Verbal-and-nonverbal-emotional-behaviour-of-staff-A-first-attempt-in-the-development-of-an-observation-instrument_2011_Research-in-Developmental-Disabilities.pdf

Research in Developmental Disabilities 32 (2011) 2408–2414

Contents lists available at ScienceDirect

Research in Developmental Disabilities

Verbal and nonverbal emotional behaviour of staff: A ﬁrst attempt in the

development of an observation instrument

Wietske M.W.J. van Oorsouw a , b , * , Petri J.C.M. Embregts a , c , d , Jody Sohier e

a Tilburg University, Department of Developmental and Clinical Psychology, Tilburg, The Netherlands

b Amarant, Tilburg, The Netherlands

c HAN University of Applied Sciences, Faculty of Health and Social Studies, Nijmegen, The Netherlands

d SZ Dichterbij, Oostrum, The Netherlands

e Radboud University Nijmegen, The Netherlands

A R T I C L E I N F O

A B S T R A C T

Article history:

Received 11 July 2011

Accepted 15 July 2011

Available online 15 August 2011

It is common to use questionnaires and interviews to assess the emotions of staff who serve

clients with intellectual disabilities. Remarkably, observations of actual staff behaviour and

assessments of nonverbal expressions are usually not involved. In the present study, we have

made a ﬁrst start in the development of an observation instrument that registers both verbal

and nonverbal expressions of emotion. Following extensive literature reviews, interviews

with clinical experts and pilot observations, we have created an observation system and

observed a team of staff members during a period of three months. Results revealed

sufﬁcient to good interrater reliabilities regarding the observation of the three forms of

emotional expression: body movement, facial expression, and verbal utterances. Low levels

of explained variances between the different forms of expression showed that predictive

values were weak. This means that there is a surplue value of the observation of nonverbal

emotional expressions as compared to existing instruments that base their assessment on

verbal and/or written responses only.

Keywords:

Staff

Emotion

Behaviour

Intellectual disabilities

Observation

1. Introduction

Staff members serving people with intellectual disabilities (ID) frequently report negative emotions that among others

appear to be related to job demands, team climate within the organization, and challenging behaviour of clients (( Moore,

Ball, & Kuipers, 1992; Rose, 1993; Rose & Schelewa-Davies, 1997; Rose, David, & Jones, 2003 ). In the last decade, results

regarding negative effects of staff emotions on staff behaviour towards their clients have been reported frequently. High

levels of so called ‘‘Expressed Emotions’’ from staff negatively affect both challenging behaviour and well-being of client

( Tattan & Tarrier, 2000; Van Humbeeck & Van Audenhove, 2003 ). Staff members who experience negative emotions

regarding challenging behaviour of clients are less willing to help their client, use less adequate coping strategies and as a

consequence, show staff behaviour that maintains or even increases the persistence of challenging behaviour of clients in the

long term ( Hastings & Remington, 1994; Lambrechts, Petry, & Maes, 2008 ). Apart from the conclusion whether expressed

emotions of staff are either positive or negative, high levels of emotional expression towards clients with ID are undesirable.

We need to be aware of the fact that people with ID show deﬁciencies in their executive verbal functions and social

* Corresponding author at: Room P104, Tilburg University, Faculty of Social and Behavioural Sciences, P.O. Box 90153, 5000 LE Tilburg, The Netherlands,

Tel.: +31 13 466 8922; fax: +31 13 466 2067.

E-mail addresses: W.M.W.J.vanOorsouw@UvT.nl (Wietske M.W.J. van Oorsouw), P.J.C.M.Embregts@UvT.nl (Petri J.C.M. Embregts).

doi: 10.1016/j.ridd.2011.07.019

W.M.W.J. van Oorsouw et al. / Research in Developmental Disabilities 32 (2011) 2408–2414

2409

information processing, which leads to problems in the perception, understanding and control of emotions (e.g., Schuengel &

Janssen, 2006 ).

Several measures are used in research regarding emotions of staff serving clients with ID. Within studies regarding

emotions, attributions and beliefs of staff, it is common to use questionnaires like the Emotional Reactions to Aggressive

Challenging Behaviour Scale ( Mitchell & Hastings, 1998 ) and the Thoughts and Feelings Questionnaire ( Fletcher, 1989 ).

Within studies regarding expressed emotions of staff, investigators frequently use interviews (e.g., Camberwell Family

Interview [CFI], The Five Minute Speech Sample [FMSS]) as well as questionnaires (e.g., Level of Expressed Emotion [LEE], The

Perceived Criticism Scale [PCS]; Van Humbeeck et al., 2001; Weigel, Langdon, Collins, & O’Brien, 2006 ). Beyond all doubt,

these studies lead to a broad range of useful and interesting knowledge about emotional staff behaviour. It is however

remarkable, that most measures are based on written responses to questionnaires and verbal utterances during interviews,

instead of observations of actual staff behaviour (for a review, see Lambrechts et al., 2008 ). Measures that are solely based on

verbal or written expressions have some limitations. Emotions can be expressed in several ways and the dominant inﬂuence

of nonverbal expressions might be underestimated ( Navarro & Karlins, 2008 ). Besides, it is not yet clear, if and how different

forms of expression (verbal, face, body) cohere to each other. At present, a simple observation instrument that enables the

observer to register both verbal and nonverbal expressions of emotional behaviour is not available.

In the present study we made a ﬁrst attempt in the development of an observation instrument that measures both verbal

and nonverbal forms of emotional expression. Based on an extensive literature search, reviewing clinical experts and

conducting pilot observations, we developed and observation instrument. During a period of three months, we conducted

this instrument during shift change within one team of staff members serving clients with mild ID and challenging

behaviour. This exploratory study was aimed to develop an observation instrument from which the content was based on

scientiﬁc research regarding emotional behaviour. Also, we aimed to develop an instrument with sufﬁcient to good

interrater reliability scores. We did not take into account the nature (positive of negative) nor the type (fear, anger,

happiness) of emotions. When sufﬁcient interrater reliabilities would be achieved, we would additionally analyse whether

the observation of nonverbal emotional behaviours next to verbal emotional behaviours would have a surplue value over the

observation of verbal forms of emotional behaviour only.

2. Methods

2.1. Development observation instrument

2.1.1. Procedure

To our knowledge, there is no existing standardised observation instrument for the simultaneous assessment of facial,

body, linguistic and prosodic expressions of emotions. Therefore, we cautiously attempted to develop one ourselves. The

development of this instrument consisted of four phases.

First, we conducted two literature reviews. A preliminary literature review was conducted to clarify the concept of

emotions. This research revealed that aspects of emotions can be divided into three categories ( Frijda, 1986 ): the physiology

of emotions, the emotional experience, and emotional behaviour. We focused on the last one. Emotional behaviour

represents the expression of emotion, which are not mere movements, but various modes of interaction with the

environment. Interaction occurs through nonverbal and verbal expression. Nonverbal expressions are therefore divided in

facial expressions and body movements. Verbal expressions consist of both linguistic utterances (i.e., the content of a

message) and prosodic features (e.g., loudness and pitch of voice). A secondary literature review was conducted to collect

speciﬁc examples of expression of behaviour that contained verbal, facial, or body expressions. Nonverbal elements of

emotional expression were described in multiple studies. Facial movement is the most extensively studied nonverbal

expression. Movements of eyes, eyebrows, and mouth are frequently related to emotional expression (e.g., Dornaika &

Davoine, 2008; Frijda, 1986 ). The second form of nonverbal expression of emotion is interaction through body movement.

Basic emotions can be recognised according to body movements such as shaking of the ﬁsts, raising the arms, and dropping

the head (e.g., Atkinson, Dittrich, Gemmell, & Young, 2004; Frijda, 1986; Navarro & Karlins, 2008 ). Linguistic elements of

verbally expressed emotions came from the CFI scale ( Brown & Rutter, 1966 ). The CFI aims to measure expressed emotions

by assessing the number of (a) critical comments, (b) hostile comments, and (c) emotional over involvement (EOI). Prosodic

features of verbally expressed emotions were deﬁned on the basis of a large body of research in this ﬁeld (e.g., De Gelder &

Vroomen, 2000; Frick, 1985; Frijda, 1986 ). The most frequently reported features of prosody are pitch, loudness, rate, and

quality of voice. Despite large inter-speaker variability, there is general consensus that prosodic features of speech

communicate emotions ( De Gelder & Vroomen, 2000; Frick, 1985 ). The secondary literature review resulted in a list of

behaviour expressions that were found to be related to emotion. The behaviours on the list were categorized in aspects of (a)

verbal, (b) facial, and (c) body expressions.

Following the literature reviews, we conducted interviews with three senior staff members to collect descriptions about

communication outcomes of staff (phase 2). These clinical experts had experience with staffs’ communication of emotion on

a daily basis and were, therefore, considered to have valuable and useful information about speciﬁc behaviour that might be

related to staff emotion.

In the third phase, speciﬁc behaviour descriptions revealed from the interviews were added to the list of behaviours that

had been constructed in phase one. Then, we conducted direct and indirect (i.e., video) pilot observations, to test whether we

2410

W.M.W.J. van Oorsouw et al. / Research in Developmental Disabilities 32 (2011) 2408–2414

could recognise the deﬁned behaviours on the list during observation of staff. The main goal of this phase was to discuss each

behaviour on the list. Was the deﬁnition of this behaviour clear? Was this expression visible during observation of

spontaneous behaviour? And ﬁnally, we discussed whether we observed elements of behaviour that could be relevant, but

were not yet included and deﬁned on the list. This phase also provided information about the use of a video-camera in the

main study, for example the position the camera should be placed in.

In phase four, we compared all the information we had collected so far and conducted an additional literature search. All

behaviours that could not be clearly operationalized in observable terms (e.g., visible heart beat in the neck, making

controlled movements, quiet breathing) were removed from the list. Also, behaviour that had a doubtful or ambiguous

relationship with emotion (e.g., turning head aside, speak quickly, repeating sentences, putting hands on hips) were removed

from the list. This reduction resulted in a list of behaviours that were (a) observable, (b) clearly deﬁned, and (c) had a clear

relationship with emotion. Examples are: clenching hand to ﬁsts, shrugging one’s shoulders while the upper movement

shoulders is visible. This ﬁnal list was the basis of the registration form and observation manual.

2.1.2. Registration of nonverbal communication of emotion

Three main criteria were developed. Criterion one: Only behavioural expressions that were included in the ﬁnal list of

observable, clearly deﬁned behaviours with a clear relationship to emotion (compiled in phase four) were to be registered. A

description of each of these behavioural expressions (e.g., tapping ﬁngers, enlarge eyes) was provided in the manual.

Criterion two: Expressions selected on the basis of criterion one, needed to be executed with a discernible amount of force.

For example, pointing to a direction is not necessarily associated with emotion. Only when the behaviour was accompanied

with a powerful pointing together with a stretched ﬁnger it was coded as an emotional behavioural expression. Criterion

three: Expressions that were not included in the ﬁnal list compiled in phase four were only registered if these expressions

were very clearly related to emotion. We added criterion three to account for the possibility that a clearly emotional

behaviour could occur that had not been included in the list of expressions, due to its unusual character. Sometimes for

example, participants showed complex movements in the neck to indicate their disgust. Such behaviours were registered in

the category ‘Other’. These emotional behaviours also needed a visible degree of force to be coded as an emotional

expression. Fig. 1 shows an overview of the nonverbal expressions that are described in detail in the observation manual.

Fig. 1. An overview of nonverbal expressions included in the manual.

2.1.3. Observation manual

The observation manual explained the criteria for scoring with respect to every behavioural expression in more detail.

Regarding facial and body expressions, speciﬁc descriptions of visible characteristics were given for each zone of the body

(e.g., forehead, mouth, arms). For example, clenches ﬁst and hits the table; shrugging shoulders, with clearly raised visible

shoulders; eyes wide open with a size visibly exceeding its natural width. In accordance with criterion two, nonverbal

movements should be executed with force to be registered for scoring as an emotional expression. Although, a watertight

deﬁnition of forceful reactions does not exist, it was certainly possible to reliably distinguish relaxed and natural movements

from forceful movements. Exclusion criteria for each body zone were also included in the observation manual. For example,

approving, but powerless shaking the head is excluded from scoring as this movement appears to be a listening attitude

W.M.W.J. van Oorsouw et al. / Research in Developmental Disabilities 32 (2011) 2408–2414

2411

rather than an emotional response. There is no clearly deﬁned set of vocal characteristics regarding the prosodic features of

verbal expressions. We decided to include powerfully sighing and sudden change in pitch as important prosodic features.

Based on the review by Van Humbeeck, Van Audenhoven, De Hert, Pieters, & Storms (2002) , scoring criteria for linguistic

utterances (e.g., hostile and critical comments) were based on the scoring criteria of the CFI (see Leff & Vaughn, 1985 ).

2.1.4. Observation system

An example of the registration form is shown in Fig. 2 . The rows of the observation registration form represent

observational categories (i.e., body, facial, and verbal expression) and the columns represent time segments. The

observational system is based on a 10-second partial interval recording. This means that the observer registers whether each

observation category has occurred in each ten second interval.

Fig. 2. Observation registration form.

2.2. First application of the instrument: a pilot

2.2.1. Participants

The team of staff members that was involved in this study consisted of 2 men and 10 women, with an average age of 34

years (SD = 9; range: 24–49). All participants worked as staff members in mental health care systems. The average number of

2412

W.M.W.J. van Oorsouw et al. / Research in Developmental Disabilities 32 (2011) 2408–2414

working experience in mental health care ranged from 0 to 25 years (M = 6, SD = 8). All staff members worked in one team

and served 12 clients with mild ID and challenging behaviour.

2.2.2. Procedure

Permission was obtained from the manager and the staff members to video tape them during work. We chose to video

tape during shift change, because staff members who were ﬁnishing their shift and staff members who were just starting a

new shift got together to discuss the clients and the organization of the day, following regular routines. Staff proximity to

each other during shift change facilitated the collection of data.

Daily video recordings were made for a period of three months. Staff members were not informed about the aim and

content of the study, they were not aware of the fact that researchers were observing forms of emotional behaviour. During

shift change, two or more staff members sat at a table. The video camera was placed on one side of the table and the research

assistant tried to get at least two staff members on the screen. At least the face and the upper part of the body were visible

during 80% of the recording. In general, legs and feet were not on the screen, because video taping of legs and feet required

more distance from the camera, which would be detrimental to the visibility of facial expressions.

The ﬁrst three recordings of every staff member were deleted to control for habituation. Finally, we collected 11 h of video

recordings. The behaviour of individual staff members was observed for an average of 50 min (SD = 33).

In preparation to the reliability assessments, one of the authors extensively trained two Master students in Psychology in

the use of the observation system. During this training, the author and both students paid extensive attention to the

repetition of behaviour deﬁnitions and differences in pilot observations between both students.

3. Results

3.1. Interrater reliability

The ﬁrst aim of the present study was the development of an observation instrument with sufﬁcient levels of interrater

reliability. Therefore, a total of 660 min of video material was collected and interrater reliability was based on 22% (i.e.,

147 min) of the total. The interrater reliabilities regarding the assessment of the three forms of emotional behaviour were

considered to be sufﬁcient, with a Cohen’s kappa of .75 for body expressions, .75 for facial expressions, and .70 for verbal

expression ( Cohen, 1988 ). This means that our observation instrument provides the possibility to reliably observe and

3.2. Correlations between different forms of emotional behaviour

When sufﬁcient interrater reliabilities would be achieved, we would additionally analyse whether the observation of

nonverbal emotional behaviours in addition to verbal emotional behaviour would have a surplue value over the observation

of verbal forms of emotional behaviour only. Pearson correlations between different forms of emotional behaviour are

presented in Table 1 . Results show that all forms of emotional behaviour are signiﬁcantly correlated, with the exception of

the relationship between prosodic features and linguistic utterances. Generally, this means that there is a coherence

between different forms of emotional behaviour. However, it is remarkable that all correlations were relatively small and

consequently, the R 2 values were small. So, although there was a signiﬁcant relationship between most forms of emotional

behaviour, one form could explain very little of the variance of another form of emotional behaviour. This result indicates

that our observations did in fact tap different behaviours and therefore, the assessment of nonverbal forms of emotional

expression has a surplue value over the assessment of verbal of written expressions only.

Table 1

Pearson correlations and R 2 between different forms of emotional communication.

Expression

Body

Prosody

Linguistic

R 2

Facial

.376 **

.141

437

.192 **

.037

437

.102 *

.010

437

Body

.213 **

.045

459

.197 **

.039

459

Prosody

.053

.003

468

Correlation is signiﬁcant at the .01 level (2-tailed).

4. Discussion

In the present study we made a ﬁrst attempt in the development of an observation instrument that measures both verbal

and nonverbal forms of emotional expression. Our ﬁrst aim was to develop an observation instrument from which the

content was based on scientiﬁc research regarding emotional behaviour. Also, we aimed to develop an instrument with

sufﬁcient to good interrater reliability scores. We did not take into account the nature (positive of negative) nor the type

(fear, anger, happiness) of emotions. When sufﬁcient interrater reliabilities would be achieved, we would additionally

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: