Measuring explanatory style in children.

Abstract

Causal explanations that individuals use to explain events in their lives are referred to as explanatory style. Three dimensions: internal-external, stable-unstable, and global-specific have most frequently been measured. Internal, stable, global explanations for negative events represent a pessimistic style, whereas these same explanations for positive events are considered optimistic. Explanations for negative events that are stable and global are considered to reflect hopelessness. The psychometric properties of the most commonly used measure of explanatory style for children, the Children Attributional Style Questionnaire (CASQ; Kaslow, Tanenbaum & Seligman, 1978) are poor. This is a limitation to research and theoretical advancement. Four studies were conducted in this project to investigate the measurement of explanatory style in 9-12 year old children. In Study 1, children (N = 173) completed the CASQ in a group to investigate the psychometric properties of the composite scales and subscales and the relationship between explanatory style and depressive symptoms. Internal consistency and inter-item correlations of the composite scales and subscales were poor. Regression analyses showed explanatory style for negative events (pessimism or hopelessness) made weak but significant unique contributions to the explanations of depressive symptoms. Study 2 (N = 72) investigated the stability of the CASQ scales longitudinally. The internal consistency and inter-item correlations for the CASQ scales were poor. The stability of explanatory style was low. The predicted relationship between depressive symptoms and explanatory style was found to be inconsistent, emerging at Time 1 but not at Time 2, 12 months later. Study 3 (N = 79) examined the forced-choice response scale of the CASQ using a fuzzy set approach. A fuzzy set scale which uses a Likert-type response that ranged from completely true to completely false was used to determine how well a child’s response of choice, their natural response, matched both the selected and non-selected response from the CASQ. Items on the CASQ that measure both pessimism and hopelessness were found to be a poor match to the natural responses of children. Little separation was found between the selected and non-selected responses for all items. The internal consistency of the CASQ was poor when the forced choice scoring approach was used. When Likert-type fuzzy values were used, good internal consistency was obtained. Providing a wider range of responses, obtained using fuzzy values, produced a more sensitive measure of the components of explanatory style. When the CASQ was scored according to the forced choice protocol weak, significant relationships were found between explanatory style and depressive symptoms, and explanatory style and neuroticism. There were no significant relationships found for either pessimism or hopelessness, with either depression or neuroticism using Likert-type fuzzy values. Study 4 elicited spontaneous causal explanations following success or failure on tasks that were familiar or unfamiliar. Task familiarity was manipulated. Using an interview format, children (N = 111) responded to questions, eliciting causal explanations, following task success or failures. Likert-type scales measured the internality, stability or globality of the explanation. Results showed that, following failure on two familiar tasks, acceptable levels of internal consistency were obtained on the subscales used to produce the measure of hopelessness and for the composite measure of hopelessness. This same pattern did not emerge following failure on combinations of familiar and unfamiliar events or on two tasks that were unfamiliar. Stable and global explanations and the composite measure of hopelessness, following failure on familiar tasks, were also positively related with depressive symptoms but not neuroticism. These results show that a reliable measure of hopelessness can be obtained from spontaneous explanations for failure at familiar events. Under these conditions the theoretically predicted relationship between explanatory style and depressive symptoms emerges. Conclusions were drawn about the theoretical conceptualisation of explanatory style and measurement recommendations were made that apply to 9- to 12-year-old children. Explanations for familiar events produced a consistent measure of explanatory style. The use of a Likert-type response scale to assess agreement with internal, stable, global components were shown to improve scale reliability. The findings are discussed in relation to theory and the measurement of explanatory style in children.