The most effective way in which to guard against threats to internal validity while honoring the necessity in some cases for the use of a longitudinal studies is to assess events that have occurred in the environment. The study could also be supplemented with one-time interviews with individuals who had recently lost their spouses and those who had not, matching the two groups as closely as possible in other ways.
Threats to external validity are less important in quasi-experimental than in experimental design; this said, the primary threat to experimental validity in this case is the small sample size.
Deskovitz, M., Key, D.Hill, E.M.,...
(2004). A long-term family-oriented treatment for adolescents with substance-related disorders: An outcome study. Child and adolescent social work journal 21(3): 265-284.
The above study surveyed individuals who had completed a long-term day treatment facility for substance abuse over a five-year period. It utilized an experimental technique by including all of the graduates in the subject pool. The authors examined the series of survey results to derive what factors pushed some subjects into abstinence while others relapsed. They concluded that those who felt that they had "hit bottom" were the ones who could maintain sobriety.
