Using SAS PROC MIXED to Fit Multilevel Models, Hierarchical Models, and Individual Growth Models (1998) by Judith Singer is “a step-by-step tutorial that shows how to use SAS to fit the two most common multilevel models: (1) two-level models, designed for data on individuals nested within naturally occurring hierarchies (e.g., students within classes) and (2) individual growth models, designed for exploring longitudinal data (on individuals) over time.” (Singer, 1998, p. 323) The main findings of the paper are the code for three level models and the code for working with multilevel data in SAS provided in the conclusion and the appendix. The paper states that various models cannot be applied in education and the behavioral science without appropriate software.
While reading the paper some questions might arise. One of them concerns the validity of the hypothesis tests used. On the one hand, such questions rely on large sample approximation, on the other, “variance components are known to have skewed sampling distributions that render normal approximations such as these questionable.” (p. 351) What remains unclear is how empowered the author was to resort to these very tests in her research.
Another remark that arises is that the paper lacks somehow the discussion of the differences between various methods used and the consequences of these differences. We suppose that the discussions of the type would have made the paper sound more objective.
The article Multilevel Theory Building: Benefits, Barriers, and New Developments (1999) by Klein et al. reviews the benefits of and barriers to multilevel theory building and describes new development in this theory. The paper contributes significantly to the reader’s knowledge on multilevel theory and outlines perspectives for further research in the field.
We believe that the main advantage of the paper is that it investigates the multilevel theory building from various angles: from the point of view of micro- and macro- organizational scholars who studied behavior within and of organizations respectively and those who have examined “the influence of the organization on individual and group behavior or the influence on individual and group behavior on the organization as a whole.” (Klein, 1999, p. 247) The main findings of the reviewed works are rendered by the authors in a clear language that makes the reader involved into further studying the problem under analysis.
A shift from single-level to multilevel research appears to be daunting. The abundance of literature discussing various ways of conceptualizing multilevel models, justifying aggregation and evaluating the models engenders some confusion with the readers. The article From Micro to Meso: Critical Steps in Conceptualizing and Conducting Multilevel Research (2000) by Klein and Kozlovski aims at cutting through this confusion by providing guidance with respect to theory, research methods, and analyses (Klein & Kozlovski, 2000, p. 212). Regarding the complex nature of the research done we should also admit that it seems too complicated for achieving the purpose set. We are inclined to believe that the paper lacks the authors’ explanations of numerous aspects of the problem discussed. Sometimes only a mere overview of the existing literature is presented without adequate evaluation of it. Therefore, the paper does not seem to help the researcher in his/her search, but, on the contrary, it only enhances one’s confusion as far as the existing research on the problem is concerned.
Various researches on the problem of service performance have taken only one side of analysis: they have focused either on organizational level or on individual level of analysis. The research A Multilevel Investigation of Factors Influencing Employee Service Performance and Customer Outcomes (2004) conducted by H. Liao and A. Chuang bridged up the gap between the two approaches. Actually, this is the first study that brought the two perspectives together and tested a multilevel model of the antecedents and consequences of employee service performance. One of the main benefits of the paper is that it both investigates the existing knowledge on the problem and presents its own, new, findings. As for the drawbacks, we believe that the paper needs a more detailed discussion of the practical implications of the study for service organizations. The paper contains a lot of theoretical information and it is desired that it should also have demonstration of how this knowledge can be implemented.
References
Klein, K. J. et al.(1999). Multilevel theory building: Benefits, barriers, and new developments. Academy of Management Review, 24(2), 243-248.
Klein, K. J.& Kozlowski, S. W. J. (2000) From micro to meso: Critical steps in conceptualizing and conducting multilevel research. Organizational Research Methods, 3(3), 211-236.
Liao, H. & Chuang, A. (2004). A multilevel investigation of factors influencing employee service performance and customer outcomes. Academy of Management Journal, 47(1), 41-58.
Singer, J. D. (1998). Using SAS PROC MIXED to fit multilevel models, hierarchical models, and individual growth models. Journal of Educational and Behavioral Statistics, 24(4), 323-355.