Premium
MANY IMPORTANT ALCOHOL PROBLEMS ARE WIDELY DISPERSED: COMMENT ON SKOG (2006)
Author(s) -
ROOM ROBIN
Publication year - 2006
Publication title -
addiction
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.424
H-Index - 193
eISSN - 1360-0443
pISSN - 0965-2140
DOI - 10.1111/j.1360-0443.2006.01378.x
Subject(s) - population , sample (material) , alcohol consumption , consumption (sociology) , inclusion (mineral) , psychology , class (philosophy) , environmental health , alcohol , medicine , social psychology , sociology , computer science , social science , chromatography , artificial intelligence , biochemistry , chemistry
Concerning Skog's comments on population surveys, it is true that those missing in surveys, at least in developed countries with the usual survey methods, are likely both to be heavier drinkers and to have more alcohol problems than average. The question is how much difference this is likely to make. Losses of representation of the whole population in surveys come from two sources: failure to respond among those sampled, and not being included in the sample. In their recent review, Gmel & Rehm (2004) conclude that ‘non‐response does not seem to be a major source of under‐coverage’ of consumption, and regard non‐inclusion as ‘a more important issue’. That those not included in household‐based samples may include many extreme cases is suggested by the contrast between the dramatic differences often found by social class in mortality for alcohol‐related causes (e.g. Mäkelä 1999) and the gentler social class gradients for survey‐reported problems (e.g. Hradilova Selin 2004, pp. 102–104). How much difference sample biases make is likely to vary with the particular problem considered; but if the main problem is with those not included in household‐based samples, there is an upper bound to the problem—in most developed societies, this would be less than 5% of the population.