How to Select Join Fields
Guidelines
Join specifications are used to select candidate pairs that agree exactly on all listed join fields. Pairs not selected by the join specifications do not participate in the linkage. So it is very important to know which cases you may be inadvertently excluding. Join fields should be changed for each pass to ensure that as many different candidate pairs as possible are selected for evaluation. For example, if you use Date of Birth to join, and your data include DOB only for drivers, then the candidate pairs will be limited to drivers. A separate pass will be necessary using specifications that pick up passengers.
Field from A: Any field from standardized Table A (Crash).
Field from B: Equivalent field from standardized Table B (EMS).
Criteria for A: Case selection criteria that apply to Table A. Syntax is the same as WHERE clauses in SQL statements.
Criteria for B: Case selection criteria that apply to Table B.
JOIN FIELDS SHOULD HAVE TOTAL INFORMATION ABOUT EQUAL TO THE CUTOFF WEIGHT. Much lower total information means you waste processing time evaluating many very poor candidate pairs that will be discarded. Much higher means you exclude marginal candidates that should be evaluated. It is OK if some of your passes have higher totals, just not all of them.
Run times depend on the number of candidate pairs selected. Join fields that do not sufficiently discriminate among the persons and crashes involved may cause very, very long run times.
