3
$\begingroup$

In a basic research design where you want to test how some intervention (let's say our IV is 'drinking' with the categories

  1. water;

  2. sprite; and

  3. rum)

affects some continuous measurement (let's say our DV is 'math performance- in time) theoretically you could run the experiment as:

participants get 1 of 3 drink types, then their math performance is measured but I'm guessing that to be able to say 'the IV affects the DV' it would be better to test the DV before the intervention AND after it, so your design would look something like:

-test for math performance (time 1)

-randomly assign to drink water, sprite, or rum

-test for math performance again (time 2)

my question then is would it be statistically better to do A), B), or C):

A) a one-way between-subjects ANCOVA with 'test for math performance at time 1' as the covariate, the 'random assignment to 1 of the 3 drinking categories' as the IV, and the 'test for math performance at time 2' as the DV

B) a two-way repeated measures ANCOVA with ''test for math performance' as a repeated factor (levels of time 1 and time 2), and 3) the 'random assignment to 1 of the 3 drinking categories' as a between-subjects IV or

C) a one-way between-subjects ANOVA with the 'random assignment to 1 of the 3 drinking categories' as a between-subjects IV, and the difference score (performance time 2 - performance time 1) as the DV

$\endgroup$
  • $\begingroup$ Since, the natural variation in between each individual's performance is likely to be high. Option C, the match pair test, comparing the change from before to after is most likely to provide a significant result. $\endgroup$ – Dave2e Feb 10 at 18:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.