Search Results

You are looking at 1 - 10 of 88 items for :

  • "statistical modeling" x
Clear All
Restricted access

Weimo Zhu, Zorica Nedovic-Budic, Robert B. Olshansky, Jed Marti, Yong Gao, Youngsik Park, Edward McAuley and Wojciech Chodzko-Zajko

Purpose:

To introduce Agent-Based Model (ABM) to physical activity (PA) research and, using data from a study of neighborhood walkability and walking behavior, to illustrate parameters for an ABM of walking behavior.

Method:

The concept, brief history, mechanism, major components, key steps, advantages, and limitations of ABM were first introduced. For illustration, 10 participants (age in years: mean = 68, SD = 8) were recruited from a walkable and a nonwalkable neighborhood. They wore AMP 331 triaxial accelerometers and GeoLogger GPA tracking devices for 21 days. Data were analyzed using conventional statistics and highresolution geographic image analysis, which focused on a) path length, b) path duration, c) number of GPS reporting points, and d) interaction between distances and time.

Results:

Average steps by subjects ranged from 1810−10,453 steps per day (mean = 6899, SD = 3823). No statistical difference in walking behavior was found between neighborhoods (Walkable = 6710 ± 2781, Nonwalkable = 7096 ± 4674). Three environment parameters (ie, sidewalk, crosswalk, and path) were identified for future ABM simulation.

Conclusion:

ABM should provide a better understanding of PA behavior’s interaction with the environment, as illustrated using a real-life example. PA field should take advantage of ABM in future research.

Restricted access

Lew Hardy, Tim Woodman and Stephen Carrington

This paper examines Hardy’s (1990, 1996a) proposition that self-confidence might act as the bias factor in a butterfly catastrophe model of stress and performance. Male golfers (N = 8) participated in a golf tournament and reported their cognitive anxiety, somatic anxiety, and self-confidence prior to their tee shot on each hole. All anxiety, self-confidence, and performance scores were standardized within participants in order to control for individual differences. The data were then collapsed across participants and categorized into a high self-confidence condition and a low self-confidence condition by means of a median split. A series of two-way (Cognitive Anxiety × Somatic Anxiety) ANOVAs was conducted on each self-confidence condition in order to fag where the maximum Cognitive Anxiety × Somatic Anxiety interaction effect size lay along the somatic anxiety axis. These ANOVAs revealed that the maximum interaction effect size between cognitive and somatic anxiety was at a higher level of somatic anxiety for the high self-confidence condition than for the low self-confidence condition, thus supporting the moderating role of self-confidence in a catastrophe model framework.

Restricted access

Kathleen F. Janz and Fatima Baptista

bone formation ( 7 , 21 ). Our second highlighted paper, Gabel et al ( 5 ), used novel approaches in accelerometer data reduction and a mixed-effects statistical model to separate the effect of total volume of vigorous physical activity from the frequency of short bouts of vigorous physical activity

Restricted access

John H. Challis

Segmental moment of inertia values, which are often required to perform mechanical analyses of human movement, are commonly computed using statistical models based on cadaver data. Two sets of equations for estimating human limb moments of inertia were evaluated: linear multivariable equations and nonlinear equations. Equation coefficients for both sets of equations were determined using the cadaver data of Chandler et al. (1975). A cross-validation procedure was used to circumvent the problem of model evaluation when there is limited data with which to develop and evaluate the model. Moment of inertia values for the longitudinal axes were predicted with similar degrees of accuracy with either set of equations, while for the transverse axes the nonlinear equations were superior. An evaluation of the influence of the accuracy of moment of inertia estimates on resultant joint moments for three activities showed that the influence of these errors was generally small.

Restricted access

Luis F. Aragón-Vargas and M. Melissa Gross

The purpose of this study was to investigate the kinesiological factors that distinguish good jumpers from poor ones, in an attempt to understand the critical factors in vertical jump performance (VJP). Fifty-two normal, physically active male college students each performed five maximal vertical jumps with arms akimbo. Ground reaction forces and video data were collected during the jumps. Subjects' strength was tested isometrically. Thirty-five potential predictor variables were calculated for statistical modeling by multiple-regression analysis. At the whole-body level of analysis, the best models (which included peak and average mechanical power) accounted for 88% of VJP variation (p < .0005). At the segmental level, the best models accounted for 60% of variation in VJP (p < .0005). Unexpectedly, coordination variables were not related to VJP. These data suggested that VJP was most strongly associated with the mechanical power developed during jump execution.

Restricted access

Asger R. Pedersen, Peter W. Stubbs and Jørgen F. Nielsen

The aim was to challenge the assumptions of standard statistical analyses of average surface electromyography (sEMG) data as a measurement of response magnitudes following the generation of a reflex. The ipsilateral tibial nerve was stimulated at three stimulation intensities and the response sEMG was measured in the contralateral soleus (cSOL) muscle. The magnitude of the cSOL response was measured at a set time window following ipsilateral tibial nerve stimulation. The averaged and trial-by-trial response magnitudes were assessed and compared. The analysis of the averaged and trial-by-trial response revealed significantly different results as the trial-by trial response magnitudes were log-normally distributed with between subject variance heterogeneity violating assumptions of standard statistical analyses. A statistical model has been suggested for the analysis of the responses. By ignoring trial-by-trial response variability and distribution, erroneous results may occur. This may change the interpretation of the results in some studies.

Restricted access

Deborah L. Feltz, Graig M. Chow and Teri J. Hepler

The Feltz (1982) path analysis of the relationship between diving efficacy and performance showed that, over trials, past performance was a stronger predictor than self-efficacy of performance. Bandura (1997) criticized the study as statistically “overcontrolling” for past performance by using raw past performance scores along with self-efficacy as predictors of performance. He suggests residualizing past performance by regressing the raw scores on self-efficacy and entering them into the model to remove prior contributions of self-efficacy imbedded in past performance scores. To resolve this controversy, we reanalyzed the Feltz data using three statistical models: raw past performance, residual past performance, and a method that residualizes past performance and self-efficacy. Results revealed that self-efficacy was a stronger predictor of performance in both residualized models than in the raw past performance model. Furthermore, the influence of past performance on future performance was weaker when the residualized methods were conducted.

Restricted access

Liz Wanless and Jeffrey L. Stinson

While managing the intercollegiate athletic development office is critical to contributions generation, the nearly 40 years of research modeling intercollegiate athletic fundraising emphasized limited factors external to this department. Both theoretical and statistical justification warrants a broader scope in contemporary factor identification. With a resource-based view as the theoretical foundation, a list of 43 variables both internal and external to the intercollegiate athletic development office was generated through an extensive literature review and semistructured interviews with athletic and nonathletic fundraising professionals. Based on the factors identified, random and fixed effects regression models were developed via test statistic model reduction across a 5-year panel (FY2011–FY2015). Ninety-three schools were included, representing 73% of the Football Bowl Subdivision (FBS) membership (85% of public FBS institutions). The results highlight the role of both internal and external factors in explaining intercollegiate athletic fundraising procurement.

Restricted access

G.V. Kondraske and P.J.H. Beehler

Traditional human performance research methods have consisted of multiple regression statistical models based on data such as physical size parameters, reaction times, running speeds, and jumping power. Despite widespread use over many years, the success achieved with these methods has been poor to mediocre. Robust methods for prediction and discovery of insights into human performance remain illusive. The purpose of this paper is to introduce General Systems Performance Theory (GSPT) and the Elemental Resource Model (ERM) for human performance into the fields of physical education and sport. This theory and model collectively represent a new methodological approach with unique features that include: 1) modeling and measurement of all aspects of performance using resource constructs, 2) the use of cause-and-effect resource economic principles (i.e., the idea of threshold “costs” for achieving a given level of performance in any given high level task), and 3) the concept of monadology (i.e., the use of a set of “elements” to explain a complex phenomenon). Although the ERM is intended to encompass all attributes of performance of all human subsystems and to apply to any circumstance involving a human and task, we focus here on relevance and application to gender-related issues in physical activities. This is achieved, after presenting an overview of the ERM, by means of a description and discussion of a set of hypothetical experiments that may be used as a guide for conducting future research. Based on our preliminary investigations, we suggest that it may be appropriate to question the common practice of anticipating and seeking correlations between high level task performance and routinely acquired measures of more basic aspects of performance (e.g., the resources). In contrast to traditional statistical modeling methods, the new concepts and methods represent a cause-and-effect approach that is more similar to the process that an engineer uses to design a system capable of performing a specified task. We believe that the ERM and its associated methods offer a promising basis for a broad spectrum of research into often controversial, gender-related human performance issues and we encourage more widespread investigation, refinement, and implementation of the ERM and GSPT.

Restricted access

Greg Lindsey, Yuling Han, Jeffrey Wilson and Jihui Yang

Purpose:

To model urban trail traffic as a function of neighborhood characteristics and other factors including weather and day of week.

Methods:

We used infrared monitors to measure traffic at 30 locations on five trails for periods ranging from 12 months to more than 4 y. We measured neighborhood characteristics using geographic information systems, satellite imagery, and US Census and other secondary data. We used multiple regression techniques to model daily traffic.

Results:

The statistical model explains approximately 80% of the variation in trail traffic. Trail traffic correlates positively and significantly with income, neighborhood population density, education, percent of neighborhood in commercial use, vegetative health, area of land in parking, and mean length of street segments in access networks. Trail traffic correlates negatively and significantly with the percentage of neighborhood residents in age groups greater than 64 and less than 5.

Conclusions:

Trail traffic is significantly correlated with neighborhood characteristics. Health officials can use these findings to influence the design and location of trails and to maximize opportunities for increases in physical activity.