statistics

# Against statistical significance

Amrhein V, Greenland S, McShane B. Scientists rise up against statistical significance.

Should we retire statistical significance? Read Deborah Mayo before deciding!

*Nature*. 2019; 567: 305-307.Should we retire statistical significance? Read Deborah Mayo before deciding!

# Sample size for multinomial logistic regression

de Jong VMT, Eijkemans MJC, van Calster B

"In most cases, the calibration and overall predictive performance of the multinomial prediction model is improved by using penalized maximum likelihood regression"

*et al.*Sample size considerations and predictive performance of multinomial logistic prediction models.*Stat Med*. 2019 (Open Access)."In most cases, the calibration and overall predictive performance of the multinomial prediction model is improved by using penalized maximum likelihood regression"

# Graphics for uncertainty

Bowman A. Graphics for uncertainty. Journal of the Royal Statistical Society: Series A (Statistics in Society) (Open Access). 2019;182:403-418.

Density strips and other tools to graph uncertainty.

Density strips and other tools to graph uncertainty.

# Sample size for logistic regression

van Smeden M, Moons KG, de Groot JA, Collins GS, Altman DG, Eijkemans MJ, and Reitsma JB. Sample size for binary logistic prediction models: Beyond events per variable criteria.

The event per variable (EPV) criterion has limited value for calculating sample size.

*Stat Methods Med Res*962280218784726, 2018.The event per variable (EPV) criterion has limited value for calculating sample size.

# Design of growth charts

Ohuma EO and Altman DG. Design and other methodological considerations for the construction of human fetal and neonatal size and growth charts.

Methodological considerations for the development of growth charts,

*Stat Med*2018.Methodological considerations for the development of growth charts,

# In gentle praise of significance tests

# ⭐️ Cargo-cult statistics

Stark PB, Saltelli A. Cargo-cult statistics and scientific crisis. Significance. 2018;15:40-43. 10.1111/j.1740-9713.2018.01174.x [html]

"They [practitioners] invoke statistical terms and procedures as incantations, with scant understanding of the assumptions or relevance of the calculations, or even the meaning of the terminology"

Wondering what is cargo-cult science ?

"They [practitioners] invoke statistical terms and procedures as incantations, with scant understanding of the assumptions or relevance of the calculations, or even the meaning of the terminology"

Wondering what is cargo-cult science ?

# Sample size for multivariable prediction - 1

11 18 statistical methods

Riley RD, Snell KIE, Ensor J, Burke DL, Harrell FE, Moons KGM, Collins GS. Minimum sample size for developing a multivariable prediction model: Part I - Continuous outcomes. Stat Med. 201810.1002/sim.7993

"In this article, we build on the previous work of Harrell et al to propose how to calculate a suitable sample size for development of a prediction model using linear regression" (part 1).

"In this article, we build on the previous work of Harrell et al to propose how to calculate a suitable sample size for development of a prediction model using linear regression" (part 1).

# Sample size for multivariable prediction - 2

11 18 statistical methods

Riley RD, Snell KI, Ensor J, Burke DL, Harrell FE, Moons KG, Collins GS. Minimum sample size for developing a multivariable prediction model: PART II - binary and time-to-event outcomes. Stat Med. 201810.1002/sim.7992.

How to calculate a suitable sample size for development of a prediction model using binary or time-to-event outcomes (part 2).

How to calculate a suitable sample size for development of a prediction model using binary or time-to-event outcomes (part 2).

# What makes a biostatistician?

Zapf A, Huebner M, Rauch G, Kieser M. What makes a biostatistician?. Stat Med. 201810.1002/sim.7998

"We discuss the required professional expertise for the main areas of applications in the medical field as well as the necessary soft skill competencies of a biostatistician"

"We discuss the required professional expertise for the main areas of applications in the medical field as well as the necessary soft skill competencies of a biostatistician"

# Inappropriate Analysis and Reporting

10 18 statistical methods

Wang MQ, Yan AF, Katz RV.

"Researchers frequently make inappropriate requests of their biostatistical consultants regarding the analysis and reporting of their data"

__Researcher Requests for Inappropriate Analysis and Reporting: A U.S. Survey of Consulting Biostatisticians.__Ann Intern Med. 2018;169:554-558. 10.7326/M18-1230"Researchers frequently make inappropriate requests of their biostatistical consultants regarding the analysis and reporting of their data"

# Predictive Value of New Measurements

"When the outcome variable Y is continuous, there are only three measures of added value that are commonly used: increase in R2, decrease in mean squared prediction error, and decrease in mean absolute prediction error. Why have so many measures been invented when Y is binary or censored?" Blog post

# Lack of group-to-individual generalizability

Fisher AJ, Medaglia JD, Jeronimus BF. Lack of group-to-individual generalizability is a threat to human subjects research. Proceedings of the National Academy of Sciences. 2018;115:E6106-E6115.

"This suggests that literatures in social and medical sciences may overestimate the accuracy of aggregated statistical estimates".

I recently gave a talk on this *big* problem.

"This suggests that literatures in social and medical sciences may overestimate the accuracy of aggregated statistical estimates".

I recently gave a talk on this *big* problem.

# Perils of the NNT

07 18 statistical methods NNT

# Bootstrap Inference using boottest

06 18 statistical methods bootstrap

Roodman D, MacKinnon JG, Nielsen MO, Webb MD. Fast and Wild: Bootstrap Inference in Stata Using boottest. Queen’s Economics Department. Working Paper No. 1406

"We review the main ideas of the wild cluster bootstrap, oﬀer tips for use, explain why it is particularly amenable to computational optimization, state the syntax of boottest, artest, scoretest, and waldtest, and present several empirical examples for illustration"

"We review the main ideas of the wild cluster bootstrap, oﬀer tips for use, explain why it is particularly amenable to computational optimization, state the syntax of boottest, artest, scoretest, and waldtest, and present several empirical examples for illustration"

# Study size based on precision

Rothman KJ, Greenland S. Planning Study Size Based on Precision Rather than Power.

"Focusing on precision may help to avert some of the problems related to reliance on statistical hypothesis testing"

*Epidemiology*. 2018"Focusing on precision may help to avert some of the problems related to reliance on statistical hypothesis testing"

# Big Data and Predictive Analytics

Shah ND, Steyerberg EW, Kent DM. Big Data and Predictive Analytics: Recalibrating Expectations.

On the importance of model calibration.

*JAMA*. 2018On the importance of model calibration.

# Linear regression and the normality assumption

Schmidt AF, Finan C. Linear regression and the normality assumption.

Outcome transformations may change the target estimate and hence bias results.

*Journal of Clinical Epidemiology*. 2018;98:146-151.Outcome transformations may change the target estimate and hence bias results.

# Initial data analysis

Huebner M, le Cessie S, Schmidt C, Vach W. A Contemporary Conceptual Framework for Initial Data Analysis.

Proposal of a framework for IDA from the IDA group of the STRATOS initiative.

*Observational Studies*. 2018;4:171-192.Proposal of a framework for IDA from the IDA group of the STRATOS initiative.

# Case-control matching

Mansournia MA, Jewell NP, Greenland S. Case-control matching: effects, misconceptions, and recommendations.

On the limitations of matching in case-control studies.

*Eur J Epidemiol*. 2018;33:5-14.On the limitations of matching in case-control studies.

# Abandon statistical significance

Amrhein V, Trafimow D, Greenland S. Abandon statistical inference.

"We reccomend that we should use, communicate, and teach our statistical methods as being descriptive of logical relations between assumptions and data, rather than allowing specified generalized inferences about universal populations".

*Peer J Preprints*. 2018"We reccomend that we should use, communicate, and teach our statistical methods as being descriptive of logical relations between assumptions and data, rather than allowing specified generalized inferences about universal populations".