All pcurve tests are just a simple sum of transformed pvalues. There is a fundamental disconnect between the null hypotheses being tested by p-curve and the claims being made.
August 9, 2025 at 9:18 PM
All pcurve tests are just a simple sum of transformed pvalues. There is a fundamental disconnect between the null hypotheses being tested by p-curve and the claims being made.
If you are teaching an introduction to statistics course this spring and are looking for a fun way to illustrate p-values to your students, I've found this to be effective for undergrad crowds. #statistics #education #pvalues
December 27, 2024 at 1:10 AM
If you are teaching an introduction to statistics course this spring and are looking for a fun way to illustrate p-values to your students, I've found this to be effective for undergrad crowds. #statistics #education #pvalues
Lots of people get radicalized by pvalues of 0.051 or 0.11.
November 27, 2024 at 10:42 AM
Lots of people get radicalized by pvalues of 0.051 or 0.11.
tbh i am quite proud of myself for (i think) being able to understand pvalues in a way that i think really pretty much anyone could understand it and with a good chance of sticking
July 30, 2025 at 4:29 AM
tbh i am quite proud of myself for (i think) being able to understand pvalues in a way that i think really pretty much anyone could understand it and with a good chance of sticking
#statstab #583 Statistical methods: Basic concepts, interpretations, and cautions
Thoughts: A bit of history and a bit of theory. Useful overview of p-values.
#pvalues #nhst #estimationstatistics #fisher #neyman #compatibility #history
arxiv.org/pdf/2508.10168?
Thoughts: A bit of history and a bit of theory. Useful overview of p-values.
#pvalues #nhst #estimationstatistics #fisher #neyman #compatibility #history
arxiv.org/pdf/2508.10168?
arxiv.org
July 29, 2026 at 6:47 PM
#statstab #583 Statistical methods: Basic concepts, interpretations, and cautions
Thoughts: A bit of history and a bit of theory. Useful overview of p-values.
#pvalues #nhst #estimationstatistics #fisher #neyman #compatibility #history
arxiv.org/pdf/2508.10168?
Thoughts: A bit of history and a bit of theory. Useful overview of p-values.
#pvalues #nhst #estimationstatistics #fisher #neyman #compatibility #history
arxiv.org/pdf/2508.10168?
What looks great to me is the option to add pvalues with a single call that both perform the test and add the values
tidyplots
Time to say goodbye to ggplot2?😜
"a significant reduction of code complexity" vs ggplot2
cran.r-project.org/web/packages...
Jan Broder Engler bioRxiv 2024
www.biorxiv.org/content/10.1...
Time to say goodbye to ggplot2?😜
"a significant reduction of code complexity" vs ggplot2
cran.r-project.org/web/packages...
Jan Broder Engler bioRxiv 2024
www.biorxiv.org/content/10.1...
November 19, 2024 at 11:38 AM
What looks great to me is the option to add pvalues with a single call that both perform the test and add the values
I'm working on an exploratory analysis of clinical trial data and am trying to see if I can make everyone (stats and clinical coauthors) mad by putting pvalues in Table 1 but nowhere else in the manuscript.
June 24, 2025 at 7:16 AM
I'm working on an exploratory analysis of clinical trial data and am trying to see if I can make everyone (stats and clinical coauthors) mad by putting pvalues in Table 1 but nowhere else in the manuscript.
The basic idea of p-curve rests on the idea that the skew of a set of p-values is informative about whether QRPs are occurring. As we show, the p-curve tests have nothing to do with skew. It is trivial to create left skewed pvalues that p-curve would confidently label as right skewed.
August 9, 2025 at 9:18 PM
The basic idea of p-curve rests on the idea that the skew of a set of p-values is informative about whether QRPs are occurring. As we show, the p-curve tests have nothing to do with skew. It is trivial to create left skewed pvalues that p-curve would confidently label as right skewed.
Now knock your socks off and calculate the pvalues!
December 17, 2025 at 2:21 PM
Now knock your socks off and calculate the pvalues!
someone please stop me from re-writing the pvalues section of the script over and over -_- I just need to call it
November 22, 2025 at 3:43 PM
someone please stop me from re-writing the pvalues section of the script over and over -_- I just need to call it
Aargh. The curse of the underflow p value strikes the MVP study.
In picking the lead snps the authors (inadvertantly) picked the first SNP with an underflow pvalue. This can be up to 500kb from the true lead snp.
All the strong pvalues listed from verma et al in @gwascatalog.bsky.social are suspect
In picking the lead snps the authors (inadvertantly) picked the first SNP with an underflow pvalue. This can be up to 500kb from the true lead snp.
All the strong pvalues listed from verma et al in @gwascatalog.bsky.social are suspect
See now this is what I'm talking about. A massive study (MVP) with 80,000 GWAS associations and the 1,000 most significant are corrupted because the authors didn't provide their actual p-values.
www.ebi.ac.uk/gwas/publica...
Don't do this
www.ebi.ac.uk/gwas/publica...
Don't do this
February 11, 2025 at 2:52 PM
Aargh. The curse of the underflow p value strikes the MVP study.
In picking the lead snps the authors (inadvertantly) picked the first SNP with an underflow pvalue. This can be up to 500kb from the true lead snp.
All the strong pvalues listed from verma et al in @gwascatalog.bsky.social are suspect
In picking the lead snps the authors (inadvertantly) picked the first SNP with an underflow pvalue. This can be up to 500kb from the true lead snp.
All the strong pvalues listed from verma et al in @gwascatalog.bsky.social are suspect
What this means is that a significant result for either test only allows one to claim that “at least one” study (out of the set) doesn’t have the property being considered. Why does this happen? Because pcurve completely ignores the configuration of the pvalues being considered.
August 9, 2025 at 9:18 PM
What this means is that a significant result for either test only allows one to claim that “at least one” study (out of the set) doesn’t have the property being considered. Why does this happen? Because pcurve completely ignores the configuration of the pvalues being considered.
#statstab #597 Distribution of p-values when comparing two groups: an interactive visualization
Thoughts: Teaching about p-values is challenging, so a visual can go a long way.
#education #dataviz #plots #pvalues #nhst #visualization
rpsychologist.com/d3/pdist/
Thoughts: Teaching about p-values is challenging, so a visual can go a long way.
#education #dataviz #plots #pvalues #nhst #visualization
rpsychologist.com/d3/pdist/
August 18, 2026 at 5:17 PM
#statstab #597 Distribution of p-values when comparing two groups: an interactive visualization
Thoughts: Teaching about p-values is challenging, so a visual can go a long way.
#education #dataviz #plots #pvalues #nhst #visualization
rpsychologist.com/d3/pdist/
Thoughts: Teaching about p-values is challenging, so a visual can go a long way.
#education #dataviz #plots #pvalues #nhst #visualization
rpsychologist.com/d3/pdist/
I don’t bother teaching stats without resampling anymore. The use of pvalues is actively harmful in many settings. Bootstrapping and model sensitivity analysis, by contrast, will never not be relevant. Doesn’t matter if your model is an LLM or a simple binary threshold. If it isn’t robust, it sucks.
February 26, 2025 at 7:22 PM
I don’t bother teaching stats without resampling anymore. The use of pvalues is actively harmful in many settings. Bootstrapping and model sensitivity analysis, by contrast, will never not be relevant. Doesn’t matter if your model is an LLM or a simple binary threshold. If it isn’t robust, it sucks.
#statstab #309 The statistical significance filter leads to overoptimistic expectations of replicability
Thoughts: Not sure how many researchers interpret p-values are indexes of replicability, but they shouldn't.
#replication #pvalues #TypeMerror #meta
doi.org/10.1016/j.jm...
Thoughts: Not sure how many researchers interpret p-values are indexes of replicability, but they shouldn't.
#replication #pvalues #TypeMerror #meta
doi.org/10.1016/j.jm...
Redirecting
doi.org
March 27, 2025 at 6:08 PM
#statstab #309 The statistical significance filter leads to overoptimistic expectations of replicability
Thoughts: Not sure how many researchers interpret p-values are indexes of replicability, but they shouldn't.
#replication #pvalues #TypeMerror #meta
doi.org/10.1016/j.jm...
Thoughts: Not sure how many researchers interpret p-values are indexes of replicability, but they shouldn't.
#replication #pvalues #TypeMerror #meta
doi.org/10.1016/j.jm...
#statstab #135 {p-checker} The one-for-all p-value analyzer
Thoughts: Why not try your hand at some tools for detecting publication bias (mileage may vary). Useful teaching demo tools.
#shinyapp #r #rstats #phacking #pvalues #NHST #education #edu
shinyapps.org/apps/p-check...
Thoughts: Why not try your hand at some tools for detecting publication bias (mileage may vary). Useful teaching demo tools.
#shinyapp #r #rstats #phacking #pvalues #NHST #education #edu
shinyapps.org/apps/p-check...
Experience Statistics
shinyapps.org
July 12, 2024 at 3:35 PM
#statstab #135 {p-checker} The one-for-all p-value analyzer
Thoughts: Why not try your hand at some tools for detecting publication bias (mileage may vary). Useful teaching demo tools.
#shinyapp #r #rstats #phacking #pvalues #NHST #education #edu
shinyapps.org/apps/p-check...
Thoughts: Why not try your hand at some tools for detecting publication bias (mileage may vary). Useful teaching demo tools.
#shinyapp #r #rstats #phacking #pvalues #NHST #education #edu
shinyapps.org/apps/p-check...
They already have the pvalues
RFK Jr: "By September we will know what has caused the autism epidemic and we will be able to eliminate those exposures."
April 10, 2025 at 11:15 PM
They already have the pvalues
#statstab #289 The meaning of significance in data testing
Thoughts: Fisherian significance testing =/= Neyman-Pearson statistical hypothesis testing. Many debates on p-values and frequentist stats are due to this confusion.
#pvalues #NHST #frequentist
doi.org/10.3389/fpsy...
Thoughts: Fisherian significance testing =/= Neyman-Pearson statistical hypothesis testing. Many debates on p-values and frequentist stats are due to this confusion.
#pvalues #NHST #frequentist
doi.org/10.3389/fpsy...
Frontiers | The meaning of significance in data testing
Recent developments in psychology (e.g., Woolston, 2015a; Trafimow, 2014; Nuzzo, 2014) are showing apparently reasonable but inherently flawed positions agai...
doi.org
February 27, 2025 at 8:00 PM
#statstab #289 The meaning of significance in data testing
Thoughts: Fisherian significance testing =/= Neyman-Pearson statistical hypothesis testing. Many debates on p-values and frequentist stats are due to this confusion.
#pvalues #NHST #frequentist
doi.org/10.3389/fpsy...
Thoughts: Fisherian significance testing =/= Neyman-Pearson statistical hypothesis testing. Many debates on p-values and frequentist stats are due to this confusion.
#pvalues #NHST #frequentist
doi.org/10.3389/fpsy...
This is unfortunately true. I would say the main reasons are that the subject is hard and deep understanding is not incentivized enough.
But note understanding UMAP is much harder than understanding pvalues.
But note understanding UMAP is much harder than understanding pvalues.
December 23, 2024 at 2:54 PM
This is unfortunately true. I would say the main reasons are that the subject is hard and deep understanding is not incentivized enough.
But note understanding UMAP is much harder than understanding pvalues.
But note understanding UMAP is much harder than understanding pvalues.
I am currently revamping small sample defaults of pyfixest (the goal is to match fixest SEs out of the box). Now some unit tests for Poisson comparing Python vs R fail for derived statistics (tstats, pvalues, CIs) because (I think) of restrictive tolerances.
March 29, 2025 at 10:00 PM
I am currently revamping small sample defaults of pyfixest (the goal is to match fixest SEs out of the box). Now some unit tests for Poisson comparing Python vs R fail for derived statistics (tstats, pvalues, CIs) because (I think) of restrictive tolerances.
#statstab #47 Randomization Tests
Thoughts: More researchers should know about this specific approach to analysing experimental data. But, also of the issues around it. Nice overview.
#randomization #NHST #pvalues #research #statistics
www.uvm.edu/~statdhtx/St...
Thoughts: More researchers should know about this specific approach to analysing experimental data. But, also of the issues around it. Nice overview.
#randomization #NHST #pvalues #research #statistics
www.uvm.edu/~statdhtx/St...
www.uvm.edu
March 12, 2024 at 6:16 PM
#statstab #47 Randomization Tests
Thoughts: More researchers should know about this specific approach to analysing experimental data. But, also of the issues around it. Nice overview.
#randomization #NHST #pvalues #research #statistics
www.uvm.edu/~statdhtx/St...
Thoughts: More researchers should know about this specific approach to analysing experimental data. But, also of the issues around it. Nice overview.
#randomization #NHST #pvalues #research #statistics
www.uvm.edu/~statdhtx/St...
#statstab #518 Before p < 0.05 to Beyond p < 0.05: Using History to
Contextualize p-Values and Significance Testing
Thoughts: An appropriate paper for today.
#pvalues #teaching #history #pedagogy #nhst #fisher
www.tandfonline.com/doi/pdf/10.1...
Contextualize p-Values and Significance Testing
Thoughts: An appropriate paper for today.
#pvalues #teaching #history #pedagogy #nhst #fisher
www.tandfonline.com/doi/pdf/10.1...
www.tandfonline.com
April 1, 2026 at 7:28 PM
ae-pvalues 1.0
A tool to explain anomalies identified by an auto-encoder based on pvalues
Author: Maxime LANVIN
🐛Bug Tracker
🏠Homepage
A tool to explain anomalies identified by an auto-encoder based on pvalues
Author: Maxime LANVIN
🐛Bug Tracker
🏠Homepage
March 31, 2025 at 9:00 AM
ae-pvalues 1.0
A tool to explain anomalies identified by an auto-encoder based on pvalues
Author: Maxime LANVIN
🐛Bug Tracker
🏠Homepage
A tool to explain anomalies identified by an auto-encoder based on pvalues
Author: Maxime LANVIN
🐛Bug Tracker
🏠Homepage
#statstab #298 Replication: Do not trust your p-value, be it small or large
Thoughts: Even under exact replications, a p-values is not very good at predicting the p-value in a future study. p=.05 ~ 50% rep
#pvalues #replication #estimation #metascience
physoc.onlinelibrary.wiley.com/doi/10.1113/...
Thoughts: Even under exact replications, a p-values is not very good at predicting the p-value in a future study. p=.05 ~ 50% rep
#pvalues #replication #estimation #metascience
physoc.onlinelibrary.wiley.com/doi/10.1113/...
March 12, 2025 at 2:10 PM
#statstab #298 Replication: Do not trust your p-value, be it small or large
Thoughts: Even under exact replications, a p-values is not very good at predicting the p-value in a future study. p=.05 ~ 50% rep
#pvalues #replication #estimation #metascience
physoc.onlinelibrary.wiley.com/doi/10.1113/...
Thoughts: Even under exact replications, a p-values is not very good at predicting the p-value in a future study. p=.05 ~ 50% rep
#pvalues #replication #estimation #metascience
physoc.onlinelibrary.wiley.com/doi/10.1113/...