Details of probabilistic linkage process in LinkSolv V.9.0
Linkage specifications | ||
Construct | Set at: | |
False positive rate | 0.01 | |
Cut-off probability | 0.60 | |
Join criteria | Pass 1 data elements | Pass 2 data elements |
Date of birth | Age | |
Initial letter of first name | Initial letter of first name | |
Sex | Soundex of last name* | |
Sex | ||
Match criteria | ||
Data element | Comparison method | |
First name, soundex | InList† | |
Last name, soundex | InList | |
Middle initial | InList | |
Month and day of birth | Exact | |
Residential street name, soundex | InList | |
Residential ZIP code | InList | |
Sex | Exact | |
Social Security Number (SSN) | Typos, 1‡ | |
Driver licence number | InList | |
Age at death | Exact | |
Event date | InList | |
Event municipality | InList | |
Race | Exact |
*Soundex is a phonetic algorithm for indexing names by sound, as pronounced in English.26
†The InList comparison method (LinkSolv V.9.0 (2015 Strategic Matching)) allows a single value in one record to be compared with multiple, concatenated values in another record. For example, crash-involved driver records had a single driver licence number listed, while the licensing records contained multiple driver licence numbers per person. With this comparison method, the single driver licence number in the crash-involved driver record was compared against each driver licence number in the licensing record.
‡The typos comparison method (LinkSolv V.9.0) allows up to X differences between two values to be considered a match; for example, with X=1 typo specified, the SSN values of 123 456 789 and 723 456 789 would be considered a match.