After reviewing Initiative Impact details, click the Matching Details tab to confirm that students were successfully matched for analysis.
Recall that participant students are matched with comparison students on two metrics:
- Persistence prediction: The likelihood of the student to persist to the next term and stay enrolled past the add/drop date.
- Propensity score: The similarity of the comparison student to students who participated in the initiative.
The charts on the Matching Details tab show the distributions of these scores for participants and the comparison group before and after matching. You should see the matching details for Persistence Prediction by default.
Before Matching chart shows the scores for all submitted participants and the eligible comparison group. Before matching, the distribution of scores may be very different. You may see some selection bias, such as a participant group with much higher persistence predictions for an opt-in initiative such as tutoring. It follows that students who chose to attend voluntary tutoring would be more likely to persist at your institution. The matching process takes care of this bias before evaluating the impact of the initiative.
The Similarity indicated on the chart shows how close the scores of the participants and the comparison group are. The closer this measure is to 100%, the more similar the scores of the student groups used for analysis. Before matching, this measure will include natural separation between the scores. After matching, the score distributions should be much more similar.
The After Matching chart shows the scores for the selected participants and the matched students from the comparison group.
Click the Propensity Score tab to switch to the distributions for those scores.