1. “cause-and-effect relationship … the average rate of beer drinking among studied subjects is not much different from the Czech average”. Stop and think for a while – what has population average (for any data set) to do with the relationship (between any variables)? Nothing, of course. What matters is the spread of data which you cannot determine from the Fig. 1 (because I used Box-Cox transformation; see http://en.wikipedia.org/wik…).

2. “it seems unlikely that the heaviest drinkers in the study rose to the level of alcohol abusers”. But “unlikely” just means that you just do not know. Moreover, for a particular person drinking just 2 or beers may have the effect as 10 beers for another person. Anyway, I would be happy to see your mental productivity with a hangover after a night spent with some 5 liters of beer (I mean Czech beer, not the US crap:-). Just one night per week like that would decrease your productivity by some 20% in comparison (assuming you work 5 days/week) to another similar person that would not drink (so much or at all). I am eager to hear an argument that rejects that logic:-) (or the results of hundreds of studies of alcohol-related defects in human cognition).

3. “five highly influential beer guzzlers controlled the fit”. Incorrect: I (just for fun) removed those people, recalculated the regression and the relationship was just close to the sacred cow named 0.05. But in my sample I would prefer to remove a different data points than you – I am joking of course.

4. “the total number of people involved in the study looks to be 18”. Wrong. In 2006 I got 34 data points INCLUDING 18 old and 16 NEW researchers. See the bloody Fig. 1 properly, man!:-)

5. “The correlation coefficients are just barely large enough to claim with 95% confidence that the results are statistically different from an R^2 of 0”. This is bullshit. PLEASE, read some statistical textbook first…

6. “There is no way to know whether all confounding factors (read systematic errors) have been eliminated.” Agreed – just like in ANY ecological/behavioural study:-)

7. “this small group of Czech ornithologists socialized together”. I checked those 5 people you were so much interested in – they work at 5 different institutions in 3 different towns. So they unlikely drink every night together:-)

8. “small” data set. The sample of that magnitude (34) is quite good for an ecological study,
in fact! (see any behavioural ecology journal and you will see that my sample is perhaps even above average:-) More importantly, I surveyed Czech avian ecologists and I had 90% (!) success in getting response from my respondents – thus, I had data for almost the whole population I studied! From this "sample quality" view my study is very much above average, of course:-) But still my work was just a preliminary survey which should be followed by wider sampling across countries and scientific professions, of course.

9. The last paragraph is the only one in you text which is close to reality, sorry:-)

Cheers, Tom