Statistical Analysis of Student Test Scores before and after an Intervention


Please help me ensure that I am conducting the proper statistical analysis. I am trying to determine if my student test score results are statistically significant after having applied a learning intervention.

The data I am using are student test scores on TEST 1 from my 3 previous classes vs. student test scores on TEST 1 (the exact same test) from my current class which I taught using a different method.

I entered the scores for each class in separate columns on Excel. I asked Excel to run a T-test. For Array 1 I chose the 3 columns of student scores from the pre-intervention classes. For Array 2 I chose the column of test scores from my current class that had the learning intervention. I then chose 2 tails and type 2. Did I set this T-Test up correctly? Can I use 3 columns of data for Array 1? Please also note that I did not have the exact same number of students per class. Below is an example of my spreadsheet (these are not the actual scores just a representation).

Column 1 Column 2 Column 3 Column 4
Class 2013 Class 2014 Class 2015 Class 2016(with intervention)
86 87 85 92
88 92 88 92
91 82 90 90
84 87 83 88
90 86 87 93
86 85 89 87
86 87 85 92
88 92 88 92
91 82 90 90
84 87 83 88
90 87 93

Thank you for the help
the simplest would be to run a non-parametric test like Kruskal-Wallis.

Thank you for the advice, but I'm not familiar with the Kruskal-Wallis. Can you explain? It that a test I can run on Excel?
its a non-parametric anova, and yes, you can run it on excel (ANOVA at least, not sure about the KW test).

anova assumes normality, but a KW test does not. Usually ANOVA is fairly robust against normality but when sample sizes are relatively small-ish (in your case 10-12 per group) its beneficial to run a non-parametric if normality is not there.