yellow brick road to stats heaven

~ a loose collection of statistical and quantitative research material for fun and enrichment ~

by roland b. stark

recently published in The American Statistician

"Does Home Health Care increase the probability of 30-day hospital readmissions? Interpreting coefficient sign reversals, or their absence, in binary logistic regression analysis." A study of statistical methods and of the surprising link between home health care and hospital readmission. By Alecos Papadopoulos and Roland B. Stark.

Non-Technical Summary


Data for 30-day readmission rates in American hospitals often show that patients that receive Home Health Care (HHC) have a higher probability of being readmitted to hospital than those that did not receive such services, but it is expected that when control variables are included in a regression we will obtain a “sign reversal” of the treatment effect. We map the real-world situation to the binary logistic regression model, and we construct a counterfactual probability metric that leads to necessary and sufficient conditions for the sign reversal to occur, conditions that show that logistic regression is an appropriate tool for this research purpose. This metric also permits us to obtain evidence related to the criteria used to assign HHC treatment. We examine seven data samples from different USA hospitals for the period 2011- 2017. We find that in all cases the provision of HHC increased the probability of readmission of the treated patients. This casts doubt on the appropriateness of the 30-day readmission rate as an indicator of hospital performance and a criterion for hospital reimbursement, as it is currently used for Medicare patients.

analyzing not just for correlation but for causation

A free, non-technical guide. You've heard that predicting an outcome (finding indicators) is not the same as explaining it (establishing cause and effect). But how to interpret these sorts of research findings -- or how to get started with your own explanatory analysis? You can start here. Guaranteed to help "or your money back."


Statistics & Pooh Bear

Where I praise the invoking of Pooh Bear and pan some research fit for Captain Obvious.

what a Batman Graph can show about wealth and poverty in America

A short piece on understanding socioeconomics by way of scatterplots.

'Batman' Graph of US Wealth and Poverty

the most insidious statistical mistakes

Commentary on interpretation errors that are terribly difficult to avoid, involving conventional significance testing.

learning from research questions

Sometimes it's eye-opening--and encouraging--to see the range of questions one might address using statistics.

applying behavioral economics to the business of higher education

This article contrasts two approaches to uncovering the reasons behind enrollment, application, and other key decisions made by students and their parents. You may be surprised at how much more effective some innovative "derived importance" techniques prove as compared to traditional "stated importance" methods. Abstract and link to paper.

of fishbones and college enrollments

A quick summary of a brainstorming tool. It's a helpful way to start a process of investigating the factors that could drive decisions.

Fishbone Diagram

statistical detection of cheating

This paper of 30-odd pages explores the ingenious yet usually flawed methods that have been devised for statistical detection of copying on multiple-choice exams. The methods of Angoff, Crawford, Belleza & Belleza, and Kling are examined in depth, with some attention also given to the work of Frary, Wollack and Cohen, and others. Abstract and link to paper.

Comparison of Cheating Methods

illustrating partial correlation vs. interaction

It's easy to confuse these two. Unable to find a graphical representation that explained the difference to my satisfaction, I created the set of charts and brief commentary available here.

anova / ancova decision guide

Think it's easy to sort out the differences between ANOVA and ANCOVA? This look at empirical and conceptual considerations may make you think again. It includes a flow chart, glossary, and commentary and is available in Word.

Anova/Ancova Flow Chart

factor analysis decision guide - *updated* nov. 2021

"How can I find patterns in my data?" This ten-page guide introduces the sometimes wondrous method of factor analysis and offers guidance on its classic decision points: on data applicability, method of extraction, number of factors, rotation method, how to display results, and more. Available in pdf.

what's wrong with value-added models in education

Eleven reasons why teachers shouldn't be judged based on student test scores.

my favorite books on statistics

In case you're not laughing at the very idea, here are some recommendations.

recommended sites


Periodic Table of Visualization Methods

Ralph Lengler and Martin J. Eppler's Periodic Table of Visualization Methods is a gorgeous, ingenious, highly informative single-page display of about a hundred types of charts and diagrams for visualizing data, concepts, strategies, and more. Hovering your cursor over any "element" in this table will bring up a colorful and instructive example.

art from data: the R graph gallery

The R Graph Gallery of Yan Holtz

An excellent resource for data graphics you can create in the R software, including very useful code examples. I was happy to be able to contribute in small ways to this repository created by Yan Holtz.

gallery of data visualization

All of Inflation's Little Parts

Michael Friendly of Toronto's York University presents a nice smorgasbord of visual confections. Some good lessons and also some good laughs.

Q & A on statistics: the crossvalidated community

crossvalidated site

A fine place to ask targeted questions on statistics or data visualization, or to browse through searchable questions, answers, and commentary.

"every single cognitive bias"

A fun, intricate infographic for fans of psychology and behavioral economics research. From the people at

I welcome any comments you would like to make, publicly or privately.

my statistical and research consulting

copyright 2007-2021 by roland b. stark.

my statistical and research consulting