Socioeconomic factors and suicide rates at large-unit aggregate levels : a comment

Can socioeconomic factors seemingly explain variation in suicide rates at large-unit aggregate levels only due to an ecological fallacy? This is what Kunce and Anderson (2002) suggest based on fixed-effects estimation of US state suicide rates, in which they find little evidence that socioeconomic factors matter. We demonstrate that this result does not hold true for other large-unit aggregate levels in our analysis of suicide at the cross-national level. We find that many socioeconomic factors have a statistically significant impact. We conclude that sociological and economic theories explaining variation in suicide rates at the large-unit aggregate level with the help of aggregate socioeconomic factors cannot simply be dismissed because of an alleged ecological fallacy.