Abstract
Data from the Statistics Canada 2005 Canadian Community Health Survey is used to test the hypothesis that classification errors of the type noted by Fillmore et al. (Citation) could invalidate the statistical results on the effects of alcohol consumption on self-rated health and the incidence of heart disease and diabetes. The results obtained in this study show that the beneficial effects of moderate alcohol use that so many studies have found, still appear even when the correct classification of alcohol use is employed. However, parameter biases and inferential errors can occur when researchers fail to distinguish between former drinkers and never drinkers within the non-drinking group.
Notes
Notes
1. Fillmore et al. (Citation2006) raised the possibility that the results in MMY indicating a beneficial effect of moderate alcohol use on various health measures might actually be wrong. This was, of course, the motivation for the current study when the 2005 data became available. But it quickly became clear that this issue was one that involved a large body of important research and, therefore, needed to be carefully investigated.
2. If moderate alcohol use reduces the risks of death from other serious illnesses like cancer, stroke or diabetes this will bias the effects of alcohol use on coronary heart disease towards being less effective or even being positively associated with heart disease since some of the moderate drinkers were prevented from dying from the other diseases. When specific mortalities are the subject of the investigation the analysis should be carried out in a competing risk framework. See van den Berg (2005).
3. To make the model identified q1 must be set equal to 0.
4. A drink is 12 oz of 5% beer, 5 oz of 12% wine or 1.5 oz of 40% spirit.
5. The results for this simulation were based only on drinkers, that is, respondents who claimed to be occasional or regular drinkers. The optimum number of drinks was calculated using the formula where and sd are the mean and standard deviation of the number of drinks per day consumed by drinkers.
6. For this case it is interesting to note that failing to distinguish former drinkers from abstainers would have generated erroneous beneficial effects for moderate alcohol use.