In Sweden, the use of statement analysis for evaluating children's accounts of abuse is accepted within the legal system. One advantage of this assessment method is the more structured evaluation procedure, with hypothesis building and testing in a logical order. As a consequence it is possible for an external audience (i.e. the court) to follow and evaluate the assessor's line of reasoning. However, despite its status within the legal system, the Swedish tradition of witness psychology/statement analysis has not been without critics (Christianson, 1992; Lindblad, 1991). Thus, as a part of the research project, one study focused particularly the current use of statement analysis in Swedish courts.
This study (Gumpert and Lindblad, 1999) revealed that different assessors varied with regard to what information related to statement characteristics and contextual circumstances was pointed out and how it was judged. For example, one expert judged the fact that a child's account of abuse expanded between interviews as a sign of unreliability, whereas another concluded that the presence of an expanding story was logical and coherent with other information in the case. Given the basic principle of the Swedish way of using statement analysis (Trankell, 1971; Holgerson, 1990), this diversity may not be in opposition with the theoretical model. Within the hermeneutic framework it is emphasised that 'the question at issue in a particular case must be considered in the light ofthe context ofthat particular case' (Holgerson, 1990, p. 127). However, the courts must be aware of possible variation among experts and cases, and should not regard witness psychologists as homogenous group applying criteria in a uniform way.
Hypothesis testing has strong support in the literature (Dammeyer, 1998; Munro, 1999; Holgerson, 1990; Plous, 1993), and is seen as an 'effective debiasing technique' (Plous, 1993, p. 256). Most experts applying statement analysis worked according to such a procedure. However, a possible logical miscalculation appeared to be related to the way in which experts formulated their hypotheses. The first hypothesis usually suggested that the child's statement referred to either 'self-experienced' or 'real' events, thus indirectly making propositions as to the underlying background. By contrast, many of the alternative hypotheses touched on both the issue of underlying background ('not based on real events'; 'some other person') and on issues of possible disturbing events related to statement history and context, such as the presence or absence of previous influence, leading questions etc. Despite the different focus of the hypotheses, they appeared to be regarded as equally weighted and mutually exclusive (Figure 4.3.1). Such a practice appeared to allow for confusion with regard to what circumstances should be regarded as causal factors as opposed to conditions capable of complicating the evaluation of the child's statement. The way experts posed hypotheses seemed to assume a 'mono-causal' explanatory model, where all the hypotheses were viewed as theoretical explanations for the mere existence of a statement of sexual abuse.
A recurrent discussion in the research on Statement Validity Analysis is the difficulty to interpret low scores, that is, how to evaluate statements that are vague or of low quality in some sense. Suggestive or leading interview techniques may produce answers from children that go beyond their own experiences (e.g. Ceci and Bruck, 1993),
but other studies suggest that such improper interview techniques may also elicit 'low-quality' statements (e.g. lacking detail) (Hershkowitz et al., 1997; Lamb et al., 1997). The need to evaluate interview quality prior to drawing conclusions based on the quality of the child's account has been pointed out.
Interview quality was indeed an issue among Swedish witness psychologists, but in a slightly different way. Several of the witness psychologists in this study observed associations between leading questions and the presence of 'weak' or ambiguous statements, and suggested that the leading or suggestive question had produced a false answer, thus indicating causality. Given possible alternative explanations of the existence of a weak statement, a more reasonable approach would be to evaluate interview technique prior to analysing a statement, and avoid making strong conclusions based on the statement if the interview quality was judged to be low.
Was this article helpful?