Research methods and statistics
Dec 13, 2015
Learning outcomes
At the end of this session and with additional reading you will be able to: Critically evaluate a research paper
Internal validity is concerned about the causal-effect relationship in a study◦ Can observed changes be attributed to your
programme/intervention or are the results influenced by alternative causes?
History – some historical event that occurred Maturation – events over time that aid
learning Testing – pre-testing primes the participants
for what is to come (practice effect) Instrumentation – is there any difference in
the pre-test/post-test instrument Mortality – have any participants dropped out Regression- pre-test scores are always likely
to improve on a post-test
Have a control group that does not have the programme/intervention◦ Placebo effect in medical trials
Selection bias – the degree to which the groups are comparable before the study◦ Selection history threat◦ Selection testing threat◦ Selection maturation threat◦ Selection instrumentation threat◦ Selection mortality threat◦ Selection regression threat
Diffusion of treatment – participants share their experiences with other participants
Compensatory rivalry – groups try and compete with one another
Resentful demoralisation - the comparison group under achieves due to jealousy with the other group
Compensatory equalization of treatment - the researcher tries to make one or both groups feel special
External validity is concerned with generalisation◦ Can you generalise your results to a wider
population Sample modelling – you first identify the
population you want to generalise to and then draw a sample from that population
Proximal similarity (Gradient of similarity) – identify populations that are more or less similar to the original study
Use a random sample Keep drop out rates low Replicate the study with different people at
different times in different places