Marital Quality

Marital happiness and satisfaction are often measured by single, straightforward questions that ask respondents to rate their marriages on a scale of happiness or satisfaction. There may be up to ten points on the scale, but often there are only three or four. It should be noted that individual satisfaction might be composed of different dimensions in different cultures. For example, a culture where arranged marriages and strong residential extended kinship is practiced might have satisfaction with in-laws as a significant dimension of marital quality and this dimension might not show up in cultures that practice voluntaristic mate selection and neolocal residence.

The prevailing view in family social science is that single-item indicators of marital quality are unsophisticated, and they are shunned by many researchers in favor of scales and multiple-item indices. Nevertheless, the best evidence on trends in and correlates of marital quality are based on responses to one, two, or three questions, since more complex measures have very rarely been used with large and representative samples. Furthermore, several critics have argued for the use of global measures of marital quality rather than multidimensional scales and indices that include variables that may influence or be influenced by spouses' evaluations of their marriages (e.g., Fincham and Bradbury 1987; Huston, McHale, and Crouter 1986; Huston and Robins 1982; Norton 1983). One may go beyond these critics and argue that one-item measures are sometimes superior to even two- and three-item scales and indices that might be multi-dimensional rather than unidimensional.

If a question has high face validity, as the straightforward questions about marital happiness and satisfaction do, then any other questions will have lower face validity and perforce must deal with something other than simply happiness or satisfaction. The purpose of multiple-item scales is to measure "latent" variables for which no direct measurement is possible and for which several indirect measures produce a higher degree of validity than a single one can. The usual assumption seems to be that there can be no simple, direct, and straightforward measure of feelings or other psychological characteristics, although single-item measures of date of birth, gender, and various demographic characteristics are routinely used. However, the correctness of this assumption is not self-evident, and the preference for multiple-item indicators for all psychological characteristics may grow primarily out of the researchers' need to feel sophisticated.

Nevertheless, most questions used to gauge marital happiness and satisfaction provide only crude measurement, if only because they offer only a few response alternatives, and the distribution of responses is usually highly skewed. For instance, the question about marital happiness most often used on national surveys in the United States offers only three degrees of happiness—"very happy," "pretty happy," and "not too happy"—and up to two-thirds of the respondents select the highest degree. Much of the variance in marital happiness must be among those who select the "very happy" alternative, but the measure is not finely calibrated enough to capture that variance. Furthermore, there is probably a systematic over-reporting of marital happiness and satisfaction, due not only to social desirability considerations—the most commonly discussed source of response bias—but also due to denial and a stoic tendency to put up a happy front. The extent of any such bias is unknown, and perhaps unknowable, but the likelihood that it is substantial is high enough to make it unwise to take reports of marital happiness and satisfaction at face value. Generally, only trends in the reports, and differences among categories of married persons, are worthy of interpretation. Of course, changes and differences in response bias may occasionally affect trend and comparative data.