Multcompare function for post-hoc test in 2-way repeated measures ANOVA

37 次查看（过去 30 天）
Sofia Moraresku
Sofia Moraresku 2021-5-7
评论： Scott MacKenzie 2021-7-28
Hello, I would like to ask about the correctness of the results of multcompare function for 2-way repeated measures ANOVA.
I performed 2 way repeated measures ANOVA and used function multcompare for post-hoc tests in MATLAB (by default it uses 'ComparisonType', 'Tukey-Kramer'). I also performed the same analysis in the program STATISTICA (from TIBCO) using post-hoc Tukey HSD, which should be the same as the default in mulcompare. I've got the same F-statistic and p-values in ANOVA results in both programs, but very different outputs of post-hoc Tukey tests.
Here is the part of the code I used in MATLAB (two within-subject factors: Time and Condition).
% specify that all variables are in the same group - repeated measures
repeatFactors = table(factor1,factor2, 'VariableNames',{'Conditions' 'Time'});
% create a model to be fit with repeated measures
rm = fitrm(data,'control_0_2-allo_1_0 ~ 1','WithinDesign',repeatFactors);
% fit the model
ranova_table = ranova(rm,'WithinModel','Conditions*Time');
% multiple comparisons across groups
Tukey_table = multcompare(rm,'Conditions', 'By', 'Time');
So, I wonder what the reason could be for the different post-hoc results and if I'm using the function multcompare correctly with 2-way repeated measures ANOVA.
Thanks!

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Analysis of Variance and Covariance 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by