Broken boundaries or broken marriages? Racial intermarriage and divorce in the United States.

OBJECTIVE: Several recent studies have investigated the consequences of racial intermarriage for marital stability. None of these studies properly control for first-order racial differences in divorce risk, therefore failing to appropriately identify the effect of intermarriage. Our article builds on an earlier generation of studies to develop a model that appropriately identifies the consequences of crossing racial boundaries in matrimony. METHODS: We analyze the 1995 and 2002 National Survey of Family Growth using a parametric event-history model called a sickle model. To appropriately identify the effect of interracial marriage we use the interaction of wife's race and husband's race. RESULTS: We find elevated divorce rates for Latino/white intermarriages but not for black/white intermarriages. Seventy-two percent of endogamous Latino marriages remain intact at 15 years, but only 58 percent of Latino husband/white wife and 64 percent of white husband/Latina wife marriages are still intact. CONCLUSIONS: We have identified an important deficiency in previous studies and provide a straightforward resolution. Although higher rates of Latino/white intermarriage may indicate more porous group boundaries, the greater instability of these marriages suggests that these boundaries remain resilient.

Age, expressed emotion, and interpersonal control patterning in families of persons with schizophrenia.

Expressed emotion (EE) is a measure of a relative's attitude toward a person with a mental disorder as reflected by comments made to an interviewer. Over the years, an impressive body of research has been generated in attempts to explain the relationship of EE and course of illness, particularly in regard to schizophrenia. Past analyses have demonstrated two common patterns of interpersonal control in families with high-EE relatives. The first was that the relative and patient competed for 'who's in charge', with both rigidly asserting control. The other was that the relatives responded rigidly in one control mode no matter what the offspring was saying. In this study, we tested whether the difference in patterning was due to the patients being older. Study participants were 71 relative-patient dyads. EE was measured with the Five-Minute Speech Sample, the Family Problem-Solving Task was used to generate interaction, and interaction data were coded with the Relational Control Coding System. Multilevel modeling was used to estimate a model with verbal exchanges as Level 1, relatives' EE as Level 2, and age of the patient as Level 3. Results showed that age attenuated the relationship between the patient's message and the parent's response, and this effect was stronger in families with high-EE relatives. This effect may reflect disengagement on the part of the relative, the patient, or both. The over-responsive families may need advice on how to create more distance, or how to be more objective, but the distant families may need more support and encouragement to communicate and problem-solve.