PubHawk
April 2015 PMID 25886883

Sample size calculations for skewed distributions.

Bonnie Cundill, Neal D E Alexander

Full text PDF Download social card

Mentions

8
Peak: 3 in Nov 2019
across 2 sources since Nov 2019

Extended reach

34,141
potential audience

Total likes

186

Video views

3,894

Mentions per month

Nov 2019 – Sep 2025

Where it's discussed

  • YouTube5 videos
  • X / Twitter3 posts

AI synopsis

What people are saying

The discussions reflect enthusiasm and support for the article's relevance to applications like healthcare claims and GLM gamma models, with examples such as technical mentions of 'power analysis' and 'log link,' and hashtags like #rstats and #biometry. The tone emphasizes practical benefits and encourages others to read for methodological insights, indicating high interest and perceived significance.

AgreementStrong agreement

The posts explicitly endorse the article's utility, especially for applying GLM gamma models in healthcare costs, indicating strong agreement.

InterestHigh interest

Posts demonstrate high interest by highlighting specific applications like power analysis and including technical details and hashtags.

EngagementModerate engagement

While sharing practical insights, the discussions are relatively straightforward with limited critical depth, showing moderate engagement.

ImpactHigh impact

These discussions underscore the importance of accurate sample size calculations in healthcare and statistical modeling, hinting at high potential impact on research practices.

Abstract in brief

  • Sample size calculations for skewed distributions often rely on normal approximations, even when data are to be analyzed with generalized linear models (GLMs), which include families like negative binomial, Poisson, and gamma.
  • Using simulations, the study assesses the effectiveness of these normal approximations, finding that for negative binomial and gamma distributions, calculations on the log link scale are accurate and often better than simple normal approximations, while they offer little benefit for Poisson and binomial cases.
  • The method described is particularly useful for determining sample sizes when comparing means of highly skewed data, such as in vaccine trials or studies involving insecticide-treated materials.]
Read the full abstract

Top posts

Most liked on X and Bluesky
Rstats
@rstatstweet · X

RT @AgEconomist: Well worth a read if you need to do power analysis for healthcare claims costs via GLM gamma with log link https://t.co/a3…

Nov 11, 2019 · 2 shares · 28,803 followers View post →
Matt Bogard
@AgEconomist · X

Well worth a read if you need to do power analysis for healthcare claims costs via GLM gamma with log link https://t.co/a3r01A8EY3 #rstats code included #healthcare #econometrics #biometry

Nov 11, 2019 · 2 shares · 1,371 followers View post →

All mentions

السَّـلاَمُ و عَلَیكُم
@Awan_slk77 · X

RT @AgEconomist: Well worth a read if you need to do power analysis for healthcare claims costs via GLM gamma with log link https://t.co/a3…

Nov 11, 2019 · 2 shares · 61 followers View post →
Rstats
@rstatstweet · X

RT @AgEconomist: Well worth a read if you need to do power analysis for healthcare claims costs via GLM gamma with log link https://t.co/a3…

Nov 11, 2019 · 2 shares · 28,803 followers View post →
Matt Bogard
@AgEconomist · X

Well worth a read if you need to do power analysis for healthcare claims costs via GLM gamma with log link https://t.co/a3r01A8EY3 #rstats code included #healthcare #econometrics #biometry

Nov 11, 2019 · 2 shares · 1,371 followers View post →