Estimating Macro–Relationships Using Micro–Data: A One–Stage Approach

TY - JOUR

T1 - Estimating Macro–Relationships Using Micro–Data

T2 - A One–Stage Approach

AU - Krivo, Lauren J.

AU - Kaufman, Robert L.

PY - 1990/11

Y1 - 1990/11

N2 - Widespread recognition that many sociological processes occur at multiple levels has led researchers to search for valid forms of cross-level modeling such as contextual analysis. In this article, we propose an alternative cross-level method for examining the effects of macro-factors on macro-processes while controlling for processes at a lower-level of aggregation. This method applies existing statistical techniques but in a new fashion. It has the advantages of more precise control for lower-order effects, reduction of problems with degrees of freedom, and more error-free prediction of the higher-level relationships than using alternative estimation strategies. As a result, parameter estimates of the macro-level effects are more efficient. An empirical example studying macro-determinants of unemployment rates in states in the United States demonstrates that estimation of this model leads to more valid conclusions than more commonly used methods.

AB - Widespread recognition that many sociological processes occur at multiple levels has led researchers to search for valid forms of cross-level modeling such as contextual analysis. In this article, we propose an alternative cross-level method for examining the effects of macro-factors on macro-processes while controlling for processes at a lower-level of aggregation. This method applies existing statistical techniques but in a new fashion. It has the advantages of more precise control for lower-order effects, reduction of problems with degrees of freedom, and more error-free prediction of the higher-level relationships than using alternative estimation strategies. As a result, parameter estimates of the macro-level effects are more efficient. An empirical example studying macro-determinants of unemployment rates in states in the United States demonstrates that estimation of this model leads to more valid conclusions than more commonly used methods.

UR - http://www.scopus.com/inward/record.url?scp=84970260119&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84970260119&partnerID=8YFLogxK

U2 - https://doi.org/10.1177/0049124190019002003

DO - https://doi.org/10.1177/0049124190019002003

M3 - Article

SN - 0049-1241

VL - 19

SP - 196

EP - 224

JO - Sociological Methods & Research

JF - Sociological Methods & Research

IS - 2

ER -