2015
DOI: 10.1515/jos-2015-0003
|View full text |Cite
|
Sign up to set email alerts
|

Linear Regression Diagnostics in Cluster Samples

Abstract: An extensive set of diagnostics for linear regression models has been developed to handle nonsurvey data. The models and the sampling plans used for finite populations often entail stratification, clustering, and survey weights, which renders many of the standard diagnostics inappropriate. In this article we adapt some influence diagnostics that have been formulated for ordinary or weighted least squares for use with stratified, clustered survey data. The statistics considered here include DFBETAS, DFFITS, and… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

0
13
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
5

Relationship

1
4

Authors

Journals

citations
Cited by 12 publications
(13 citation statements)
references
References 30 publications
0
13
0
Order By: Relevance
“…Therefore, we did not report any regression diagnostic for our models using multiple imputation. However, interested readers may find a complete description of regression diagnostics for the subsample of models for which regression diagnostics were feasible in Supplemental Analysis S4: linear regressions with complete case analysis (43)(44)(45) . The 'svydiags' package in R was used for these analyses (46) .…”
Section: Discussionmentioning
confidence: 99%
“…Therefore, we did not report any regression diagnostic for our models using multiple imputation. However, interested readers may find a complete description of regression diagnostics for the subsample of models for which regression diagnostics were feasible in Supplemental Analysis S4: linear regressions with complete case analysis (43)(44)(45) . The 'svydiags' package in R was used for these analyses (46) .…”
Section: Discussionmentioning
confidence: 99%
“…Furthermore, we tested if the outlying observations are influential cases or not by performing the following tests: Difference in Fits (DFFITS), Difference in Betas (DFBETAS), and the Cook's Distance. 40 Due to the defined sample size, the strongest explanatory variables were chosen from the literature reviewed. Power analysis using G*Power revealed that a medium-to-large effect size (f 2 = 0.27), models with 6 tested explanatory variables, and a sample of 55 are sufficient to attain power of .8 given, P = .05.…”
Section: Discussionmentioning
confidence: 99%
“…Furthermore, we tested if the outlying observations are influential cases or not by performing the following tests: Difference in Fits (DFFITS), Difference in Betas (DFBETAS), and the Cook’s Distance. 40 …”
Section: Methodsmentioning
confidence: 99%
“…These assumptions were accepted when the p ‐values of the S‐W test and Bartlett's test were higher than the level of significance (Stewart, 1987). Autocorrelation was not verified because the authors did not use time‐series data (Li & Valliant, 2011). All results were calculated using SPSS Statistics.…”
Section: Aim Methodology and Datamentioning
confidence: 99%