Size performance of some tests in one-way anova

Gamage, Jinadasa; Weerahandi, Samaradasa

doi:10.1080/03610919808813500

Cited by 83 publications

(34 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used Chi Square and ANOVAs to compare the demographic characteristics of segments and one-way between-segment MANOVAs across Bullying Prevention Attitudes Scale scores, importance scores, and utility values. When MANOVAs were statistically significant, Brown-Forsythe ANOVAs (Gamage and Weerahandi 1998) and post hoc Dunnett's C tests were used to compare the means for each segment.…”

Section: Discussionmentioning

confidence: 99%

Modeling the Bullying Prevention Program Preferences of Educators: A Discrete Choice Conjoint Experiment

Cunningham

Vaillancourt

Rimas

et al. 2009

J Abnorm Child Psychol

View full text Add to dashboard Cite

We used discrete choice conjoint analysis to model the bullying prevention program preferences of educators. Using themes from computerized decision support lab focus groups (n=45 educators), we composed 20 three-level bullying prevention program design attributes. Each of 1,176 educators completed 25 choice tasks presenting experimentally varied combinations of the study's attribute levels. Latent class analysis yielded three segments with different preferences. Decision Sensitive educators (31%) preferred that individual schools select bullying prevention programs. In contrast, Support Sensitive educators (51%) preferred that local school boards chose bullying prevention programs. This segment preferred more logistical and social support at every stage of the adoption, training, implementation, and long term maintenance processes. Cost Sensitive educators (16%) showed a stronger preference for programs minimizing costs, training, and implementation time demands. They felt prevention programs were less effective and that the time and space in the curriculum for bullying prevention was less adequate. They were less likely to believe that bullying prevention was their responsibility and more likely to agree that prevention was the responsibility of parents. All segments preferred programs supported by the anecdotal reports of colleagues from other schools rather than those based on scientific evidence. To ensure that the bullying prevention options available reflect the complex combination of attributes influencing real world adoption decisions, program developers need to accommodate the differing views of the Decision, Support, and Cost Sensitive segments while maximizing the support of parents and students.

show abstract

Section: Discussionmentioning

confidence: 99%

Modeling the Bullying Prevention Program Preferences of Educators: A Discrete Choice Conjoint Experiment

Cunningham

Vaillancourt

Rimas

et al. 2009

J Abnorm Child Psychol

View full text Add to dashboard Cite

show abstract

“…One-way ANOVA analyses were conducted due to the size of the sample, and the robustness of this procedure. 8,11 …”

Section: Discussionmentioning

confidence: 99%

Are there similarities in physical fitness characteristics of successful candidates attending law enforcement training regardless of training cohort?

Lockie¹,

Stierli²,

Dawes³

et al. 2018

Journal of Trainology

View full text Add to dashboard Cite

Objectives:The aim of this study was to determine if there were differences in physical fitness performance across different cohorts of successful candidates attending law enforcement agency (LEA) training. Design and Methods:Retrospective, non-identifiable, data from three training cohorts, comprising 226 LEA candidates (♂ = 196: ♀ = 30) were analyzed. Data from a standard testing battery used to screen new candidates on entry were used to inform physical fitness performance. This battery included: maximal number of push-up and sit-up repetitions in 60 seconds (s); a 75-yard pursuit run (75PR) around a pre-determined course designed to mimic a foot pursuit; an arm ergometer test where candidates completed as many revolutions in 60 s; and the 2.4 kilometer (km) run. A one-way analysis of variance with Bonferroni post hoc adjustments compared age and test results across the classes. Alpha levels were set at p < 0.05 a priori.Results: There were no significant (p = 0.091-0.458) differences between the three cohorts for age, maximal number of push-ups and sit-ups completed in 60 s, time to complete the 75PR, number of revolutions completed in the 60 s arm ergometer test, or time to complete the 2.4 km run. Conclusions:The level of physical fitness for new candidates attending LEA training, as measured by the testing battery, was similar across cohorts attending training and as such physical training programs to prepare new candidates for LEA duties may not need to be different. However, trainers should be aware of individual variations in physical characteristics within classes to optimize individual gains.(Journal of Trainology 2018;7:5-9)

show abstract

“…The actual size of this test is equal to P PVALW < H 0 . A series of studies (see Gamage and Weerahandi, 1998;Thursby, 1992; and the references therein) have demonstrated that the Weerahandi test is stable in size and exhibits good power. In particular, Thursby (1992) constructed an extensive simulation study which considered a wide variety of tests and rated the Weerahandi test as a "preferred" test in his study.…”

Section: A Heuristic Procedures For Estimating the Critical Value Of F Amentioning

confidence: 97%

A Test for the Equality of Parameters for Separate Regression Models in the Presence of Heteroskedasticity

Oberhelman¹,

Kadiyala²

2007

Communications in Statistics - Simulation and Computation

View full text Add to dashboard Cite

Testing for the equality of regression coefficients across two regressions is a problem considered by analysts in a variety of fields. If the variances of the errors of the two regressions are not equal, then it is known that the standard large sample F -test used to test the equality of the coefficients is compromised by the fact that its actual size can differ substantially from the stated level of significance in small samples. This article addresses this problem and borrows from the literature on the Behrens-Fisher problem to provide some simple modifications of the large sample test which allows one to better control the probability of committing a Type I error. Empirical evidence is presented which indicates that the suggested modifications provide tests which are superior to well-known alternative tests over a wide range of the parameter space.

show abstract

Size performance of some tests in one-way anova

Cited by 83 publications

References 11 publications

Modeling the Bullying Prevention Program Preferences of Educators: A Discrete Choice Conjoint Experiment

Modeling the Bullying Prevention Program Preferences of Educators: A Discrete Choice Conjoint Experiment

Are there similarities in physical fitness characteristics of successful candidates attending law enforcement training regardless of training cohort?

A Test for the Equality of Parameters for Separate Regression Models in the Presence of Heteroskedasticity

Contact Info

Product

Resources

About