Dear,
I am using Linear Mixed Modeling in SPSS. I have a significant interaction between two variables: correctness (2 levels: correct vs error) and condition (3 levels: drug a, drug b, drug c). In order to follow up on the significant interaction I run pairwise comparisons and ask SPSS to give me the Bonferroni corrected p-values. As I have 6 comparisons (drug a vs b, b vs c, a vs c within correct and within error) I was expecting that the provided p-values would be six times higher compared to the p-values when I do not do a correction (the LSD p-values)
Instead, the Bonferroni corrected values are 3 times higher. Does anyone know what causes this difference? And what is the best 'correction' to report? Should I report the LSD p-values * 6? Or are the Bonferroni corrected p-values provided by SPSS sufficient?
Many thanks,
Des
I am using Linear Mixed Modeling in SPSS. I have a significant interaction between two variables: correctness (2 levels: correct vs error) and condition (3 levels: drug a, drug b, drug c). In order to follow up on the significant interaction I run pairwise comparisons and ask SPSS to give me the Bonferroni corrected p-values. As I have 6 comparisons (drug a vs b, b vs c, a vs c within correct and within error) I was expecting that the provided p-values would be six times higher compared to the p-values when I do not do a correction (the LSD p-values)
Instead, the Bonferroni corrected values are 3 times higher. Does anyone know what causes this difference? And what is the best 'correction' to report? Should I report the LSD p-values * 6? Or are the Bonferroni corrected p-values provided by SPSS sufficient?
Many thanks,
Des