Would you happen to know if there is an alternative statistical test (e.g., a non-parametric test) that could be used in place of an ANCOVA if your data-set violates the homogeneity of regression slopes assumption?
connorism6 Quade’s test is what you need. 1) Rank the dependent variable and any covariates, using the default settings in the SPSS RANK procedure. This is done for all cases, ignoring the grouping variable. 2) Run a linear regression of the ranks of the dependent variable on the ranks of the covariates, saving the (raw or Unstandardized) residuals, again ignoring the grouping factor. 3) Run a one-way analysis of variance (ANOVA), using the residuals from the regression in the prior step as the dependent variable, and the grouping variable as the factor. The F test resulting from this ANOVA is the F statistic Quade used.
Hi. Could I do one analysis with two covariants? For example: put as covariants the variables of weight and body_height instead of "body_size_factor"? thank you
Thank you for your helpful video! I would like to know how to check the assumption of homogeneity of regression slopes for two-way ancova? Do I need to add IV1*IV2*covariate in the model?
What is the issue if the sex * body_size is significant ? It means that the homogeinity of regression does not hold but I do not really understand if it means that we cannot apply a "proper" Ancova in that situation. Thanks.
+daxatak If you violate the homogeneity of regression slopes assumption, the ANCOVA p-value will not be accurate. The question is can you violate the assumption "a little bit" and get a reasonably accurate ANCOVA p-value? I answer this question in my upcoming stats textbook. Stay tuned.
+how2stats First of all, thanks a lot for your feedback and for all your really helpfull videos! you're doing a great job! So, when the interaction effect is significant, I should be carefull when interpreting the main effects, but this could mean that the interaction effect is actually real/interesting, right? In that way, is it ok to split my covariate in subgroups in order to run an 2 way ANOVA and checking the interaction effect and post-hocs statistics ?
HELP NEEDED! Can someone please inform me of an analysis that can run a three-way ancova? Just found out today that SPSS will not run a three-way. Any suggestions?
SPSS can definitely run a three-way ANCOVA. Just put the variables in the appropriate boxes (via the GLM menu option), run the analysis, and you'll see a three-way interaction term in the 'Tests of Between-Subjects Effects' table.
@@how2stats THANK YOU SO MUCH. I am having trouble finding a video of how to do this. Do you know of any that I can access or do you have plans of ever doing one on a three way ANCOVA ? Really soon? Just asking. lol
Hi, If I want to does gender control (affect ) the relationship between language achievement and language anxiety, should I use this to do the statistics? thx
whay version of spss is this? When I do it I don't get Levene's test, I had to do it separately. Also I didn't get the eta square value, how do Iget that? (Thank you so much for these videos by the way, you are basically writing my dissertation)
Hello I am doing ANCOVA with 2 factors that have three levels. Main effects are significant but when I do pairwise comparison none of them are significant. How can this be possible? And is there a way to plot the interaction effects between factors and the covariate? Thanks.
20 minutes later: and the answer is yes! haha! Thanks for the video!
oh my. that took a long time to find out how to do ANCOVA. but at least it's detailed!
Would you happen to know if there is an alternative statistical test (e.g., a non-parametric test) that could be used in place of an ANCOVA if your data-set violates the homogeneity of regression slopes assumption?
connorism6 Quade’s test is what you need.
1) Rank the dependent variable and any covariates, using the default settings in the SPSS RANK procedure. This is done for all cases, ignoring the grouping variable.
2) Run a linear regression of the ranks of the dependent variable on the ranks of the covariates, saving the (raw or Unstandardized) residuals, again ignoring the grouping factor.
3) Run a one-way analysis of variance (ANOVA), using the residuals from the regression in the prior step as the dependent variable, and the grouping variable as the factor. The F test resulting from this ANOVA is the F statistic Quade used.
Thanks a lot. I got one more question about whether we can use Multiple linear regression analysis for this type of data
Hi. Could I do one analysis with two covariants? For example: put as covariants the variables of weight and body_height instead of "body_size_factor"? thank you
Yes, no problem.
Thank you for your helpful video! I would like to know how to check the assumption of homogeneity of regression slopes for two-way ancova? Do I need to add IV1*IV2*covariate in the model?
What is the issue if the sex * body_size is significant ? It means that the homogeinity of regression does not hold but I do not really understand if it means that we cannot apply a "proper" Ancova in that situation.
Thanks.
+daxatak If you violate the homogeneity of regression slopes assumption, the ANCOVA p-value will not be accurate. The question is can you violate the assumption "a little bit" and get a reasonably accurate ANCOVA p-value? I answer this question in my upcoming stats textbook. Stay tuned.
+how2stats First of all, thanks a lot for your feedback and for all your really helpfull videos! you're doing a great job!
So, when the interaction effect is significant, I should be carefull when interpreting the main effects, but this could mean that the interaction effect is actually real/interesting, right? In that way, is it ok to split my covariate in subgroups in order to run an 2 way ANOVA and checking the interaction effect and post-hocs statistics ?
HELP NEEDED! Can someone please inform me of an analysis that can run a three-way ancova? Just found out today that SPSS will not run a three-way. Any suggestions?
SPSS can definitely run a three-way ANCOVA. Just put the variables in the appropriate boxes (via the GLM menu option), run the analysis, and you'll see a three-way interaction term in the 'Tests of Between-Subjects Effects' table.
@@how2stats THANK YOU SO MUCH. I am having trouble finding a video of how to do this. Do you know of any that I can access or do you have plans of ever doing one on a three way ANCOVA ? Really soon? Just asking. lol
Hi, If I want to does gender control (affect ) the relationship between language achievement and language anxiety, should I use this to do the statistics? thx
whay version of spss is this? When I do it I don't get Levene's test, I had to do it separately. Also I didn't get the eta square value, how do Iget that? (Thank you so much for these videos by the way, you are basically writing my dissertation)
You'd have to select them in the "options" menu
+Mébh Conneely McInerney Go back to part 1 toward the end: 3.54
Hello I am doing ANCOVA with 2 factors that have three levels. Main effects are significant but when I do pairwise comparison none of them are significant. How can this be possible? And is there a way to plot the interaction effects between factors and the covariate? Thanks.
Do you breath statistics?
Before "yes," however, your covariate is not significant, right?
Thank you so much for saving my life! :D
Thank you very much for the video. I liked, commented, and subscribed. Will also preach your math gospel.
Thanks!
thanks for this! it really helped!
Thanks, very helpful
great
Love this, funny subject to choose... I'm facebooking this
marry me
wasting of time you talk about a lot of irrelevant things