Reader Comments

Post a new comment on this article

Self-selection and sampling

Posted by pst1 on 19 Sep 2014 at 20:39 GMT

Opt-in online surveys are infamous for their propensity for self-selection. Can you explain how you control for that?
If self-selection is not an issue how do you explain 77.5% female participation, which is far from the known gender distribution in academic and research institutions in the US (http://www.nsf.gov/statis...)?

How are the respondents distributed across the institutions and places of work surveyed? Is there a particular campus with high number of respondents?

Can you comment on the number of research institutions that were included in the survey?

No competing interests declared.