Analysing Experimental Design and Analysis Flashcards
most rigorous/least biased evidence-based designs
systematic reviews with meta-analyses (also RCTs)
most rigorous form of experimental study
RCT
when are RCTs used?
to measure effect of an intervention by randomly assigning individuals to intervention or control group
allocation in RCTs
random
unbiased
strengths of RCTs
reduced bias (randomised and blinding)
construct validity
weaknesses of RCTs
generalisation (specific populations)
ethical considerations
when to used ANOVA in research
- Comparing 3 different categories of ppts
- Comparing 3 experimental groups and one control group (on a treatment outcome)
what does ANCOVA look at?
the effect of one or more factors on an outcome/DV
ANCOVA acknowledges what?
the influence of one or more covariates
additional assumptions in ANCOVA
homogeneity of regression slopes and independence of the covariate
characteristics of a covariate
- Chosen based on existing theory and research
- Measured at ratio or interval (ordinal possible)
advantages of ANCOVA
Reduces within-group error variance
Eliminates confounds
In pre-post treatment designs the use of covariates yields more precise estimates of effect size
when is ANCOVA usually used in research?
variables
- Often used to compare results in a pre-/post- test study design
○ Pre-test scores are used as the covariate
○ Post-test scores are our outcome measure (DV)
○ Treatment group is the factor we are manipulating (IV)
issues with change scores
- Outcome-baseline=change score
- Pre-test scores highly correlate with change/ difference scores.
- Regression to the mean: people with low scores at baseline will generally improve more than those with high scores.
assumption of independence of the covariate in ANCOVA
- Covariate should not differ sig. between groups
- Should be independent of the treatment effect
○ Otherwise treatment effect can be obscured
§ Often issue when random-assignment is over-looked
- Should be independent of the treatment effect
