Its clear there is a warning that said SPSS cant permorf Bonferroni Pos Hoc in Repeated Measure for the split group because there is less than 3 groups being tested. When you running Repeated Measure with Bonferroni you put 'treatment' for being tested with Bonferroni not the 'time'. But why when you show the output there is a Bonferroni Pos Hoc table for Repeated Measure? Its so confusing. Please answer this question
Which one is the output table for pos hoc bonferroni test repeated measure ANOVA on SPSS? because at the end you put the screenshot table but I dont know which table it is on SPSS. It said BONFERRONI at the title of the table from your screenshot but I cant find the same table like you
Hi, would it be possible to add predictors of succes of treatment to this model? say in this example analysing whether the severity of depression at baseline influences the effect of group (treatment) or maybe age at baseline influences the effect of treatment, or comorbid symptoms of anxiety etc. ? If not, what model would you recommend?
My project involves taking activity measures before and 6 weeks and 6 months after a knee surgery. I will also be comparing measures by participants' characteristics. What statistical approach is appropriate for comparing the effects of the surgery on activity? Thanks
Hi Emmanuel, I would suggest you use a 3 time (baseline, 6 weeks and 6 months) repeated measures Analysis of Variance (ANOVA). The activity measure will be your dependent variable. If you find a significant effect of knee surgery on activity measures, then run further post-hoc tests to identify where the significance lies (for example it could be baseline and 6 weeks, or baseline and 6 months, or 6 weeks and 6 months). Importantly, check your data is normally distributed, as this is an assumption of an ANOVA. Good luck!
Thank you Dr Sando. I was also thinking if Mixed Model could be an option? Some participants may not have complete data from baseline to 6 months follow up. Thank you doc
Your welcome! A repeated measures ANOVA should be fine as you only have one group. A mixed model would only be necessary if you had another group (e.g. control group). Then you would be analysing the group by time effect for the dependent variable. Don't worry about loss of data at 6 month follow-up. As long as you still have reasonably sufficient participants you can still run the repeated measures ANOVA, but just make clear that the n will be lower (as compared to 6 weeks). I hope this makes sense.
Thank you Dr Sando, I am very grateful for your assistance and patience!
Thank you so much, you literally saved my 3rd year project!
Your very welcome Shantrece! I am glad the information helped and good luck with the rest of your studies.
Its clear there is a warning that said SPSS cant permorf Bonferroni Pos Hoc in Repeated Measure for the split group because there is less than 3 groups being tested. When you running Repeated Measure with Bonferroni you put 'treatment' for being tested with Bonferroni not the 'time'. But why when you show the output there is a Bonferroni Pos Hoc table for Repeated Measure? Its so confusing. Please answer this question
Which one is the output table for pos hoc bonferroni test repeated measure ANOVA on SPSS? because at the end you put the screenshot table but I dont know which table it is on SPSS. It said BONFERRONI at the title of the table from your screenshot but I cant find the same table like you
Hi, would it be possible to add predictors of succes of treatment to this model? say in this example analysing whether the severity of depression at baseline influences the effect of group (treatment) or maybe age at baseline influences the effect of treatment, or comorbid symptoms of anxiety etc. ?
If not, what model would you recommend?
Great one!
Thanks Emmanuel!
thank you very helpful, is there a similar way (mixed ANOVA) to perform analysis for non parametrics?
Hi, Non-parametric alternative for mixed ANOVA: None.
Resource: Pallant, J. (2016). SPSS survival manual. McGraw-Hill Education (UK).
My project involves taking activity measures before and 6 weeks and 6 months after a knee surgery. I will also be comparing measures by participants' characteristics. What statistical approach is appropriate for comparing the effects of the surgery on activity? Thanks
Hi Emmanuel, I would suggest you use a 3 time (baseline, 6 weeks and 6 months) repeated measures Analysis of Variance (ANOVA). The activity measure will be your dependent variable. If you find a significant effect of knee surgery on activity measures, then run further post-hoc tests to identify where the significance lies (for example it could be baseline and 6 weeks, or baseline and 6 months, or 6 weeks and 6 months). Importantly, check your data is normally distributed, as this is an assumption of an ANOVA. Good luck!
Thank you Dr Sando. I was also thinking if Mixed Model could be an option? Some participants may not have complete data from baseline to 6 months follow up. Thank you doc
Your welcome! A repeated measures ANOVA should be fine as you only have one group. A mixed model would only be necessary if you had another group (e.g. control group). Then you would be analysing the group by time effect for the dependent variable. Don't worry about loss of data at 6 month follow-up. As long as you still have reasonably sufficient participants you can still run the repeated measures ANOVA, but just make clear that the n will be lower (as compared to 6 weeks). I hope this makes sense.