These statistical jargons are presented as how I understood them in terms of research for psychological treatments but is generally applicable to the field. Tough learning especially when quantitative research/statistical analysis is not my strong field at all. Slightly uncomfortable about reducing individuals to subjects to be studied and talking about them as though they are something to be measured rather than each having their idiosyncracities.
Effect size : small, medium, large-
-more commonly reported in the social science research when reviewing literature’s results (if I understood correctly) in addition to statistical significance ie. p values.
between and within subjects paradigms
- between subjects is the measure of outcome for examples, across different psychological treatment
- within subjects is the measure of process for example, difference pre and post treatment
- important for issues of internal and external validity. At the moment, I only understand external validity as whether the results is generalisable to the population studied.
Edit: internal validity is whether the effect studied is indeed the one affecting the results and not some other unwanted variable. For example, in comparing a psychological therapy to waiting list- results may show a positive outcome for those being seen by a therapist but is this due to the psychological treatment per se or other factors such as further deterioration of mental health due to the passing of time being on the waiting list.
Participant used for final analysis : intention-to-treat vs completers
- Intention-to-treat is an inclusion of all participants that were originally assigned to an experimental condition into final analysis. .
- Completers analysis includes only participants that completed the course of psychological treatment. This may present bias results as individuals who do not complete a psychological treatment for example may have dropped out because they don’t feel it was working ie. negative outcome
Ok, this is as far as I have gotten at the moment in what I use in my review of the literature. I am aware there are so much more to learn. For example, different kinds of statistical analysis may have their own pitfalls for example, random effects analysis may yield a more conservative measure of combined effects. Not even going there at the moment.