my suggestion: now compute the 'additive' model (the one with just the two main effects) and then evaluete both models with AIC. If the AIC of the additive model is lower then it's all cake and you can evalulate the additive model.
R can calculate, in pos hoc tests, the difference in CHANGES between groups (as an ANOVA of deltas)? (Instead of multiple comparsions isolated across different groups and/or time points)?
Hey thanks for this video, I'm wondering about seeing "group" in the error term. Isn't group in your data set a between-subject factor, one subject is either in group "male" or in group "female". How should the model formula look like for this case? aov(resp.var ~ group*time + Error(subID/time + group)) or better without between-subject factors aov(resp.var ~ group*time + Error(subID/time)) ?
Thankyou for this informative video. I have a question here. I am working on data of pest study in crops and was wondering if I can apply two-way repeated measure analysis. I have three treatments (control, insecticide 1, insecticide 2) and I am measuring three response variables (LAI, abundance, defoliation) at 5 days interval for a total of 6 weeks. I am selecting 3 plants randomly per plot to record the response variables. So every time point (5 days interval), I select three random plants. Would repeated Two-way repeated measure analysis apply to that data set?
Sorry to report that, but this tutorial is incorrect. Subject is taken as integer and will be treated as a quantitative variable, where it should be a qualitative one. If you change these integers by "subject1", "subject2"... (which should be used) you will get different results.
Thanks for your criticism. The model used in this example as aov() function which indicates response variable is separated by group and time and not the subID. It depends on the hypothesis or objectives of the research. The results may be wrong if subID is a factor and I am using it as integer as shown by str() function. However one can change it to factor by using as.factor() command but again it depends on objectives of the research whether to take this categorical variable as factor.
@@AGRONInfoTech Hello, Great Video Agron! Was very helpful. Would you be able to provide any examples or reasons one would choose subID to be a factor variable over an integer? I am currently looking at plants droughted and non droughted over multiple time periods.
Your data contains replication. I have data without replication i.e. Augmented block design. In this design,, I have check cultivars that are repeated. Please make a video on how can I calculate two way ANOVA for augmented block design. Hoping to hear from you soon.
@@AGRONInfoTech Thanks for your efforts. Please also calculate genotype by environment interaction GxE for augmented block as well. It will be very helpful. Please consider it on top priority.👍😊🤝.
Imma be bold and ask you, since you seem to know quite a lot about R. I hope you don't mind. I am trying to make a stacked plot of the next data: 6 different tests on x-axis, average behaviour durations on y-axis. Every test has 2-4 behavioural observations, i have 5 subjects. So i can get at most 20 different behavioural durations of one behaviour. Do i calculate the average duration for every behaviour and put that in a table and make a stacked plot of that, or is it possible to make a stacked plot with average durations (so not the sum of the durations of behaviours) with just the 'raw' data (the 20 different behavioural durations of one behaviour)? I hope you understand and have a few spare minutes to help this R-Dummy!
two way anova
how to get the standard error table?
Great video. I just cannot undestood how i can to interpret the interections of group:time with the output kruskal-wallis (final video)... thank u
If my group and time dependent variable show significant p values, however, my interaction is not... how do you recommend proceeding from there?
my suggestion: now compute the 'additive' model (the one with just the two main effects) and then evaluete both models with AIC. If the AIC of the additive model is lower then it's all cake and you can evalulate the additive model.
R can calculate, in pos hoc tests, the difference in CHANGES between groups (as an ANOVA of deltas)? (Instead of multiple comparsions isolated across different groups and/or time points)?
Hey thanks for this video, I'm wondering about seeing "group" in the error term. Isn't group in your data set a between-subject factor, one subject is either in group "male" or in group "female". How should the model formula look like for this case? aov(resp.var ~ group*time + Error(subID/time + group)) or better without between-subject factors aov(resp.var ~ group*time + Error(subID/time)) ?
Thankyou for this informative video. I have a question here. I am working on data of pest study in crops and was wondering if I can apply two-way repeated measure analysis. I have three treatments (control, insecticide 1, insecticide 2) and I am measuring three response variables (LAI, abundance, defoliation) at 5 days interval for a total of 6 weeks. I am selecting 3 plants randomly per plot to record the response variables. So every time point (5 days interval), I select three random plants. Would repeated Two-way repeated measure analysis apply to that data set?
Yes you can apply repeated measures analysis of variance.
Is it possible to add a covariate to this model? I.e. Some initial condition associated with each subject? If so, what would the syntax be? Thank you.
I am not sure about it. However I shall try to find some way and will let you know.
Sorry to report that, but this tutorial is incorrect.
Subject is taken as integer and will be treated as a quantitative variable, where it should be a qualitative one.
If you change these integers by "subject1", "subject2"... (which should be used) you will get different results.
Thanks for your criticism. The model used in this example as aov() function which indicates response variable is separated by group and time and not the subID. It depends on the hypothesis or objectives of the research. The results may be wrong if subID is a factor and I am using it as integer as shown by str() function. However one can change it to factor by using as.factor() command but again it depends on objectives of the research whether to take this categorical variable as factor.
@@AGRONInfoTech Absolutely. Thanks!
@@AGRONInfoTech Hello,
Great Video Agron! Was very helpful.
Would you be able to provide any examples or reasons one would choose subID to be a factor variable over an integer? I am currently looking at plants droughted and non droughted over multiple time periods.
Your data contains replication. I have data without replication i.e. Augmented block design. In this design,, I have check cultivars that are repeated. Please make a video on how can I calculate two way ANOVA for augmented block design. Hoping to hear from you soon.
OK thanks for your suggestion. I shall make a video on augmented block design soon.
@@AGRONInfoTech Thanks for your efforts. Please also calculate genotype by environment interaction GxE for augmented block as well. It will be very helpful. Please consider it on top priority.👍😊🤝.
Imma be bold and ask you, since you seem to know quite a lot about R. I hope you don't mind. I am trying to make a stacked plot of the next data:
6 different tests on x-axis, average behaviour durations on y-axis.
Every test has 2-4 behavioural observations, i have 5 subjects. So i can get at most 20 different behavioural durations of one behaviour. Do i calculate the average duration for every behaviour and put that in a table and make a stacked plot of that, or is it possible to make a stacked plot with average durations (so not the sum of the durations of behaviours) with just the 'raw' data (the 20 different behavioural durations of one behaviour)?
I hope you understand and have a few spare minutes to help this R-Dummy!