Inferences for Fixed Effects Based Regression Parameters in a Finite Population Setup Using Two-stage Cluster Sample

Brajendra C. Sutradhar ()
Additional contact information
Brajendra C. Sutradhar: Memorial University

Sankhya A: The Indian Journal of Statistics, 2024, vol. 86, issue 2, No 10, 991 pages

Abstract: Abstract In a clusters based infinite/super population (SP) setup, the inference complexity for regression parameters depend on the cluster correlations structure oriented marginal fixed or mixed effects based mean models, where as opposed to the fixed effects based models the mixed models exhibit both regression and cluster variance/correlation parameters in the mean structure especially for discrete such as clustered counts and binary data. On the other hand, as recently discussed by Sutradhar (2022, Sankhya B, 84, 259-302), in the fixed models setup the modelling of cluster correlations can be complex. In the finite population (FP) setup where FP data follow a SP model but the regression effects estimation is done using a two-stage cluster sample, the inferences become much more complex. As opposed to the recent mixed models based FP inferences (Sutradhar 2023b, Annals of the Institute of Statistical Mathematics, 75, 425-462), there does not appear any studies for fixed models based analysis for exponential family data where true correlation structures play important role in defining FP regression parameters. To resolve this inference issue, in this paper we make following specific contributions. First, the cluster correlations for linear, counts and binary data are developed in a SP setup such that the marginal mean models depend only on the fixed regression effects. Second the FP data which are hypothetical or unobserved until a sample is taken to observe a part, are utilized to develop hypothetical estimating equations for the SP regression parameters, which subsequently define the FP regression and correlation parameters. Third, the design weighted unbiased estimators are obtained for the FP regression and correlation parameters, where the specific formulas for correlation estimators depend on the nature of the response data whether linear, counts or binary. Also, as an additional application of the regression effects estimation, the FP total prediction is discussed. Next, the design consistency of the regression estimators is developed in details.

Keywords: Cluster correlation structure yielding fixed effects based means; Design unbiasedness and design consistency; Finite population parameters arising from a super population model; Two-stage cluster sample from a finite population; Primary 62F10; 62H20; Secondary 62F12 (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://link.springer.com/10.1007/s13171-024-00362-w Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:sankha:v:86:y:2024:i:2:d:10.1007_s13171-024-00362-w

Ordering information: This journal article can be ordered from
http://www.springer.com/statistics/journal/13171

DOI: 10.1007/s13171-024-00362-w

Access Statistics for this article

Sankhya A: The Indian Journal of Statistics is currently edited by Dipak Dey

More articles in Sankhya A: The Indian Journal of Statistics from Springer, Indian Statistical Institute
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().