Most data involving organizations are hierarchical in nature and often contain variables measured at multiple levels of analysis. Hierarchical linear modeling (HLM) is a relatively new and innovative statistical method that organizational scientists have used to alleviate some common problems associated with multilevel data, thus advancing our understanding of organizations. This article presents a broad overview of HLM’s logic through an empirical analysis and outlines how its use can strengthen sport management research. For illustration purposes, we use both HLM and the traditional linear regression model to analyze how organizational and individual factors in Major League Baseball impact individual players’ salaries. A key implication is that, depending on the method, parameter estimates differ because of the multilevel data structure and, thus, findings differ. We explain these differences and conclude by presenting theoretical discussions from strategic management and consumer behavior to provide a potential research agenda for sport management scholars.
Samuel Y. Todd, T. Russell Crook and Anthony G. Barilla
Shane N. Sweet, Michelle S. Fortier and Chris M. Blanchard
Because motivation has been deemed a key barrier to physical activity, it is imperative that we know how motivational levels change over time and how that change relates to physical activity. Based in Self-Determination Theory, this study investigated fluctuations in physical activity and motivational regulations over 25 weeks and tested the relationship between these 2 variables.
Data from the Physical Activity Counseling trial were examined. Inactive adults recruited from a primary care center (N = 120) answered motivation and physical activity questionnaires during the intervention and postintervention phases. Hierarchical linear modeling was used to test the hypotheses.
Quadratic changes were found for external regulation (γ20= 0.02, P < .05) and physical activity (γ20 = –2.64, P < .001), while identified (γ10= 0.04, P = .03) and intrinsic (γ10= 0.04, P = .01) regulations increased linearly over the course of the 25 weeks. Only identified regulation (γ30= 3.15, P = .01) and intrinsic motivation (γ30= 4.68, P < .001) were significantly and positively related with physical activity.
Physical activity, external and identified regulations and intrinsic motivation changed over the 25 weeks. Intervention should aim at fostering identified regulation and intrinsic motivation as greater levels of these regulations were related with physical activity.
Xihe Zhu and Justin A. Haegele
affiliation because of the de-identified nature of the existing dataset. Data Analysis Because student-level as well as school-level data are encompassed in the study, we used hierarchical linear modeling (HLM, ver. 6.08; Scientific Software International, Inc., Skokie, IL) for data analysis ( Raudenbush
Rose M. Angell, Stephen A. Butterfield, Shihfen Tu, E. Michael Loovis, Craig A. Mason and Christopher J. Nightingale
of individual growth over time. Statistical Methods Primary analyses focused on hierarchical linear modeling (HLM) and were conducted using HLM 7 (Scientific Software International, Skokie, IL). As previously reported by Butterfield et al. ( 2012 ), HLM explicitly models individual variation in
Denny Meyer, Madawa W. Jayawar, Samuel Muir, David Ho and Olivia Sackett
missing data are addressed using inverse probability weights and hierarchical linear models fitted using maximum likelihood procedures. In the qualitative analysis, we use text mining to analyze the results of an open-ended description of the VPGC program, and then, using machine-learning tools, we
research makes several other contributions. It makes use of a novel and comprehensive data set of player performance from the 2008 NFL season. Measuring player quality is difficult, and even key scoring metrics are inadequate. This study uses an empirical Bayesian hierarchical linear model (HLM) that
Sarah Burkart, Jasmin Roberts, Matthew C. Davidson and Sofiya Alhassan
change in the dependent variables of interest over time, Hierarchical Linear Modeling (version 6; Scientific Software International, Inc, Skokie, IL) 24 was used. This type of analysis estimates individual growth curves, uses them to assess changes in behavior over time, and develops a trajectory
Timothy Martinson, Stephen A. Butterfield, Craig A. Mason, Shihfen Tu, Robert A. Lehnhard and Christopher J. Nightingale
recorded for analysis. Analysis Preliminary analyses examined simple descriptive statistics and comparisons between students at the start of the school year. Given the multiple assessments and potential confounding of some student covariates, primary analyses focused on hierarchical linear modeling (HLM
Paddy C. Favazza, Gary N. Siperstein, Susan A. Zeisel, Samuel L. Odom, John H. Sideris and Andrew L. Moskowitz
This study examined the effectiveness of the Young Athletes program to promote motor development in preschool-aged children with disabilities. In the study, 233 children were randomly assigned to a control group or the Young Athletes (YA) intervention group which consisted of 24 motor skill lessons delivered 3 times per week for 8 weeks. Hierarchical Linear Modeling (HLM) showed that children who participated in the YA intervention exhibited mean gains of 7–9 months on the Peabody Developmental Motor Subscales (PDMS) compared with mean gains of 3–5 months for the control group. Children in the YA intervention also exhibited significant gains on the gross motor subscale of the Vineland Teacher Rating Form (VTRF). Teachers and parents reported benefits for children not only in specific motor skills, but also kindergarten readiness skills and social/play skills. The necessity for direct and intentional instruction of motor skills, as well as the challenges of involving families in the YA program, are discussed.
Susan C. Duncan, Terry E. Duncan, Lisa A. Strycker and Nigel R. Chaumeton
Typical studies of youth physical activity ignore the dependence among family members, examining only individual levels of data rather than individual and family levels. The current study examined physical activity among siblings (mean age = 12.2 years), using hierarchical linear modeling. Individual-and family-level covariates of physical activity were included in the model. Data from 930 siblings nested within 371 families were analyzed in a four-level multilevel design. Results indicated that siblings were similar in their levels of physical activity, and that levels of physical activity varied across families. At the individual level, age was a significant predictor of physical activity. At the family level, higher levels of family support were related to higher levels of sibling physical activity, as were single-parent status and higher income. Perceptions of neighborhood opportunities and observed neighborhood physical activity facilities were negatively related to family levels of physical activity.