This paper examines biases that may occur when models of couple behavior are estimated using one partner’s report of own and of spouse’s desired family size rather than independent reports obtained from both partners. When estimates of random measurement error are incorporated in a model of couple fertility expectations, proxy reports are valid indicators of spouse desires. In particular, there is little evidence that proxy reports are affected by systematic errors arising from projection of own beliefs onto the spouse. For desired family size, random measurement error in both proxy and self reports is of far greater concern than is systematic error.

The text of this article is only available as a PDF.
You do not currently have access to this content.