Pre and Post with Treatment group - ANOVA , ANCOVA, Regress differences???

[Marvin85]

Hello friends,

My main task is to "test" if staff member who participated in the counseling sessions had fewer "Use of Force (UOF)" incidents in the 6 moths after the counseling session. I have a data set with the number of UOF incidents for the staff member in the 6 months before (pre) and after the counseling (post). I also have a control group (staff who didn't participate in the counseling sessions). So in total I have the pre UOF count the post UOF count and a group variable.

1. Which test should I used?

By conducting Paired sample tttests and sign rank tests for each group separately, I concluded that UOF incidents significantly decreased in the two groups from pre to post.

Now I would like to know if the staff who received the counseling (experimental group) reduced the number of UOF in a greater rate that staff in the control group. I just want to see if the counseling is more effective in reducing UOF incidents than no counseling at all.

2. Should I use ANOVA, ANCOVA or any other technique?

To try to do this, I created a variable that is the difference of post and pre UOF, and conduct a linear regression with the Group variable (experimental vs control) as the independent variable and the pre scores as an independent variable as well- more as a control. Is this valid? Any suggestions?

Thank you so much!
Marvin

 

[Karabiner]

Some information on the study and the data would be useful.
Was allocation to groups randomized?
Was the baseline level of both group equal?
How large were the groups?
Which median and range does the dependent
variable have at baseline and at follow-up?

With kind regards

K.

 

[Marvin85]

Hi Karabiner,

Thank you a lot for your reply.

1. No, the groups were not randomized.Basically, the criteria is that if a staff member had 3 UOF incidents in a year, they had to attend the counseling session. For the control group, I just chose staff that had 3 UOF but didn't received the training. I have compared the demographic characteristics of the two groups and they are very similar. Same mean age, race, years in the job, etc. The only big and very important difference is that the control group have fewer UOF in the pre and post times (UOF is my dependent variable). I guess this happened because I didn't assign staff to the groups randomly.
2. No, as mentioned in 1. control group have ion average fewer UOF than experimental.
3.345 each group.
4. UOF pre = 2.8 post 1.7 for both group. As I said this vari by group.

I was doing some reading and I guess Mixed or repeated measure ANOVA or ANCOVA, can work here???

 

[Karabiner]

The mixed ANOVA with 1 within-subject factor and 1 grouping factor could perhaps be considered. The problem is that you don't have an interval scaled variable. Instead you have a count variable with small counts. I don't know whether an ANOVA or a linear regression approach is valid in that case. Sample size is ok and could perhaps alleviate problems with assumptions.

In addition, interpretation has to be cautious because groups are compared who seemingly showed different behaviour before inclusion and therefore have different baseline values.

With kind regards

K.

 

[Marvin85]

Thank you! What is the differences between Mixed ANOVA and Factorial ANOVA? Do you use Stata? Is so, What would the appropriate syntax for a Mixed ANOVA?? I would appreciate any help.