Development and validation of a new scale to assess air quality knowledge (AQIQ)

Air pollution is a leading public health concern around the world. Assessing the public's knowledge about air quality is critical to calibrate public health interventions. However, previous efforts to measure knowledge about air quality (AQIQ) have not relied on consistent and validated measures, thus precluding cross-country comparisons. We aimed to develop a robust scale to assess AQIQ and tested it in multiple countries. To evaluate the psychometric properties and select the best performing items out of 10 AQIQ questions, we used methods from classical test theory and item response theory. We evaluated the scales using several scalability measures, including the Kuder-Richardson Formula 20 (KR-20), Loevinger's H, as well as trace lines. Volunteers from the United States (US, n = 400), India (n = 403), and China (n = 443) were recruited to validate the scale. Multiple linear regression was used to estimate the association between demographic factors and AQIQ. We found that participants from India had the highest AQIQ. In addition, not all questions performed well in each country. The scale was pruned and shorter subscales were validated. In the US, we obtained a 4-item scale (KR20 = 0.53, Loevinger's H = 0.34). In India, we obtained a 6-item scale (KR20 = 0.56; Loevinger's H = 0.48 for just 2 items). In China, we obtained a 5-item scale (KR20 = 0.39; Loevinger's H = 0.41 for just 2 items). Compared to the 10-item scale, the pruned scales showed stronger associations between measures of socioeconomic status and AQIQ. The results were robust to the scale used. Overall, general knowledge questions measured AQIQ more effectively in the US and India whereas knowledge of the air quality index better measured AQIQ in China. The findings suggest that careful measurement and validation are essential to develop knowledge scales for use in public health and environmental research.