Abstract
In many areas of research, the round-robin design is used to study interpersonal judgments and behaviors. The resulting data are analyzed with the social relations model (SRM), whereby almost all previously published studies have used ANOVA-based methods or multilevel-based methods to obtain SRM parameter estimates. In this article, the SRM is embedded into the linear mixed model framework, and it is shown how restricted maximum likelihood can be employed to estimate the SRM parameters. It is also described how the effect of covariates on the SRM-specific effects can be estimated. An example is presented to illustrate the approach. We also present the results of a simulation study in which the performance of the proposed approach is compared to the ANOVA method.
Similar content being viewed by others
References
Bond, C. F., & Lashley, B. R. (1996). Round-robin analysis of social interaction: Exact and estimated standard errors. Psychometrika, 61, 303–311.
Bonito, J. A., & Kenny, D. A. (2010). The measurement of reliability of social relations components from round-robin designs. Personal Relationships, 17, 235–251.
Demidenko, E. (2004). Mixed models: Theory and Applications with R (2nd ed.). Hoboken, NJ: Wiley.
Genz, A., Bretz, F., Miwa, T., Mi, X., Leisch, F., Scheipl, F., & Hothorn, T. (2015). Mvtnorm: Multivariate normal and t distributions. Retrieved February 24, 2015, from http://cran.r-project.org/web/packages/mvtnorm.
Gill, P. S., & Swartz, T. B. (2001). Statistical analyses for round robin interaction data. Canadian Journal of Statistics, 29, 321–331.
Horn, E. M., Collier, W. G., Oxford, J. A., Bond, C. F., & Dansereau, D. F. (1998). Individual differences in dyadic cooperative learning. Journal of Educational Psychology, 90, 153–161.
Jiang, J. (2007). Linear and Generalized Linear Mixed Models and Their Applications. New York: Springer.
Kenny, D. A. (1994). Interpersonal Perception: A Social Relations Analysis. New York: Guilford Press.
Kenny, D. A. (1998). SOREMO (Computer Software and Manual). Storrs, CT: University of Connecticut.
Kenny, D. A. (2007). Estimation of SRM Using Specialized Software. Storrs, CT: University of Connecticut.
Kenny, D. A., & Livi, S. (2009). A componential analysis of leadership using the social relations model. In F. J. Yammarino & F. Dansereau (Eds.), Multi-level Issues in Organizational Behavior and Leadership. Bingley: Emerald.
Kenny, D. A., Albright, L., Malloy, T. E., & Kashy, D. A. (1994). Consensus in interpersonal perception: Acquaintance and the big five. Psychological Bulletin, 116, 245–258.
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). The Analysis of Dyadic Data. New York: Guilford Press.
Küfner, A. C. P., Nestler, S., & Back, M. D. (2012). The two pathways to being an (un-)popular narcissist. Journal of Personality, 81, 184–195.
Lam, C., Van der Vegt, G. S., Walter, F., & Huang, X. (2011). Harming high performers: Social comparison and interpersonal harming in work teams. Journal of Applied Psychology, 96, 588–601.
Lashley, B. R., & Bond, C. F. (1997). Significance testing for round robin data. Psychological Methods, 2, 278–291.
Lashley, B. R., & Kenny, D. A. (1998). Power estimation in social relations analyses. Psychological Methods, 3, 328–338.
Li, H., & Loken, E. (2002). A unified theory of statistical analysis and inference for variance component models for dyadic data. Statistica Sinica, 12, 519–535.
Lüdtke, O., Robitzsch, A., Kenny, D. A., & Trautwein, U. (2013). A general and flexible approach to estimating the social relations model using Bayesian methods. Psychological Methods, 18, 101–119.
McCulloch, C. E., Searle, S. R., & Neuhaus, J. M. (2004). Generalized, Linear, and Mixed Models (2nd ed.). Hoboken, NJ: Wiley.
Nestler, S., Grimm, K. J., & Schönbrodt, F. D. (2015). The social consequences and mechanisms of personality: How to analyse longitudinal data from individual, dyadic, round-robin and network designs. European Journal of Personality, 29, 272–295.
R Core Team. (2014). R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing.
Schaalje, G. B., McBride, J. B., & Fellingham, G. W. (2002). Adequacy of approximations to distributions of test statistics in complex linear mixed models. Journal of Agricultural, Biological and Environmental Statistics, 7, 512–524.
Schönbrodt, F. D., Back, M. D., & Schmukle, S. C. (2012). TripleR: An R package for social relations analyses based on round-robin designs. Behavior Research Methods, 44, 455–470.
Schönbrodt, F. D., Back, M. D., & Schmukle, S. C. (2015). TripleR: A package for round robin analyses using R. Retrieved February 24, 2015, from http://cran.r-project.org/web/packages/TripleR.
Searle, S. R., Casella, G., & McCulloch, C. E. (1992). Variance Components. Hoboken, NJ: Wiley.
Snijders, T. A. B., & Kenny, D. A. (1999). The social relations model for family data: A multilevel approach. Personal Relationships, 6, 471–486.
Verbeke, G., & Molenberghs, G. (2009). Linear Mixed Models for Longitudinal Data. Berlin: Springer.
Warner, R. M., Kenny, D. A., & Stoto, M. (1979). A new round robin analysis of variance for social interaction data. Journal of Personality and Social Psychology, 37, 1742–1757.
Wong, G. Y. (1982). Round robin analysis of variance via maximum likelihood. Journal of the American Statistical Association, 77, 714–724.
Acknowledgments
We would like to thank Sarah Humberg and three anonymous reviewers for very helpful comments on an earlier draft of this manuscript.
Author information
Authors and Affiliations
Corresponding author
Additional information
This article is dedicated to Irmgard Laufer.
Appendices
Appendix 1: Variance–Covariance Matrix of the Judgment Vector y
Equation 5, the basic assumptions outlined above, and some standard results concerning the variance–covariance matrix of a vector of random variables are used to derive the variance–covariance matrix of the judgment vector y. Please note that
Furthermore,
and
The variance–covariance matrix of the judgment vector y then is
Appendix 2: The Other Social Relation Models
The vector of round-robin judgments y for the basic SRM (see Equation 1) is
where \(y_{1} = (y_{12}, \ldots , y_{(m-1)m})\) and \(y_{2} = (y_{21}, \ldots , y_{m(m-1)})\). The vector of actor effects is \(\alpha = (\alpha _{1}, \ldots , \alpha _{m})\) and the vector of partner effects is \(\beta = (\beta _{1}, \ldots , \beta _{m})\). \(\gamma _1\) and \(\gamma _2\) denote the relationship effect vectors. Both are defined analogously to \(y_1\) and \(y_2\), respectively. b denotes the general mean, X is a matrix that contains 1s only, and \(Z_i\), \(i = 1, \ldots , 3\), are design matrices containing 0s and 1s (indicating the presence of a random effect). In the basic SRM, relationship effects cannot be separated from error, hence \(Z_3 = I_N\) (where N is the number of judgments). The variance–covariance matrix V of the model is
For model estimation it has to be assumed that there are at least four group members, \(m \ge 4\), and at least one measure per dyad.
The vector of judgments y for the single-measure, multiple group SRM is
where \(y_{1} = (y_{121}, \ldots , y_{(m-1)mk})\) and \(y_{2} = (y_{211}, \ldots , y_{m(m-1)k})\). The vector of group effects is \(g = (g_1,\ldots ,g_k)\), the actor effect vector is \(\alpha = (\alpha _{1k}, \ldots , \alpha _{mk})\), and the vector of partner effects is \(\beta = (\beta _{1k}, \ldots , \beta _{mk})\). The relationship effect vectors, \(\gamma _1\) and \(\gamma _2\), are again defined analogously to \(y_1\) and \(y_2\), respectively. b contains the general mean, X is a matrix of 1s, and \(Z_g\), \(Z_i\), \(i = 1, \ldots , 3\), are the design matrices. Similar to the basic SRM, relationship effects cannot be separated from error in this SRM, thus \(Z_3 = I_N\). The variance–covariance matrix of this SRM is
To estimate the model parameters, it has to be assumed that there are at least two groups, \(k \ge 2\), with four group members each, \(m \ge 4\), and one measure per dyad, \(l = 1\).
Finally, the judgment vector y for the multiple-measure, one group SRM (see Equation 3) can be represented by
\(y_{1} = (y_{121},\ldots ,y_{12l} \ldots , y_{(m-1)ml})\) and \(y_{2} = (y_{211}, \ldots ,y_{21l},\ldots , y_{m(m-1)l})\). The actor effect vector is \(\alpha = (\alpha _{1}, \ldots , \alpha _{m})\) and the vector of partner effects is \(\beta = (\beta _{1} \ldots , \beta _{m})\). The vector \(\gamma _1 = (\gamma _{12}, \ldots , \gamma _{(m-1)m})\) and the vector \(\gamma _2 = (\gamma _{21}, \ldots , \gamma _{m(m-1)})\) denote the relationship effects. The vector of errors are defined analogously to \(y_1\) and \(y_2\), respectively. Again, b represents the general mean, X is a matrix containing 1s, and \(Z_i\), \(i = 1, \ldots , 3\), are design matrices. In this model, relationship effects can be separated from error, hence \(Z_3 \ne I_N\). The variance–covariance matrix of this model is
To estimate the model, there have to at least four group members, \(m \ge 4\), and two measures per dyad, \(l \ge 2\).
Rights and permissions
About this article
Cite this article
Nestler, S. Restricted Maximum Likelihood Estimation for Parameters of the Social Relations Model. Psychometrika 81, 1098–1117 (2016). https://doi.org/10.1007/s11336-015-9474-9
Received:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11336-015-9474-9