Each group's participants have taken same assessment and have pre test and post test scores. The participants did a pre test assessment and post test assessment before and after a service delivery. This was done in year 1 and year 2. I want to see which group did better than other as in showing improved performance between year 1 and 2. Which test should I use?
hi,
Paired t test for each of the groups and compare the changes.
pretest minus post-test divides by pre-test score . Can be converted to percentages
Use ANOVA and do post hoc analysis by using multiple comparison. For multiple comparison, define contrasts as per your interest and use t-test.
Hey Mansi Agnihotri
as always, we would need more information on your research design in order to provide better feedback. 11 groups is a lot which is why you should not use multiple t-tests as suggested by Bhogaraju Anand . This will inevitably lead to what is known as alpha error accumulation. ANOVA with post-hoc tests will serve the purpose of your study much more adequately as Nirajan Bam has pointed out.
Best
Marcel
I thought of doing same but did not find any references@ where t-test value have been compared between groups. also I forgot to add that the groups did assessment 2 times once in year 2020 other in 2021. I need to see which group better performed as in getting highest positive jump from pre test to post test score.
Marcel Grieger and Nirajan Bam . Thank you . I forgot to add that there are time lines in my question and group participants took tests once in year 1 and other in year 2. I want to compare the groups on basis of max improvement shown from
year 1 to year 2. Will anova with post hoc t test be applicable here?
In my opinion, when conducting research, you should have used repeated measurements to better understand the process of weakening or strengthening scores. But this has not been done. My suggestion for you is to use analysis of covariance (ANCOVA) so that by removing the pre-test effect, you can compare the post-test scores between eleven groups.
Now your data situation looks like repeated measures. Thus, you can use repeated measure ANOVA. Groups are your between subject and time is within subject. In addition, put interaction effect in the model so that you can be able to examine effect over time across the groups. If you wish instead of repeated measure ANOVA, you can use linear random slope model.
Hi,
For ANOVA It is an Independent variable with different levels but here there are several groups that are not levels of an IV.
So, Paired t-test is applicable.
As i wrote previously you can see the rate of change and convert to a percentage.
You will get percentages for each of the groups.
Place them in order and draw conclusions.
Ex:
C>D>a>H,,,,,,
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size