The development and testing of an explanatory model of recovery from disaster loss are described. The sample consisted of 155 persons at Time1 – 1981, 101 persons at Time2–1983, and represented five magnitudes of disaster loss: bereaved of disaster victims presumed dead; bereaved of disaster victims confirmed dead; persons whose permanent homes were destroyed; persons whose recreational property was destroyed; and a no disaster–loss comparison group. Data tested in the model were collected 1 and 3 years following the volcanic eruption of Mt. St. Helens in 1980 on eight variables: magnitude of loss, mass media coverage, negative rating of change, change in income, self-efficacy, social support, satisfaction with financial settlement, and mental distress. Path analysis was selected to estimate the magnitude of hypothesized direct and indirect linkages between variables presumed as causes of variables treated as effects. Results showed an improvement of goodness of fit by testing the model with the 1981 data, respecifying the model, and testing it again with the 1983 data. The greatest explanatory effects on mental distress 3 years postdisaster were mental distress reported 1 year postdisaster and negative ratings of change associated with disaster loss.