PREDICTING WITH NETWORKS: NONPARAMETRIC MULTIPLE REGRESSION ANALYSIS OF DYADIC DATA *

Abstract This paper argues that the quadratic assignment procedure (QAP) is superior to OLS for testing hypothesis in both simple and multiple regression models based on dyadic data, such as found in network analysis. A model of autocorrelation is proposed that is consistent with the assumptions of dyadic data. Results of Monte Carlo simulations indicate that OLS analysis is statistically biased, with the degree of bias varying as a function of the amount of structural autocorrelation. On the other hand, the simulations demonstrate that QAP is relatively unbiased. The Sampson data are used to illustrate the QAP multiple regression procedure and a general method of testing whether the results are statistically biased.

[1]  L. Freeman.  Centrality in social networks conceptual clarification , 1978 .

[2]  S. Boorman,et al.  Social Structure from Multiple Networks. I. Blockmodels of Roles and Positions , 1976, American Journal of Sociology.

[3]  P. Pattison,et al.  Cumulated social roles: The duality of persons and their algebras☆ , 1986 .

[4]  R. Burt,et al.  Applied Network Analysis: A Methodological Introduction , 1983 .

[5]  Walter Krämer,et al.  Finite Sample Efficiency of Ordinary Least Squares in the Linear Regression Model With Autocorrelated Errors , 1980 .

[6]  Lawrence Hubert,et al.  Inference Procedures for the Evaluation and Comparison of Proximity Matrices , 1983 .

[7]  Lawrence Hubert,et al.  The Analysis of Social Interaction Data , 1981 .

[8]  K. Reitz,et al.  Galton's Problem as network autocorrelation , 1984 .

[9]  Reginald G. Golledge,et al.  A heuristic method for the comparison of related structures , 1981 .

[10]  Robert F. Engle,et al.  The Specification of the Disturbance for Efficient Estimation , 1974 .

[11]  G. Judge,et al.  The Theory and Practice of Econometrics , 1981 .

[12]  Katherine Faust,et al.  The effect of skewed distributions on matrix permutation tests , 1985 .

[13]  James R. Lincoln,et al.  Analyzing Relations in Dyads , 1984 .

[14]  David Krackardt,et al.  QAP partialling as a test of spuriousness , 1987 .

[15]  R. Burt.  Toward a structural theory of action , 1982 .

[16]  K. Ord.  Estimation Methods for Models of Spatial Interaction , 1975 .

[17]  N. Mantel.  The detection of disease clustering and a generalized regression approach. , 1967, Cancer research.

[18]  Peter Schmidt,et al.  The Theory and Practice of Econometrics , 1985 .

[19]  L. Hubert,et al.  Quadratic assignment as a general data analysis strategy. , 1976 .

[20]  Lawrence Hubert,et al.  Combinatorial data analysis: Association and partial association , 1985 .

[21]  H. Kraemer,et al.  Statistical analysis of dyadic social behavior. , 1979 .

[22]  PATRICK DOREIAN,et al.  Network Autocorrelation Models , 1984 .

[23]  Paul W. Mielke,et al.  On asymptotic non-normality of null distributions of mrpp statistics , 1979 .