Revisiting Mothers’ Pensions: A Critique of a Social Science Classic and a New Analysis

Abstract

We identify methodological weaknesses in a paper by Skocpol, Abend-Wein, Howard, and Lehmann (1993) on the origins of mothers’ pensions in the American states in the early twentieth century. These include a sub-optimal and potentially biased strategy for assessing the impact of state characteristics on the time to adoption of pensions, as well as the use of a backward stepwise regression procedure for selecting independent variables. To determine whether Skocpol et al.’s conclusions remain valid, we recreated most of their dataset and used methods that are more appropriate for the analysis of duration data, including the Cox and complementary cloglog event history procedures. While we find support for several of their claims, our findings allow for a more straightforward interpretation of the role of explanatory variables, and the temporal dependence of the adoption process.

Citation

Olasky, S.J. and Greenberg, D.F. (2017), "Revisiting Mothers’ Pensions: A Critique of a Social Science Classic and a New Analysis", On the Cross Road of Polity, Political Elites and Mobilization (Research in Political Sociology, Vol. 24), Emerald Group Publishing Limited, Leeds, pp. 93-119. https://doi.org/10.1108/S0895-993520160000024010