Figure 1
The figure shows a multi-stage flowchart that begins with a text box titled “Death Certificate Records from Vital Statistics (n equals 2,371,848)”, at the top, and contains the list “First and Last Name”, “Last 4 Digits of S S N”, “D O B and D O D”, “Manner of Death”, and “Cause of Death (I C D codes)”. From “Death Certificate Records from Vital Statistics (n equals 2,371,848)”, a line extends downward and connects to a Venn diagram. This line is connected to a box labeled “Filtered by suicide as manner of death”, on the left. The Venn diagram has two circles overlapping each other at the centre. The left circle is labeled “Vital Statistics Suicide Cases (n equals 36,970)”. The right circle of the Venn diagram is labeled “F D L E’s A T M S Database”. Where the two circles overlap, the central overlap region contains the heading “Joined on: First Name”, “Last Name”, “D O B”, and “Last 4 S S N”, followed by a small chain-link icon. From this Venn diagram, a line extends downward and connects to a box labeled “Profiles Identified in A T M S Database (n equals 773)”. From this box, two lines extend downward and connect to two boxes arranged horizontally. The left box is labeled “A T M S Profiles Verified with Officer Certification via Manual Examination (n equals 440)”. The right box is labeled “A T M S Profiles Verified with Officer Certification via Computational Text Mining (n equals 444)”. From these two boxes, a line extends downward and connects to a box labeled “4 Unique Profiles from the Computational Method Manually Examined and Verified”. Another vertical line continues downward to a final box labeled “A T M S Profiles Verified with Officer Certification via Manual and Computational Methods (n equals 444)”. To the right of this figure, a vertical column of process labels appears, reading from top to bottom as follows: “Raw Data Source”, “Data Reduction”, “Data Integration (records joined on select features)”, “Preprocessed Data”, “Data Mining and Quality Control”, and “Processed Data”. A down-pointing dashed arrow runs vertically down each phrase and links it to the next.Schematic for suicide case selection. ATMS = Automated Training Management System (database of certified sworn officers in Florida); DOB = Date of Birth; DOD = Date of Death; FDLE = Florida Department of Law Enforcement; ICD = International Classification of Diseases; SSN = Social Security Number. Source: Authors' own work