A pooled Bayes test of independence for sparse contingency tables from small areas

Nandram, Balgobin; Kim, Dalho; Zhou, Jingran

doi:10.1080/00949655.2019.1574792

Cited by 4 publications

(1 citation statement)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Non-Bayesian test of independence for weighted and correlated data is discussed by Choi and McHugh (1999). Bayesian tests for independence in contingency tables are discussed by Nandram, Kim and Zhou (2019), Bhatta, Nandram and Sedransk (2018), Yu, Bhadra and Nandram (2017), , Nandram, Bhatta and Sedransk (2013), Nandram and Choi (2007); the paper by Nandram and Choi (2007) is our pioneering work on hypothesis testing. The Bayesian test, which is based on the Bayes factor, uses both hypotheses directly where the non-Bayesian test, based on the p-value, mostly uses just the null hypothesis.…”

Section: Introductionmentioning

confidence: 99%

Large Sample Problems

Choi¹,

Nandram²

2021

IJSP

Self Cite

View full text Add to dashboard Cite

Variance is very important in test statistics as it measures the degree of reliability of estimates. It depends not only on the sample size but also on other factors such as population size, type of data and its distribution, and method of sampling or experiments. But here, we assume that these other fasctors are fixed, and that the test statistic depends only on the sample size. When the sample size is larger, the variance will be smaller. Smaller variance makes test statistics larger or gives more significant results in testing a hypothesis. Whatever the hypothesis is, it does not matter. Thus, the test result is often misleading because much of it reflects the sample size. Therefore, we discuss the large sample problem in performing traditional tests and show how to fix this problem.

show abstract