Articles to 2017-08-17

Zum Seitenende      Übersicht Artikel      Home & Impressum

First the link to this week’s complete list as HTML and as PDF.


Curtis et al give us a) no data, b) no diagram, and c) no dimension on their regression coefficients. The latter makes it impossible to compare effect sizes. The age effect for example looks especially small but highly significant. The answer here seems to be a large abscissal range of up to ± 20 a versus coding variables ranging from zero to two. This lack also means we get no information about real effect sizes, is it cases per year or what? Most importantly we get no baseline to compare the small predictive value to. It’s said to be about half, but half of what? We’re not told. Second semester exercises are summarily thrown out for fewer faults.

Obesity and blood pressure are eliminated from the analysis and we are not told their relative effect sizes. Interestingly living in a partnership seems to be twice as bad for the heart as smoking and needs nearly two extra hours of sleep to compensate for. Is the whole result perhaps an artefact of fine grained data? Smoking on a yes/no scale is not significant, exercise and sleep with a resolution of minutes per day are.


The abstract of Farr & Patterson 2013 can easily be summarised in one sentence: “Women are good, men are bad, and homosexual men are just as bad as heterosexual ones.” Reading the full results we find a lot about parenting and family life but very little on how it affects the children. Talking about differing levels of contribution “parent A” in heterosexual couples is always the woman and in homosexual ones “the parent who replied first”, an essentially random or at least unrelated variable. So it comes as no surprise, that distribution of labour and satisfaction with that distribution look more even in homosexual couples with the average across both essentially the same.

One strong predictor on child welfare is dissatisfaction with the parenting role regardless of how uneven the split in contribution actually is. This holds for all three kinds of couples and in the heterosexual ones it tends always to be the mother who holds those detrimental grudges – as already said regardless of actual work load.

A second if weak and non significant result is teachers but not parents reporting difficult child behaviour with more competitive parents, a male trait. It is a well known and well documented fact that teachers tend strongly to prefer female traits in children, so that this result very probably is nothing to with child welfare but rather with teacher preference for an easy life. If anything, competitiveness and ambition are positive if sometimes taxing traits in children.


The encouraging result in Farr et al. 2017 is, that children from homosexual families do not differ from the general population in the gender conformance of their behaviour. That said they arrived at this result through rather a strange procedure. They did not compare (adoptive) children of same sex parents to those of their heterosexual control group but rather all of the children in their study, half of which constitute the control, to the general population. Still, looking at their more detailed tabulation there is nothing to contradict it.

There is an interesting if very weak secondary result. Female parents seem to adhere more strongly to a feminist agenda than males regardless of sexual orientation and tend to encourage and instigate gender non-conformance.


As so often happens in the humanities, Lehar does not tell us all the parameters and data needed to reproduce his results. The order of magnitude seems to fit, if his "November to April" comprises five months. The amount of heating needed is not implausible either, but he seems to be using one single value as the continuous power requirement over the whole of the heating season without telling us, how that average was arrived at.


Frei et al. lists 13 co-authors who, as authors, all bear full responsibility and should all have carefully proof read the draft. Not one of them noticed the femur measured at nearly four and a half meters in length. Of course this is a small mistake and easily corrected, but such a glaring dereliction of duty makes one wonder, what other, bigger, and less easy to spot errors they may also have missed.

As always I exempt the first author. Mistakes can and will happen and proofreading your own words is one of the hardest tasks of all. This is where editors, referees, and co-authors come in. You either commit to doing something or you don’t, but if you do not, then don’t go and claim the glory for yourself. There are far too many recent retractions, where co-authors deny all responsibility and try to wash their hands of it.


Bello et al. report on one single bone, the significance of which had already been suggested some time ago and dismissed on closer examination. That said their result does look sound, but it remains so singular, that I have my doubts as to its wider relevance.

Zum Anfang      Übersicht Artikel      Home & Impressum

Creative Commons Attribution-Share Alike 3.0 Unported License Viewable With Any Browser Valid HTML 4.01! Valid CSS!