Reliability of a questionnaire used to survey allergic disease in school

The reliability of a questionnaire which contained 21 items concerning asthma and allergies of the nose, skin, and eyes and other questions was evaluated by test-retest method. The questionnaire was the same as one used two years ago in a survey of allergic diseases found among elementary, junior and senior high school students in Shizuoka Prefecture, Japan. In order to evaluate its reliability, calculations were made of proportions of agreement, Cohen's kappa values and intraclass correlation coefficients. Both the proportions of agreement and the kappa values were fair, and all the intraclass correlation coefficients showed high values. The results suggest that there might be a slight effect of age, and that the articles which showed higher kappa values tended to be easy to answer, while the items in which kappa values were lower tended to consist of questions of multiple answers. This questionnaire can be regarded as useful for our original purpose of investigating allergic diseases. Among the kappa values of items, the values regarding diagnosis by doctor tended to be the largest. It was suggested that doctor's diagnosis was strongly convincing for patients. The reliability of this questionnaire survey can be regarded as satisfactory.