# One Way ANOVA/Post Hoc with Time as Independent Variable

#### thedeclineofgenius

##### New Member
I've been working on replicating/updating an old analysis on student test scores that doesn't have a very helpful methodology section. I've spent a number of days searching online for a similar situation so I can see if what I'm doing makes sense with no luck. I would greatly appreciate any insights or thoughts!

Basically, I'm using ANOVA to try to determine whether changes in 3rd grade reading scores in one district are significant over a ten year period. I'm using excel and XLSTAT for this analysis, but can switch to STATA if needed.

Since each third grade cohort is comprised of new students, it didn't seem like I could use a repeated measures ANOVA. I'm running a one-way ANOVA with weighted test scores as the dependent variable and 'year' as the independent variable. I have school-level data for test scores for the 3rd grade class at each school in the district, broken up by different subgroups of students at the school level.

I added observation weights for each subgroup/school based on the number of test takers to account for the different sizes of subgroups at each school. I chose a post-hoc test (initially a regwq because of the number of years(groups)) to see how many and which years had a significantly different mean score.

My Questions are:

Does running a one-way ANOVA with time as the independent variable to determine whether change is significant make sense statistically?

Am I right to think that a repeated measures ANOVA would not be the correct choice in this situation?

Are observation weights the right way to account for different frequencies in each school?

Is REGWQ a good choice to test multiple comparisons in this case?

How would I interpret the multiple comparison output? Would all pairwise comparisons need to be significant in order to conclude that changes in test scores have been significant over 10 years? if not, how many or which pairwise comparisons would need to be significant to indicate a larger trend?

I attached a sample ANOVA output (XL spreadsheet) if that helps. Any thoughts or feedback would be greatly appreciated! I want to make sure I'm approaching this correctly before I move forward. Cheers!