0
$\begingroup$

I am conducting research for my master's thesis regarding the phenomenon of Hindsight Bias. I have four experimental conditions in which the DVs include foreseeability and inevitability, and they are measured three times each in each experimental condition. Both DVs are measured with a 5-point Likert scale. In that case, you can say that both DVs are repeated measures.

My hypotheses make pairwise comparisons between the experimental conditions. In some cases, for instance, because some hypotheses only concern the effect of a specific manipulation on one of the DVs (for example foreseeability), I only compare the concerned experimental conditions with each other regarding that one DV. In such a case, I use a one-way repeated measure ANOVA in which the concerned DV (measured three times) is the within-subject factors and the concerned experimental conditions are the between-subject factors.

In one hypothesis regarding two experimental conditions, I expect that both DVs are being influenced at the same time by the manipulation. For such a situation, I wanted to integrate both DVs into the statistical analysis and I can't seem to find a proper statistical procedure in SPSS that considers both DVs and their potential interaction, their respective repeated measures and the between-subject factor (the experimental conditions) at the same time.

Does anyone have a piece of advice concerning the aforementioned issue? I would be really grateful for your help. I hope my explanations were somewhat clear enough.

$\endgroup$

1 Answer 1

0
$\begingroup$

Assuming you're satisfied with a linear model (as indicated by your stating that you've used a one-way repeated-measures ANOVA), the GLM procedure with repeated measures lets you specify between-subjects factors like your experimental conditions, within-subjects factors such as time (and there can be multiple of these where appropriate), as well as specifying multiple measures or dependent variables. The resulting output includes tests for each dependent measure, as well as tests applying to all measures at once (to test whether there are effects on any of them). If you have specific hypotheses you want to test about relationships between experimental conditions and their effects on different dependent measures over time, you can use the general formulation LBM=K (K usually being 0 but not necessarily) to apply a transformation matrix M to the dependent measures and get estimates of particular mean differences and tests of specific hypotheses about those mean differences. The general setup can be done via the menus in Analyze>General Linear Model>Repeated Measures (look in the initial dialog for the Measure section to specify names for multiple dependent variable sets), but applying the general linear formulation for estimates and tests for very specific situations may require using command syntax (LMATRIX, MMATRIX, and possibly KMATRIX subcommands).

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.