Didier Brassard
LightNews LightNews
didierbrassard.bsky.social
Didier Brassard
@didierbrassard.bsky.social
21 followers 54 following 20 posts
Postdoctoral research fellow.
💻 Nutritional epidemiology, aging, dietary assessment and causal inference (at least trying)
📍 Université de Montréal
Posts Replies Media Videos
Pinned
didierbrassard.bsky.social
Didier Brassard @didierbrassard.bsky.social · Nov 20
I publish blogs about data analysis and visualization, measurement error and nutrition research. Come check!

didierbrassard.github.io/year-archive/
Reposted by Didier Brassard
ianhussey.mmmdata.io
My article "Data is not available upon request" was published in Meta-Psychology. Very happy to see this out!
open.lnu.se/index.php/me...
LnuOpen | Meta-Psychology
open.lnu.se
October 4, 2025 at 12:54 PM
Reposted by Didier Brassard
chelseaparlett.bsky.social
It’s not the method that makes you causal it’s the assumptions
October 4, 2025 at 3:52 AM
Reposted by Didier Brassard
djnavarro.net
A few words on survey weights, why I'm embarrassed to have forgotten to take them into account in the past, and how I got lucky because I personally didn't get burned. Not a mistake I intend to repeat in the future

blog.djnavarro.net/posts/2025-0...
Some notes on survey weights – Notes from a data witch
An area of statistics in which the author is not strong, and really needs to up her game
blog.djnavarro.net
September 28, 2025 at 6:16 AM
Reposted by Didier Brassard
pwgtennant.bsky.social
The TARGET reporting guidelines for target trial emulation studies have arrived!

#EpiSky #CausalSky

jamanetwork.com/journals/jam...
TARGET 2025 Statement
This Special Communication introduces the Transparent Reporting of Observational Studies Emulating a Target Trial (TARGET) 2025 guideline, a consensus-based guidance for reporting observational studie...
jamanetwork.com
September 3, 2025 at 4:20 PM
Reposted by Didier Brassard
ryancbriggs.net
It's very human to only double check that a process is working when you get a weird result. It's also very bad practice, because sometimes your "right" result is due to a bad process and you will be misled. Social scientists (economists) do this kind of asymmetric checking.
arxiv.org/pdf/2508.20069
September 2, 2025 at 1:35 PM
Reposted by Didier Brassard
dingdingpeng.the100.ci
Ever stared at a table of regression coefficients & wondered what you're doing with your life?

Very excited to share this gentle introduction to another way of making sense of statistical models (w @vincentab.bsky.social)
Preprint: doi.org/10.31234/osf...
Website: j-rohrer.github.io/marginal-psy...
Models as Prediction Machines: How to Convert Confusing Coefficients into Clear Quantities Abstract Psychological researchers usually make sense of regression models by interpreting coefficient estimates directly. This works well enough for simple linear models, but is more challenging for more complex models with, for example, categorical variables, interactions, non-linearities, and hierarchical structures. Here, we introduce an alternative approach to making sense of statistical models. The central idea is to abstract away from the mechanics of estimation, and to treat models as “counterfactual prediction machines,” which are subsequently queried to estimate quantities and conduct tests that matter substantively. This workflow is model-agnostic; it can be applied in a consistent fashion to draw causal or descriptive inference from a wide range of models. We illustrate how to implement this workflow with the marginaleffects package, which supports over 100 different classes of models in R and Python, and present two worked examples. These examples show how the workflow can be applied across designs (e.g., observational study, randomized experiment) to answer different research questions (e.g., associations, causal effects, effect heterogeneity) while facing various challenges (e.g., controlling for confounders in a flexible manner, modelling ordinal outcomes, and interpreting non-linear models). Figure illustrating model predictions. On the X-axis the predictor, annual gross income in Euro. On the Y-axis the outcome, predicted life satisfaction. A solid line marks the curve of predictions on which individual data points are marked as model-implied outcomes at incomes of interest. Comparing two such predictions gives us a comparison. We can also fit a tangent to the line of predictions, which illustrates the slope at any given point of the curve. A figure illustrating various ways to include age as a predictor in a model. On the x-axis age (predictor), on the y-axis the outcome (model-implied importance of friends, including confidence intervals). Illustrated are 1. age as a categorical predictor, resultings in the predictions bouncing around a lot with wide confidence intervals 2. age as a linear predictor, which forces a straight line through the data points that has a very tight confidence band and 3. age splines, which lies somewhere in between as it smoothly follows the data but has more uncertainty than the straight line.
August 25, 2025 at 11:49 AM
Reposted by Didier Brassard
pwgtennant.bsky.social
I'm reviewing a lot of weak target trial emulation studies these days.

Like a wolf in sheep's clothes, these adopt the language & structure of target trial emulation, but don't apply the necessary care or thought.

Is this the 'doom cycle'? Where every promising new tool gets dragged into the mud.
statsepi.bsky.social Darren Dahly @statsepi.bsky.social · Aug 12
Target trial emulation is an important, useful idea, but we must remain careful to not allow novices to think it's a magical shortcut to causal inference. Nor can we lose sight of the *main* thing that separates randomized from non-randomized studies.
POTENTIAL RISKS OF TARGET TRIAL TERMINOLOGY There are risks that the term target trial emulation (TTE) might mislead readers without methodological training. We believe some readers may assume TTE represents a distinct form of observational study design, rather than a framework to help researchers identify and avoid self-inflicted errors in observational studies. We are particularly concerned that claims to have emulated an RCT will be mistaken for a claim to have successfully emulated an RCT. The latter statement implies that the inference is of a comparable standard to a well-conducted RCT. To be certain, the TTE framework (2), used correctly, will help investigators specify estimands, and sidestep avoidable selection bias, but we are probably more pessimistic than Schwarze et al. (1) about the ability of ‘‘advanced methods’’ to address the threat of other sources of bias. However, the difference in how randomization accounts for confounding vs. ‘‘advanced methods’’ like covariate adjustment, weighting, etc. is a difference in kind, not in degree. This is because randomization is guaranteed to eliminate confounding for some key estimands, and to do so in a manner that predominantly rests on our trust in a potentially verifiable randomization process. In contrast, acting on the effect estimate returned from an NRSI requires us to trust in the decisions and competence of the investigator, and even then, no guarantee is possible, even in relatively simple sets of relationships among a small set of potential covariates. This distinction will be obvious to methodological experts, but they only constitute a small proportion of researchers and research-consumers
August 12, 2025 at 11:03 AM
Reposted by Didier Brassard
emilyriederer.bsky.social
Technical writing is hard bcs "writing is thinking" but we often should tell our story not in the order we worked. Solution? I wrote a quick post on how @quarto.org 's embed shortcodes can reframe technical writing as reproducible evidence curation

www.emilyriederer.com/post/quarto-...

🧵 (1/n)
How Quarto embed fixes data science storytelling | Emily Riederer
Literate programming excels at capturing our stream of conscience. Our stream of conscience does not excel at explaining the impact of our work. Notebooks enable some of data scientists’ worst tendenc...
www.emilyriederer.com
July 27, 2025 at 1:14 PM
Reposted by Didier Brassard
kcklatt.bsky.social
It's exceedingly hard to argue that the administration & MAHA are committed to improving nutrition when they're simultaneously cutting everything from SNAP-Ed to innovative community nutrition work - cements the perception that food dyes are public health theatre.
www.healthbeat.org/newyork/2025...
Q&A: Nutrition expert discusses pioneering program terminated by USDA
Here’s a Q&A with a nutrition expert who created an after-school program for NYC middle-schoolers who take on adult responsibilities, like meal preparation. This spring, the USDA terminated it.
www.healthbeat.org
July 25, 2025 at 3:39 PM
Reposted by Didier Brassard
olegolt.bsky.social
Serious concerns about a new cortical biomarker for pain sensitivity

jamanetwork.com/journals/jam...

We (with @tspisak.bsky.social, @christianbuchel.bsky.social) published a commentary on Chowdhury, Bi et al. (2025, JAMA Neurology) raising serious concerns about their reported results.

👇 1/13
Concern About Predictive Performance of a Pain Sensitivity Biomarker
To the Editor Chowdhury et al1 evaluated a biomarker for pain sensitivity, combining peak alpha frequency and corticomotor excitability. The authors report outstanding performance (validation set area...
jamanetwork.com
July 22, 2025 at 3:24 PM
Reposted by Didier Brassard
bmj.com
This article provides an overview of the current state of handling continuous variables in healthcare research.

It discusses the potential limitations of assuming a linear relationship between independent and dependent variables
www.bmj.com/content/390/...
Linear predictor plot for three modelling approaches to analyse continuous variables in a case study of cerebrospinal fluid glucose and acute bacterial meningitis
July 21, 2025 at 9:02 AM
Reposted by Didier Brassard
ajordannafa.com
Statistics/Causal Inference folks, what are your favorite papers on why VIF is bad?
July 15, 2025 at 11:15 PM
Reposted by Didier Brassard
dingdingpeng.the100.ci
At this point, I might as well --
Here's an infographic showing different ways to include age as a predictor. The top shows two extremes, just as a plain old numerical predictor (imposes linear trajectory) vs. categorical predictor (imposes nothing whatsoever). And then three solutions in between!
Infographic illustrating different ways to model age. First panel shows two "extreme" cases; including age as a linear numerical predictor (df = 1) or including age as a categorical predictor (df = number of years of age minus 1). Second panel shows an intermediate solution in which age is categorized into broader bins (df = number of categories minus 1, here 5 - 1 = 4). Third panel shows an intermediate solution in which age is included with a polynomial (df = degrees of freedom of the polynomial, here 4). Fourth panel shows an intermediate solution in which age is modeled with the help of splines (df = degrees of freedom of the splines, here 4).
July 16, 2025 at 12:33 PM
didierbrassard.bsky.social
Fantastic article about modeling of continuous variables! ⬇️
Shameless plug: I wrote a blog about restricted cubic spline applied to nutrition and health data: didierbrassard.github.io/posts/2023/0...
July 16, 2025 at 11:36 PM
Reposted by Didier Brassard
tomayates.bsky.social
There are going to be an awful lot of analyses that need redone when people finally accept that collider bias is a major issue in unweighted analyses of UKBB data
Reweighting UK Biobank corrects for pervasive selection bias due to volunteering
AbstractBackground. Biobanks typically rely on volunteer-based sampling. This results in large samples (power) at the cost of representativeness (bias). Th
share.google
July 15, 2025 at 11:51 AM
Reposted by Didier Brassard
rmcelreath.bsky.social
How can we reform science? I have some ideas. But I am not sure you’ll like them, because they don’t promise much. elevanth.org/blog/2025/07...
Which Kind of Science Reform
What hope is there for science reform, if we can't agree on what to reform? Right now, principles are more important than practices.
elevanth.org
July 9, 2025 at 1:40 PM
didierbrassard.bsky.social
📄 New preprint! 
The main project of my postdoctoral work is now available (not peer-reviewed yet): Estimating the effect of adhering to #CanadaFoodGuide 2019 recommendations in older adults: a target trial emulation

🔗 www.medrxiv.org/content/10.1...
Estimating the effect of adhering to Canada’s Food Guide 2019 recommendations in older adults: a target trial emulation
Background The 2019 Canada’s Food Guide (CFG) provides universal recommendations to individuals aged 2 years or older. The extent to which these recommendations positively influence key health outcome...
www.medrxiv.org
July 8, 2025 at 2:39 PM
didierbrassard.bsky.social
I agree, the challenges to self-correcting science are real.
Tried to publish a “letter to the editor” which was rejected in the end. The editor mentioned the topic of the letter wouldn’t be of interest to readers! Wouldn’t readers also be interested to learn about flaws of a published study?
ianhussey.mmmdata.io Ian Hussey @ianhussey.mmmdata.io · Jul 5
I am often told that public critique of published articles must also solve the issues found. I think this frequently enforced requirement hinders scientific self-correction.

Blog post:

mmmdata.io/posts/2025/0...
July 5, 2025 at 2:33 PM
Reposted by Didier Brassard
mcxfrank.bsky.social
Experimentology is out today!!! A group of us wrote a free online textbook for experimental methods, available at experimentology.io - the idea was to integrate open science into all aspects of the experimental workflow from planning to design, analysis, and writing.
Experimentology cover: title and curves for distributions.
July 1, 2025 at 6:26 PM
Reposted by Didier Brassard
oberman.bsky.social
What is ‘Open Science’? Or, how is Open Science operationalised in survey research? Come talk to me at the #metascience2025 poster session!
Academic poster. Available as PDF via https://hanneoberman.github.io/presentations/2025/METASCIENCE/Metascience.pdf
July 1, 2025 at 8:21 AM
Reposted by Didier Brassard
jopieboy.bsky.social
Update #2, RETRACTED: 15 months after we (w @ollefolke.bsky.social and @johannarickne.bsky.social ) submitted the initial comment to the Journal, we've noticed the paper was ultimately retracted. Retraction note here: link.springer.com/article/10.1...
July 1, 2025 at 6:21 AM
Reposted by Didier Brassard
collinberke.bsky.social
Friendly #rstats package development reminder to my future self: `@inheritParams` is a nice {roxygen2} feature to help you stay DRY when writing function documentation. More here: r-pkgs.org/man.html#inh...

📦 : roxygen2.r-lib.org/index.html
16  Function documentation – R Packages (2e)
Learn how to create a package, the fundamental unit of shareable, reusable, and reproducible R code.
r-pkgs.org
June 27, 2025 at 8:27 PM
Reposted by Didier Brassard
cjwich.bsky.social
basically every research paper I’ve written, minus step 6.
June 28, 2025 at 11:46 AM
Reposted by Didier Brassard
cruwelli.bsky.social
TIL that impact factors are even weirder than I thought! Negotiated with Clarivate?! What? I didn’t think the scientific publishing system could get any stranger!
rmcelreath.bsky.social Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · Jun 19
While *ideal* impact factor is well known to be a bad proxy of article quality, it is less well known that *actual* impact factor is negotiated and gamed by journals. You think Nature earned that impact factor? You think that's air you're breathing? (see e.g. journals.plos.org/plosmedicine...)
June 24, 2025 at 9:13 AM
Reposted by Didier Brassard
miquai.bsky.social
Me: "What did you learn at journal club today?"

Intern: "That one day I'm going to publish a paper, and a bunch of people are going to sit around a table and rip it apart."
June 24, 2025 at 6:38 PM