Although the present findings can inform military policy and practice, it is important to consider these findings in light of some limitations.
Specifically, as is the case with many survey studies, the present findings may reflect self-report bias. Perceptions and reports of military personnel are important and typical sources of data for informing military policy and practice. But studies that assess objective, observable indicators of cohesion and readiness and the actual presence of gay and lesbian service members would be useful. Such research would require identifying and linking lesbian and gay service members with the observed units, but DADT is a challenge to such research.
An additional limitation is that the present data are cross-sectional. Thus, interpretations about direction of causality among the variables of interest cannot be made. The current policy precludes gathering of accurate identifying information about gay and lesbian service members or those who have served with them. Thus, tracking participants over time to collect longitudinal data that allow examination of prospective links among the variables of interest is not possible.
To address the limitations of the present study, efforts within the military to gather systematic data from randomly drawn samples about the presence of lesbian and gay personnel and their impact on objective indicators of unit cohesion, readiness, morale, and effectiveness would clearly be useful. Empirical data are critical for informing military policy and practice, and the present study represents a step in addressing the paucity of data addressing the rationale underlying DADT.
http://www.palmcenter.org/files/active/0/randstudy(3).pdf