A Caveat on the Use of the Quadratic Assignment Procedure

The Quadratic Assignment Procedure (QAP) has been suggested as a test of fit for structural data. An argument is made that this test is inappropriate on logical grounds and because its application leads to results that are difficult to interpret. By generating random samples under two simple models, this test is shown to be biased; moreover, the bias is not consistent but rather can be liberal or conservative, small or large, depending on parameters in the population from which the data are sampled: Quadratic Assignment Procedure, Goodness-of-fit, Network Analysis.

[1]  Mark S. Granovetter.  Networks of Collective Action: A Perspective on Community Influence Systems.Edward O. Laumann , Franz U. Pappi , 1978 .

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

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

[4]  Robert R. Sokal,et al.  Testing Statistical Significance of Geographic Variation Patterns , 1979 .

[5]  John A. Endler,et al.  Quantitative Matrix Comparisons in Ecological and Evolutionary Investigations , 1982 .

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

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

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

[9]  M M Dow,et al.  Nonparametric inference procedures for multistate life table analysis. , 1985, The Journal of mathematical sociology.

[10]  E. Laumann,et al.  Networks of Collective Action: A Perspective on Community Influence Systems , 1976 .

[11]  L. Hubert.  Assignment methods in combinatorial data analysis , 1986 .

[12]  Denton E. Morrison,et al.  The Significance Test Controversy , 1972 .

[13]  D. Krackhardt,et al.  Friendship Patterns and Culture: The Control of Organizational Diversity , 1990 .

[14]  Keiko Nakao,et al.  A Method for Testing Alternative Theories: An Example from English Kinship , 1984 .

[15]  P. Mielke,et al.  Clarification and Appropriate Inferences for Mantel and Valand's Nonparametric Multivariate Analysis Technique , 1978 .

[16]  Combinatorial procedures for the analysis of alternate models: An example from interregional migration , 1983 .

[17]  David Krackhardt,et al.  PREDICTING WITH NETWORKS: NONPARAMETRIC MULTIPLE REGRESSION ANALYSIS OF DYADIC DATA * , 1988 .

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

[19]  M M Dow,et al.  Comparison of distance matrices in studies of population structure and genetic microdifferentiation: quadratic assignment. , 1985, American journal of physical anthropology.

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

[21]  E. Dietz.  Permutation Tests for Association Between Two Distance Matrices , 1983 .

[22]  David Krackhardt,et al.  The snowball effect: Turnover embedded in communication networks. , 1986 .

[23]  L. Hubert,et al.  COMBINATORIAL PROCEDURES FOR THE ANALYSIS OF ALTERANTE MODELS: AN EXAMPLE FROM INTERREGIONAL MIGRATION , 1983 .