Tuesday, 19 June 2012

How not to do a scientific study.

Today the media have been talking about a study that shows there are dangers with people working too hard and becoming "screen slaves". Here is the BBC report with the usual press release sound bites.

At first glance this looks a nice example for my statistics course of carrying out a health related observational study. So I wanted to see the details. It is an online survey of 2010 office workers (online always gets me worried to start with). There is no link to the survey on the BBC page so I go to the Chartered Society of Physiotherapy page and find that the study is all related to their "Workout at Work Day". On their website there is even less detail about the study. They only report the study's findings and then only very sketchily and with no warnings about reliability, and no analysis at all.

One of the society's aims is to get physiotherapy recognised as a medical profession. Producing "puff" articles like this, where you do not give the content and you do not present data in a scientific way, is not the way to do it. It does not look professional and it is not professional from a scientific view-point. I am really disappointed in what they have done, this is a missed opportunity and reflects badly amongst medical professionals.

At least it does provide a case-study for my teaching next year, as an example of how NOT to do a survey, and how NOT to present your findings.

No comments: