Title page
Contents
Why this guide? 4
Overview of the Nebraska-based approach for identifying similar schools 6
The Nebraska-based approach for identifying similar schools 10
Identify a set of variables that meets key considerations 11
Choose a suitable distance measure and calculate the distance between every pair of schools 16
Match schools to each target school 22
Evaluate the quality of matches produced by a matching option 25
Actions that could be informed by identifying similar schools 31
Appendix A. Other example approaches for identifying similar schools 32
Appendix B. The Euclidean and Mahalanobis distance methods 35
References 42
Acknowledgments 43
Table 1. Variables used in the Nebraska-based approach for identifying similar schools, by variables category 14
Table 2. Example for three schools and four variables 18
Table 3. The 10 most similar schools for each of three target schools, by Euclidean distance 24
Table 4. Example of summary results for evaluating different matching options 29
Figure 1. Example of plotting the distance between four schools based on two variables 8
Figure 2. Steps and decision points in the Nebraska-based approach for identifying similar schools 10
Figure 3. Example of plotting data points for multiple schools showing correlation between two variables 19
Figure 4. Example of Euclidean distance measure with correlated variables and with school B as the target school 20
Figure 5. Example of Mahalanobis distance measure with correlated variables and with school B as the target school 21
Figure 6. Example of plotting 23 schools using two variables 23
Boxes
Box 1. Key terms 7
Box 2. What rationales have other education agencies used for selecting variable categories? 13
Box 3. Steps in evaluating the quality of matches 25
Table B.1. Example measures for three variables from 23 schools 37
Table B.2. Similar schools to target school A using different methods for calculating distance, by distance 41
Figure B.1. Values used to create variance-covariance matrix 38