Photo: Markus Spiske

Science about science

Photo: Markus Spiske

Science about science

I have written on ‘p-hacking’, the phenomenon wherby people fudge the numbers in various ways in order to get the coveted p < 0.05. Our work suggests that p-hacking is widespread: there are suspicious bulges in the distribution of published p-values.

I have also written about blind methods, i.e. where the experimenter (and the subjects, if they are people) cannot identify the treatment groups, and thus consciously or unconsciously add bias. Our work found that blind methods are used less often than they should be, and that it really matters: non-blind studies tend to yield results that fall in the predicted direction.

Most recently, I have been working on the ‘gender gap’ in science. See the gender gap project for more inforation.

Avatar
Luke Holman
Senior Lecturer

Publications

We find that researchers tend to co-publish with same-gendered colleagues, and that this tendency is presently increasing. We found no …

We synthesised 459 effect sizes from 65 pertinent experimental evolution studies using meta-analysis, and found that sexual selection …

We recorded the gender of 36 million authors from >100 countries publishing in >6000 journals from most STEMM, and made a web app …

Observer bias and other “experimenter effects” occur when researchers’ expectations influence study outcome. These biases are strongest …

A focus on novel, confirmatory, and statistically significant results leads to substantial bias in the scientific literature. One type …