I think repeated measures as we used for our mice may be suitable for you:

A one-way ANOVA shall suffice. In the event of statistical significance of the test, a post-hoc analysis could be done to compare the means.

Ette Etuk Thanks for your reply. When I will use the post host test for 4 groups at one time point, what is the best way to compare them? Example: compare 4 groups with each other ? or should I compare 3 groups with control (Control / Control + biological supplement, Model operation, Model operation+biological supplement)? or should I compare 2 pairs of groups (Control / Control + biological supplement, Model operation/ Model operation+biological supplement)

You arrange the four means in ascending order. You find the difference between the largest and the smallest. Ascertain whether it is significant. Next find the difference between the largest and the next least. Ascertain whether significant. And so on. Each time you compare a pair until the set is exhausted. Good luck.

Let me make sure I understand. The study has a 2x2 between subject design (bio sub or not, by model op or not) and 3 repeated measures (the time variable), which suggests a (2x2x(3)) mixed Anova design, but that depends on more information. You just say you have blood samples, but do not say what variables you have from these and also what other variables you have. I assume your hypotheses are whether the 2 between-subject variables affect some variables based on the blood and whether there effects are additive (i.e., is there an interaction). More information is needed about your outcome variable(s) before speculating on procedures.

Daniel Wright Daniel Thank you very much for your reply. This is a good method.

1.Example: Red blood cell level

-----------control(C)-additive(A)-model operation(MO)--modeloperation+additiveMO+A

before ----5,6------------- 7,2 ------------------ 6,4 -------------------------------- 6,1

10 day ----6,1------------6,4 ------------------- 5,1 ----------------------------------6,3

30 day ---7,1------------ 7,3--------------------7,4 ---------------------------------7,0

2.When I do a post host test in this method I got comparisons of 12 groups to each other. Can I present these results in an article in this form?

Before---------------------------------------10day---------------------------30 day

С------A-----MO --MO+A---- С------A---MO -MO+A- ---С--------A-----MO --MO+A

5,6a-7,2b -6,4c - 6,1d-------6,1e-6,4f-5,1b -6,3df------7,1abf- 7,3a--7,4bf -7,0ab

3. What should I call this method in my article?

Béatrice Marianne Ewalds-Kvist Thank you very much for the article. But I do not quite understand how many factors you took for the linear model of two factor analysis. Did you consider line + sex or line + sex + generation?

I was interested in the chart type "thigmotactic ratio". What is its name in statistical programs?.

Q1. Сan i use one-way anova to compare groups at one time point? or should i use two-way anova in this causes? NO, not possible, you can compare the groups overtime not one point in time

2. Сan i use two-way anova to compare groups at three time point between themselves? or should i use MANOVA in this causes ? NO, Two ANOVA is for two way comparison (two separate F-Values comparing columns and rows, so it depends on what you have on the columns and rows, in your own cas is groups and time point

3. Could I use Post hoc analysis it to compare 4 groups with each other at one time point ? NO

NOTE: A two way ANOVA will simultaneously test;

(1) Significant difference in drug effects among the groups ( control, additive , model operation, model operation+additive) and

(2) Significant difference in drug effects given time points (before surgery, 10 days after surgery, 30 days after surgery) if drug effect is your series

The Post Hoc Test particularly the LSD or Turkey is for multiple comparison

so the result will compare;

Control against additive , model operation, model operation+additive

Additive against Control , model operation, model operation+additive

Model operation against Control, additive , model operation+additive

Model operation+additive against Control, additive , model operation

See attached sample result

Godwin Lebari Tuaneh Thanks for your work. I have a few more questions.

1. When I do a post host test (Turkey) with factors: difference between groups * the difference between days. I got comparisons of 12 groups to each other. Can I present these results in an article in this form?

Before---------------------------------------10day---------------------------30 day

С------A-----MO --MO+A---- С------A---MO -MO+A- ---С--------A-----MO --MO+A

5,6a-7,2b -6,4c - 6,1d-------6,1e-6,4f-5,1b -6,3df------7,1abf- 7,3a--7,4bf -7,0ab

2. Is there a difference in the two way Anova and mixed Anova design?

Yes: a fixed (2way) ANOVA fits a separate intercept for each individual. This costs as many degrees of freedom as there are individuals in your study. A mixed model can assume the intercepts of the individuals to be a "random" sample of a normal distributed variable, of which only the mean and the variance is estimated, what saves you a lot of degrees of freedom.

Good read: http://www.stat.columbia.edu/~gelman/arm/

Jochen Wilhelm Thanks. When I do a post host test (Turkey). Can I present these results in an article in this form?

Before post host ( Turkey ).

Example: Red blood cell level

------control(C)-additive(A)-model operation(MO)--modeloperation+additive(MO+A)

before ----5,6------------- 7,2 ------------------ 6,4 -------------------------------- 6,1

10 day ----6,1------------6,4 ------------------- 5,1 ----------------------------------6,3

30 day ---7,1------------ 7,3--------------------7,4 ---------------------------------7,0

After post host (Turkey).

Before---------------------------------------10day----------------------------------30 day

С------A-----MO --MO+A---- С------A---MO -MO+A- -------------С--------A-----MO --MO+A

5,6a-7,2b -6,4c-6,1d-------6,1e-6,4f-5,1b -6,3df-------------7,1abf- 7,3a--7,4bf -7,0ab

I don't know if all these test really make sense. However, you can show the results in a Tukey confidece interval plot.

You can use a 2-way ANOVA since you have 2 independent variables; the time of sampling and the treatment modality.

Yes if it is better tabulated

The Post Hoc T(urkey) could also be resented as;

Group 1 Group 2 Group 3 Group 4

B4 10d 30d B4 10d 30d B4 10d 30d B4 10d 30d

in this case, your result explanation might focus on variation in time points in each group before the groups. in your presentation the result explanation will focus on variation in groups in each time point before the time points

on you 2nd question, I agree with Jochen Wilhelm .

A thigmotactic ratio is something you calculate by hand. Thigmotaxis means clinging to the wall. Wall-seeking behavior is an expression of emotionality in the mouse.

Godwin Lebari Tuaneh Thank. When I build linear model (minitab), I indicate the interaction of 2 factors (groups * time) is it a two-way Anova or Mixed model Anova?

When i do post host test, the statistical program compares all 12 groups (4 groups * 3 time points) with each other, so we have 66 comparisons. Right?

Jochen Wilhelm Thank. Can you recommend which test will make sense? As I understand it, I cannot describe in the article a comparison between 4 groups at one time point (12 comparisons) 3 times separately, because I have to compare these four groups at three time points at once (66 comparisons).