IDENTIFYING AND EXTRACTING CAUSAL EXPLANATIONS

Four or more events with explanations are ideally required to generate a valid style. Multiple events are the only way that a researcher can estimate a cross-situational style. Also, multiple events allow explanatory style to be more reliably measured. Peterson, Villanova, and Raps (1985) compared studies that disconfirm the reformulated learned helplessness model with those that support it, finding that the supporting studies had more attributions per subject than the disconfirming studies. Multiple events apparently minimize the effects of the reality of the situation, allowing the individual's habitual style to emerge.

In our research, we usually find bad events with explanations to be more abundant in verbatim material than good events with explanations. What this means is that researchers specifically interested in how people explain good events will have to search more diligently for suitable material for content analysis. When individuals describe events, good or bad, they often end up explaining them, if allowed or encouraged to elaborate their descriptions.