Self-generated identification codes (SGICs) are strings of information based on stable participant characteristics. They are often used in longitudinal research to match data between time points while protecting participant anonymity. However, the use of SGICs with military personnel has been infrequent, even though military personnel do not have the same privacy protections as civilians. The current paper reports results from two studies that tested the feasibility, reliability, and validity of using a SGIC to collect sensitive longitudinal data among military personnel. In study one, a team of 105 participants was tracked three times over a period of 12 weeks. The 10-item SGIC produced optimal matching over the 12-weeks. In study two, 1,844 participants were randomly assigned to a SGIC group or an anonymous control group, and then were asked to provide information about their alcohol use. Although match rates declined over time, there were no observed differences between study groups in participants' beliefs about the use of a SGIC. However, differences were identified in reported alcohol use behaviors between the groups, with controls reporting significantly more drinks per week and higher AUDIT-10 scores. While these findings raise potential concerns about using SGICs for epidemiological assessments of highly sensitive problem behaviors, these codes may still be useful in determining group differences in behavior change in randomized studies.
Keywords: alcohol use; identification codes; longitudinal research; measurement; military.