These are some abandoned ideas that are free for others to re-use. These were things that I thought about, but now no longer have the time to pursue. Some of them are getting a bit old, and I haven’t checked to see if they have already been published by others.
I stole this idea from the economist Paul Frijters who put his own abandoned ideas on his blog. I use the same approach as Paul:
“There is no copyright on these ideas. Anyone is welcome to use and develop them further. If you do use them, mentioning the origin of the idea would be greatly appreciated.”
Statistics / epidemiology
- Two-stage estimates in environmental epidemiology. I’ve used a two-stage regression model to estimate the additional effects of an environmental exposure. For example, in this paper we used a two-stage model to estimate the effects of temperature and on deaths, and then examined the additional effect of heat waves in a second stage. The problem is that the first stage estimates are carried forward with no uncertainty. This could easily be fixed by using a bootstrap or Bayesian model in the first stage and re-running the second stage multiple times using lots of first-stage estimates.
Meta-research
-
Which journal first moved materials and methods to an appendix? When did this happen and why?
-
Does the time between a paper being received and being published impact on its subsequent citation count? The idea is that papers which sail through peer review in quick time may be “better” (on average). A quick first look at this shows that errata throw a spanner in the works, as they interfere with the study’s dates.