A new technique for measuring household changes

Because many recent policy decisions have been aimed at effecting changes in the socioeconomic characteristics of families or households, it has become necessary to isolate policy-induced changes from demographic changes in households over time. To obtain such longitudinal data, the family records from three panels of the Current Population Survey that were interviewed both in March, 1964, and March, 1965, were used in a computer record-matching operation.The resulting data confirm that approximately 20 percera of all households are mobile in the period of a year as evidenced by the nonmatched households that were found in 1964 but not present in 1965. More important, the data indicate that S percent of the nonmobile family households became individual households, or the reverse, and 15 percent of all the remaining households changed in family size. These last two statistics represent the first national estimates of gross changes in the demographic characteristics of households.In addition to data on changes in households, this relatively inexpensive method can be used to match Current Population Survey persons' records and provide longitudinal data on the persons within households. Over all, this prototype technique offers policy planners an analytical tool with the necessary statistical controls for assessing the effects of policy decisions and predicting policy success.