff behaviour. In particular, critical EE was strongly associated with hostile and low friendly behaviour. Critical EE concerns criticism, being unreasonable or cynical, and it is therefore conceptually connected with hostility and low friendliness. Next, critical EE correlates with support-seeking behaviour. It is understandable that staff having a tendency to show negative emotional reactions send interpersonal signals for back-up and support in contact with their clients. This is in line with Hastings & Brown (2002), who found that higher self-efficacy – which can be seen as the opposite of support-seeking – predicted less negative emotional reactions.
Increased self-reflection proved to be associated with more support-seeking behaviour. It is likely that staff members who show high levels of self-reflection are skilful in using an important principle of change, namely, complementarity. Increases in complementary support-seeking staff behaviour may evoke more assertive behaviour in clients, thereby changing challenging behaviour, such as aggressive and oppositional behaviour, into adaptive behaviour. The positive association between support-seeking behaviour and self-reflection is in line with results from a study by Horowitz et al. (2006), showing that interpersonal behaviour is motivated by a relatively strong goal orientation, while these underlying goals or motives are not necessarily open to inspection. In case of staff acting in a support-seeking manner in response to challenging behaviour, while the opposite – exerting control – might be expected, it is likely that they will start reflecting upon their feelings, motives and expectations for their unconventional behaviour. Such conscious reflection can also lead to proactive thinking with regard to one's interpersonal behaviour in the future, which might explain the strong association between self-reflection and proactive thinking.
Finally, proactive thinking of staff was associated with assertive control as well as with educational level. It is plausible that thinking ahead in case of dealing with challenging behaviour helps staff to think of more guidance and directing interventions in their attempts to take control over challenging behaviour. Also, this finding can be considered as a powerful therapeutic tool which can be further enhanced by more training, because proactive thinking is likely to be more helpful than just restrict oneself to reacting upon challenging behaviour.
Although it is encouraging to see that comparable instruments for the assessment of caregiver–child interaction (de Schipper 2007) and the assessment of the psychotherapeutic relationship (e.g. WAI, BLRI) encompass mostly the same factors and behavioural dimensions, the SCIBI must be further validated by paying attention to discriminant, convergent and predictive validity. Because the SCIBI is a self-report measure of staff behaviour, it is imperative that scores on the SCIBI be compared with observations of staff behaviour. In future studies of interpersonal staff behaviour, it is also essential to incorporate more personal characteristics of clients, like type of challenging behaviour and client interpersonal behaviour, to examine effects on the various interpersonal staff behaviours. This inventory is a translation of a Dutch scale and therefore it is important that for use in an English setting, the psychometrics must be further examined to ensure that the translation has not affected the underlying factor structure.
In proposing a practical and research framework, Hastings (2005) argues that challenging behaviour of clients can be more fully understood when variables affecting staff behaviour are identified, focusing on their behaviour, their emotional reactions and their beliefs and attitudes towards challenging behaviour. As suggested in the introduction, the SCIBI can be used to assess some of these interpersonal and intrapersonal staff behaviours, examining the influence of staff beliefs, psychological resources, such as self-efficacy and coping style, and social support in teams.
Besides identifying aspects of interpersonal behaviour of staff members that are related to the occurrence of challenging behaviour, one can also use the SCIBI to identify the specific interpersonal behaviours that work best with individual clients. Moreover, with a focus on the differences between the ideal profile and an ideographic staff profile, coaching goals can be set. In consulting practice and in training programmes, a profile of the four interpersonal staff behaviours makes it possible to implement some powerful principles of change from interpersonal and systems-oriented therapy, e.g. symmetry and complementarity. In staff training, it is advisable to use video and verbal feedback, which has been proven to be an effective staff intervention (Embregts 2002; van Oorsouw et al. 2009). Furthermore, our study shows the need for reducing critical EE, stimulating more self-reflection and enhancing proactive thinking related to interpersonal staff behaviour. Cognitive–behavioural training and emotional intelligence training (Embregts & Gerits 2007) are some possible intervention strategies for this. For purposes of team discussion and feedback, the SCIBI might also be completed by colleagues of a staff member (i.e. other-report) for comparison with self-report, thus leading to more open communication in a team on one's behaviour towards a particular client.
Given that the SCIBI only takes about 5–10 min to administer, it certainly merits a place in the multidimensional assessment of challenging client behaviour. Other instruments can be used to assess such staff characteristics as coping style, stress, emotional intelligence and personality factors. The SCIBI can be used to assess the interpersonal characteristics of staff working in various contexts.
Acknowledgements
We would like to thank clients and staff members of the facilities (i.e. Maasveld, St. Anna, Op de Bies, Pepijn en Paulus, PSW, Daelzicht, Dichterbij, ASVZ and Radar) who participated in this study. We are also grateful to W. van der Kamp for her statistical assistance and to the team managers for their assistance with the data collection.
