The equivalence of contingency structure for intuitive covariation judgments about height, weight, and body fat

In a complete between-subjects design, 576 respondents judged covariations of problems framed around height, weight, and body fat. A verbal measure was used, consisting of propositional statements representing relationships between two of the variables, as well as two types of response options: true or false, and a 7-point, bipolar, Likert-type scale. Consistent with previous findings, weight and body fat were unambiguously judged as if positively correlated; there were indications of a perceived negative correlation of height and body fat; and there was no evidence that participants regarded height and weight as positively correlated. Furthermore, there was no reliable evidence in support of framing effects. However, there were indications of influence for conditional format. These findings are discussed relative to previous and ongoing research efforts, and it is concluded that it may be inappropriate to regard variations in conditional format as parallel items. Measures involving all possible variations are considered fundamentally unbiased and consistent with I. Ajzen's (1988) principle of aggregation.