As I am working on a major update to `delicatessen`, a catch-up #MEstimatorMonday (50/52) about a feature I am going to add
Currently, I do not automatically handle non-independent data. v4.0 is likely to change this
github.com/pzivich/Deli...
Currently, I do not automatically handle non-independent data. v4.0 is likely to change this
github.com/pzivich/Deli...
Handling Non-Independent Data · Issue #81 · pzivich/Delicatessen
Is your feature request related to a problem? Please describe. Right now (v3.2), there is not direct support for non-independent observations. Essentially, all one needs to do (with an independence...
github.com
December 19, 2025 at 2:27 PM
As I am working on a major update to `delicatessen`, a catch-up #MEstimatorMonday (50/52) about a feature I am going to add
Currently, I do not automatically handle non-independent data. v4.0 is likely to change this
github.com/pzivich/Deli...
Currently, I do not automatically handle non-independent data. v4.0 is likely to change this
github.com/pzivich/Deli...
This #MEstimatorMonday (49/52) let's look at beta regression. Beta regression is another regression model but for the case where the outcome Y is bounded (0,1)
So, it is good for modeling proportions as outcomes. While we can also use fractional logistic regression, this is another option
So, it is good for modeling proportions as outcomes. While we can also use fractional logistic regression, this is another option
December 15, 2025 at 7:16 PM
This #MEstimatorMonday (49/52) let's look at beta regression. Beta regression is another regression model but for the case where the outcome Y is bounded (0,1)
So, it is good for modeling proportions as outcomes. While we can also use fractional logistic regression, this is another option
So, it is good for modeling proportions as outcomes. While we can also use fractional logistic regression, this is another option
Since it's not actually Monday, I feel less guilty about this #MEstimatorMonday (48/52) which tells you how to do mediation analysis with estimating equations
Let consider the effect of action A on outcome Y through mediator M
Let consider the effect of action A on outcome Y through mediator M
December 9, 2025 at 6:59 PM
Since it's not actually Monday, I feel less guilty about this #MEstimatorMonday (48/52) which tells you how to do mediation analysis with estimating equations
Let consider the effect of action A on outcome Y through mediator M
Let consider the effect of action A on outcome Y through mediator M
It's finals week, so let's space out on this #MEstimatorMonday (47/52). I've read a few astronomy books lately, so I was inspired to find an astrostatistics example with `delicatessen`. The following example models quasars
deli.readthedocs.io/en/latest/Ex...
deli.readthedocs.io/en/latest/Ex...
Astrostatistics - Quasars — Delicatessen 3.2 documentation
deli.readthedocs.io
December 8, 2025 at 3:17 PM
It's finals week, so let's space out on this #MEstimatorMonday (47/52). I've read a few astronomy books lately, so I was inspired to find an astrostatistics example with `delicatessen`. The following example models quasars
deli.readthedocs.io/en/latest/Ex...
deli.readthedocs.io/en/latest/Ex...
This not-Monday #MEstimatorMonday (46/52) is to commemorate me finally getting around to finding an example of IV analysis from the econometric literature
In other words, I found Wooldridge's book and the Mroz 1987 paper
deli.readthedocs.io/en/latest/Ex...
In other words, I found Wooldridge's book and the Mroz 1987 paper
deli.readthedocs.io/en/latest/Ex...
Mroz 1987: Regression and Instrumental Variables — Delicatessen 3.2 documentation
deli.readthedocs.io
December 3, 2025 at 10:05 PM
This not-Monday #MEstimatorMonday (46/52) is to commemorate me finally getting around to finding an example of IV analysis from the econometric literature
In other words, I found Wooldridge's book and the Mroz 1987 paper
deli.readthedocs.io/en/latest/Ex...
In other words, I found Wooldridge's book and the Mroz 1987 paper
deli.readthedocs.io/en/latest/Ex...
For the next catch-up-but-not-Monday #MEstimatorMonday (45/52) I am going to sketch out an argument I've been thinking about
M-estimation, or the act of stacking estimating functions, has great potential to improve the statistical rigor of scientific research
M-estimation, or the act of stacking estimating functions, has great potential to improve the statistical rigor of scientific research
December 2, 2025 at 2:51 PM
For the next catch-up-but-not-Monday #MEstimatorMonday (45/52) I am going to sketch out an argument I've been thinking about
M-estimation, or the act of stacking estimating functions, has great potential to improve the statistical rigor of scientific research
M-estimation, or the act of stacking estimating functions, has great potential to improve the statistical rigor of scientific research
For my final #MEstimatorMonday (44/52) of today, let's look at another confidence bands example
This example ties in with #WorldAIDSDay as it relies on data from the AIDS Clinical Trial Group (ACTG) 175, comparing 2-drug versus 1-drug ART for prevention of disease progression among persons with HIV
This example ties in with #WorldAIDSDay as it relies on data from the AIDS Clinical Trial Group (ACTG) 175, comparing 2-drug versus 1-drug ART for prevention of disease progression among persons with HIV
December 1, 2025 at 7:15 PM
For my final #MEstimatorMonday (44/52) of today, let's look at another confidence bands example
This example ties in with #WorldAIDSDay as it relies on data from the AIDS Clinical Trial Group (ACTG) 175, comparing 2-drug versus 1-drug ART for prevention of disease progression among persons with HIV
This example ties in with #WorldAIDSDay as it relies on data from the AIDS Clinical Trial Group (ACTG) 175, comparing 2-drug versus 1-drug ART for prevention of disease progression among persons with HIV
For the 2nd #MEstimatorMonday (43/52) of today, let's look at those confidence bands a bit more.
For this, I am going to reach outside my subject area and pontificate about cheminformatics
For this, I am going to reach outside my subject area and pontificate about cheminformatics
December 1, 2025 at 4:19 PM
For the 2nd #MEstimatorMonday (43/52) of today, let's look at those confidence bands a bit more.
For this, I am going to reach outside my subject area and pontificate about cheminformatics
For this, I am going to reach outside my subject area and pontificate about cheminformatics
Turns out teaching really interferes with me staying on top of #MEstimatorMonday (42/52), so I need to do some catch up again
To start, I am going to note a new release of `delicatessen` v3.2
The major addition of this release is confidence bands
To start, I am going to note a new release of `delicatessen` v3.2
The major addition of this release is confidence bands
December 1, 2025 at 1:38 PM
Turns out teaching really interferes with me staying on top of #MEstimatorMonday (42/52), so I need to do some catch up again
To start, I am going to note a new release of `delicatessen` v3.2
The major addition of this release is confidence bands
To start, I am going to note a new release of `delicatessen` v3.2
The major addition of this release is confidence bands
To close my catch-up of #MEstimatorMonday (41/52), let's talk about how I use estimating equations with synthesis models
October 17, 2025 at 2:34 PM
To close my catch-up of #MEstimatorMonday (41/52), let's talk about how I use estimating equations with synthesis models
Give it up for day 4 of my #MEstimatorMonday catch-up week (40/52). Let's go over Bespoke IVs
Like proximal causal inference, this is another way to address unmeasured confounding. For this, let's go over per-protocol effects in trials
Like proximal causal inference, this is another way to address unmeasured confounding. For this, let's go over per-protocol effects in trials
a cartoon character from spongebob squarepants is holding a crab in his hand and screaming .
ALT: a cartoon character from spongebob squarepants is holding a crab in his hand and screaming .
media.tenor.com
October 17, 2025 at 12:20 AM
Give it up for day 4 of my #MEstimatorMonday catch-up week (40/52). Let's go over Bespoke IVs
Like proximal causal inference, this is another way to address unmeasured confounding. For this, let's go over per-protocol effects in trials
Like proximal causal inference, this is another way to address unmeasured confounding. For this, let's go over per-protocol effects in trials
For the 3rd day of my #MEstimatorMonday catch-up week (39/52), let's talk proximal causal inference
We can think about proximal causal inference as an extension of the standard identification assumptions to allow for more rich data structures. Specifically, we can account for unmeasured confounding
We can think about proximal causal inference as an extension of the standard identification assumptions to allow for more rich data structures. Specifically, we can account for unmeasured confounding
October 15, 2025 at 6:38 PM
For the 3rd day of my #MEstimatorMonday catch-up week (39/52), let's talk proximal causal inference
We can think about proximal causal inference as an extension of the standard identification assumptions to allow for more rich data structures. Specifically, we can account for unmeasured confounding
We can think about proximal causal inference as an extension of the standard identification assumptions to allow for more rich data structures. Specifically, we can account for unmeasured confounding
For the 2nd day of the #MEstimatorMonday week (38/52), I'm going to workshop a notion I've been thinking about regarding the idea of stacking estimating functions
October 14, 2025 at 6:25 PM
For the 2nd day of the #MEstimatorMonday week (38/52), I'm going to workshop a notion I've been thinking about regarding the idea of stacking estimating functions
I have been bad about keeping up with #MEstimatorMonday so this will be a week of M-estimator (so 37 to 41/52)
To start, let's talk computational aspects. As you might recall from the weeks on the sandwich variance, we need to compute derivatives. One option is automatic differentiation (autodiff)
To start, let's talk computational aspects. As you might recall from the weeks on the sandwich variance, we need to compute derivatives. One option is automatic differentiation (autodiff)
October 13, 2025 at 1:21 PM
I have been bad about keeping up with #MEstimatorMonday so this will be a week of M-estimator (so 37 to 41/52)
To start, let's talk computational aspects. As you might recall from the weeks on the sandwich variance, we need to compute derivatives. One option is automatic differentiation (autodiff)
To start, let's talk computational aspects. As you might recall from the weeks on the sandwich variance, we need to compute derivatives. One option is automatic differentiation (autodiff)
I'm going to do a quick #MEstimatorMonday (36/52) to show how the pooled logit EE can be used to construct inverse probability of censoring weights (IPCW)
This is based on the algorithm in my pre-print, but needs a few tweaks
arxiv.org/abs/2504.13291
This is based on the algorithm in my pre-print, but needs a few tweaks
arxiv.org/abs/2504.13291
Estimating equations for survival analysis with pooled logistic regression
Pooled logistic regression models are commonly applied in survival analysis. However, the standard implementation can be computationally demanding, which is further exacerbated when using the nonparam...
arxiv.org
September 9, 2025 at 12:29 PM
I'm going to do a quick #MEstimatorMonday (36/52) to show how the pooled logit EE can be used to construct inverse probability of censoring weights (IPCW)
This is based on the algorithm in my pre-print, but needs a few tweaks
arxiv.org/abs/2504.13291
This is based on the algorithm in my pre-print, but needs a few tweaks
arxiv.org/abs/2504.13291
Something a bit different for #MEstimatorMonday (35/52)
I'm going to take advantage of microblogging as a format and live code a new estimating equation
I'm going to take advantage of microblogging as a format and live code a new estimating equation
September 1, 2025 at 12:47 PM
Something a bit different for #MEstimatorMonday (35/52)
I'm going to take advantage of microblogging as a format and live code a new estimating equation
I'm going to take advantage of microblogging as a format and live code a new estimating equation
I hope you're wearing your hardhat because we are building bridges this #MEstimatorMonday (34/52)
More specifically, we are going to bridge different trials together to address a single question
pubmed.ncbi.nlm.nih.gov/38110289/
More specifically, we are going to bridge different trials together to address a single question
pubmed.ncbi.nlm.nih.gov/38110289/
Fusing trial data for treatment comparisons: Single vs multi-span bridging - PubMed
While randomized controlled trials (RCTs) are critical for establishing the efficacy of new therapies, there are limitations regarding what comparisons can be made directly from trial data. RCTs are l...
pubmed.ncbi.nlm.nih.gov
August 25, 2025 at 9:30 PM
I hope you're wearing your hardhat because we are building bridges this #MEstimatorMonday (34/52)
More specifically, we are going to bridge different trials together to address a single question
pubmed.ncbi.nlm.nih.gov/38110289/
More specifically, we are going to bridge different trials together to address a single question
pubmed.ncbi.nlm.nih.gov/38110289/
It's the first day of classes here for the semester, so for today's #MEstimatorMonday (33/52), let's do some math
So far we've seen how the M-estimation framework is useful computationally, so let's step back and see how it is also useful for statistical proofs
So far we've seen how the M-estimation framework is useful computationally, so let's step back and see how it is also useful for statistical proofs
August 18, 2025 at 9:42 PM
It's the first day of classes here for the semester, so for today's #MEstimatorMonday (33/52), let's do some math
So far we've seen how the M-estimation framework is useful computationally, so let's step back and see how it is also useful for statistical proofs
So far we've seen how the M-estimation framework is useful computationally, so let's step back and see how it is also useful for statistical proofs
It's Wonday (yes I know its late), so another happy #MEstimatorMonday (32/52)
Let's talk about something M-estimation has a bit of difficulty with: the median
Let's talk about something M-estimation has a bit of difficulty with: the median
a spongebob squarepants character has a purple skirt with the letter m on it
ALT: a spongebob squarepants character has a purple skirt with the letter m on it
media.tenor.com
August 13, 2025 at 11:35 AM
It's Wonday (yes I know its late), so another happy #MEstimatorMonday (32/52)
Let's talk about something M-estimation has a bit of difficulty with: the median
Let's talk about something M-estimation has a bit of difficulty with: the median
I'm at #JSM2025 (say hi if you are too), so a shorter #MEstimatorMonday (31/52)
As setup for later weeks, let's talk about data fusion and why M-estimators are a useful way to build fusion estimators. Today is based on the following paper
pubmed.ncbi.nlm.nih.gov/35388406/
As setup for later weeks, let's talk about data fusion and why M-estimators are a useful way to build fusion estimators. Today is based on the following paper
pubmed.ncbi.nlm.nih.gov/35388406/
Illustration of 2 Fusion Designs and Estimators - PubMed
"Fusion" study designs combine data from different sources to answer questions that could not be answered (as well) by subsets of the data. Studies that augment main study data with validation data, a...
pubmed.ncbi.nlm.nih.gov
August 4, 2025 at 9:21 PM
I'm at #JSM2025 (say hi if you are too), so a shorter #MEstimatorMonday (31/52)
As setup for later weeks, let's talk about data fusion and why M-estimators are a useful way to build fusion estimators. Today is based on the following paper
pubmed.ncbi.nlm.nih.gov/35388406/
As setup for later weeks, let's talk about data fusion and why M-estimators are a useful way to build fusion estimators. Today is based on the following paper
pubmed.ncbi.nlm.nih.gov/35388406/
Today for #MEstimatorMonday (30/52), I'm going to workshop something I have been trying to clarify my own thinking about: confidence bands
So, if something isn't clear, I'm incorrect, or there is something missing; passive-aggressively quote the offending skeet
So, if something isn't clear, I'm incorrect, or there is something missing; passive-aggressively quote the offending skeet
July 28, 2025 at 9:21 PM
Today for #MEstimatorMonday (30/52), I'm going to workshop something I have been trying to clarify my own thinking about: confidence bands
So, if something isn't clear, I'm incorrect, or there is something missing; passive-aggressively quote the offending skeet
So, if something isn't clear, I'm incorrect, or there is something missing; passive-aggressively quote the offending skeet
On this #MEstimatorMonday (29/52) let's see how we can implement sensitivity analyses for estimating the mean with missing data. This discussion is based on the following paper
pmc.ncbi.nlm.nih.gov/articles/PMC...
pmc.ncbi.nlm.nih.gov/articles/PMC...
Sensitivity Analyses for Means or Proportions with Missing Outcome Data
We describe an approach to sensitivity analysis introduced by Robins and colleagues, for the setting where the outcome is missing for some observations. This flexible approach focuses on the relations...
pmc.ncbi.nlm.nih.gov
July 21, 2025 at 5:31 PM
On this #MEstimatorMonday (29/52) let's see how we can implement sensitivity analyses for estimating the mean with missing data. This discussion is based on the following paper
pmc.ncbi.nlm.nih.gov/articles/PMC...
pmc.ncbi.nlm.nih.gov/articles/PMC...
This is the week that I get @malcolmbarrett.malco.io to quit social media because #MEstimatorMonday (27/52) is on g-estimation and IV
if you could kindly let me know when that week is so I could sign out of social media that would be great.
July 7, 2025 at 1:16 PM
This is the week that I get @malcolmbarrett.malco.io to quit social media because #MEstimatorMonday (27/52) is on g-estimation and IV
The temperatures are rising, but so are the doses for this #MEstimatorMonday (26/52). I'm going to return to an area a bit outside of my research area: dose-response modeling from the pharmacokinetics perspective (criticisms welcome)
bsky.app/profile/paus...
bsky.app/profile/paus...
To show how widely #MEstimatorMonday (10/52) goes, this week I will talk about pharmacokinetic models (something I know less about and don't use in my own work).
Specifically, I am going to focus on the use of dose-response models. We have some different dosages (D) and some continuous response (R)
Specifically, I am going to focus on the use of dose-response models. We have some different dosages (D) and some continuous response (R)
June 30, 2025 at 6:49 PM
The temperatures are rising, but so are the doses for this #MEstimatorMonday (26/52). I'm going to return to an area a bit outside of my research area: dose-response modeling from the pharmacokinetics perspective (criticisms welcome)
bsky.app/profile/paus...
bsky.app/profile/paus...
It's a musical #MEstimatorMonday (25/52) since we will focus on *instrumental* variables (IV)
IV analysis is nice because we can get around unmeasured confounding (at the cost of a more easily interpretable parameter)
IV analysis is nice because we can get around unmeasured confounding (at the cost of a more easily interpretable parameter)
June 24, 2025 at 12:39 AM
It's a musical #MEstimatorMonday (25/52) since we will focus on *instrumental* variables (IV)
IV analysis is nice because we can get around unmeasured confounding (at the cost of a more easily interpretable parameter)
IV analysis is nice because we can get around unmeasured confounding (at the cost of a more easily interpretable parameter)