• Keine Ergebnisse gefunden

Criteria for the review on coding systems

Im Dokument Understanding Body Movement (Seite 75-78)

3. Movement Behaviour Analysis across Scientific Disci- Disci-plines

3.2 Criteria for the review on coding systems

The fact that researchers who have no experience with movement analysis tend to develop their own coding systems has mainly negative consequences for the quality of research. Furthermore, as already discussed in Chapter 1, this trend, which leads to a multitude of different movement types, also inhibits the com-munication in the scientific community. The review in subsection 2.5.3 on later-ality preferences for different movement types in healthy individuals reveals the difficulties that emerge when trying to compare findings of different researchers.

Because of the scant interdisciplinary exchange, researchers are often not aware that a substantial number of movement analysis systems already exist. Among these, they may find a tool that suits for their purpose. While the existing sys-tems need not to be perfect, they offer at least a methodological experience and a basis to start with.

Therefore, this chapter provides a review on behavioural movement analysis systems. The review focuses on movement analysis tools that are designed to explore the movement behaviour as a reflection of emotional and cognitive, and interactive processes. Systems that directly code the interaction, i.e., between-subjects processes such as mirroring or synchronicity, (e.g. Davis & Hadiks, 1990), are not included in the review. As demonstrated in Chapter 2, the exami-nation of patients with mental disease and brain damage provides valuable in-formation about the link between movement behaviour and emotional, cogni-tive, and interactive functions. Therefore, also coding systems are reviewed here that are designed to code alterations of movement behaviour in patients with mental disease and brain damage.

As exposed in Chapter 1, movement behaviour research is spread across tific disciplines. Not surprisingly, the coding systems used in the diverse scien-tific disciplines differ substantially regarding the criteria on the grounds of which the movements are classified. In their paper "The Repertoire of Nonver-bal Behaviour", Ekman and Friesen (1969) list the following criteria: (i) use of a movement: context, relation to language, consciousness and intent of the non-verbal actor, external feedback, informative, communicative, or interactive value of the movement; (ii) origin of a movement: reflex-like, universal, cultural, spe-cific to a group, familiar, or individual; (iii) relationship between the movement and its meaning: arbitrary (there is no visible relation between the movement and its meaning, for instance, Churchill`s Victory sign), iconic (the movement is

similar to its meaning, for instance, the gesture "cutting someone`s throat"), in-trinsic (the movement is part of its meaning, for instance, showing the fist with anger). Many more criteria could be added to their list, such as the kinematics of the movement, the laterality of a limb movement, the psychodynamic function of a movement, etc.

In the following review, the coding systems are classified in three main groups, based on their predominant criterion for classifying body movements.

This order of presenting the coding systems, instead of ordering them according to academic disciplines, also serves to overcome the borders between the scien-tific disciplines. It clarifies that similar approaches to movement behaviour analysis have been developed in different academic disciplines.

(i) Comprehensive descriptive coding systems

These systems are descriptive in the sense that they refer to the visually perceiv-able aspect of movement behaviour. They describe not only what type of a movement is performed, e.g. a pointing gesture, and how it is performed, e.g.

pointing with strength. Comprehensiveness is another characteristic of these coding systems. All movements of several parts of the body or the body as a whole are coded.

(ii) Coding systems that classify movements according to function

These systems classify body movements according to specific functions, e.g.

meaning or psychodynamic function. As described in section 2.1, different func-tions are typically attributed to different movement classes, such as cognition to gesture, psychological balance to self-touch, mood to posture, etc.. These classes are, in turn, regularly associated with different parts of the body or with the body as a whole. Therefore, the function-oriented analysis systems are often re-stricted to coding one part of the body, such as the hands and arms.

(iii) Coding systems that classify alterations in movement behaviour

These systems register alterations in the quality, the quantity, and the concept of body movements, as compared to their performance in the movement behaviour of healthy persons. These movement alterations are typically observed in indi-viduals with mental disease or brain damage. Given the broad spectrum of pathological alterations, the systems vary substantially among each other.

To illustrate the practical problems that arise when ineffective coding systems are employed, this last review not only describes the coding systems but it also demonstrates the difficulties that may arise from deficient scales. That part of the review demonstrates that an ineffective methodology is not only a blemish, but that it may result in misleading or even wrong results, which in the example of a clinical context, have negative consequences for the patients' quality of life.

As concluded in Chapter 2, to fully understand movement behaviour and its link to cognitive, emotional, and interactive processes, basic empirical research is required. Therefore, the review below focuses on coding systems that are objec-tive and reliable.

To check for objectivity, the author has studied the definitions of the values in the respective literature. Note that in the review below, the short definitions of the movement categories and values7 that are placed in brackets can only sketch the concept of the respective category and values. For obtaining the detailed definitions, the reader should study the respective literature given in the refer-ences. A good measure for objectivity is interrater agreement.

With regard to reliability, it has to be noted that the control of interrater agreement in movement behaviour research is a rather recent standard. Accord-ingly, for some systems, there are no reports on interrater reliability. Even more so, studies on retest reliability are the exception, i.e., if after a certain period of time the observer gives the same value to the same movement (applies to video-taped movements). For reviewing the coding systems, another problem oc-curred: Those studies that report interrater agreement have used different study designs. The design of a study can facilitate or complicate the achievement of good interrater agreement. As an example, it is easier to achieve a good agree-ment when coding preselected units than when segagree-menting and coding the ongo-ing stream of behaviour. Furthermore, different statistical methods have been employed to measure interrater agreement. As an assessment of the equivalence of these statistical procedures is far beyond the scope of this chapter, in the re-view below it is only reported if interrater agreement and retest-reliability had been examined. If no reports on interrater agreement were available but the cod-ing system had been referred to or used by several other independent research-ers, it was nevertheless included in the review.

Concerning validity, it shall be noted that for none of the existing systems the validity can be judged as sufficiently established with regard to current psycho-metric standards. This does in no way imply that these systems would not prove validity if they were properly tested. However, the extended review of the cod-ing system for movement psychopathology illustrates the dilemma that arises when the validity of certain values is uncritically assumed without sufficient psychometric verification.

Given the aim of this book, the following review on coding systems is not ex-haustive. It is rather intended to provide an overview on the spectrum of differ-ent methodological approaches to movemdiffer-ent behaviour analysis. The coding systems are presented and discussed concerning those factors that have been identified in Chapter 2 as being relevant for basic empirical research on move-ment behaviour.

7 Throughout the book, values are written in italics.

Im Dokument Understanding Body Movement (Seite 75-78)