All Questions
Tagged with propensity-scores causality
153 questions
3
votes
2
answers
71
views
Limitations of propensity score matching
While studying propensity score matching, I was struck by the following thought:
When we are running a logistic regression model to estimate $p(Z=1∣X)$ through some form of parametrization and we are ...
0
votes
1
answer
36
views
Modeling approaches for conditional probability distribution, applied to Propensity Score estimation for IPW (causal inference)
I'm trying to understand and ideally implement the Inverse Probability Weighting approach to estimate a causal effect. My ressources so far have been Pearl's Primer and the book "What If?".
...
- 225
4
votes
1
answer
255
views
In a doubly robust learner, do the covariates need to be the same for the outcome model and the propensity model?
In a doubly robust learner/estimator, do we need to use the same feature set X when creating the outcome model and the propensity model? Or could we use a subset of X for the propensity model or even ...
- 41
3
votes
1
answer
120
views
What is the math rationale behind the inverse probability weighting?
Papers say IPTW (inverse probability weighting) is superior to PSM (propensity score matching) because it does not necessarily drop observations, whereas PSM drops those observation not paired.
IPTW ...
- 41
2
votes
1
answer
41
views
Propensity score weighting with post-treatment variables
It's been emphasized to balance pre-treatment variables when doing propensity score weighting (PSW) as balancing post-treatment variables can introduce bias.
I want to ask your insights on the case ...
- 377
1
vote
1
answer
272
views
In propensity score matching, is the estimand to be estimated the ATE or the ATT?
In the propensity score matching literature (Central Role of the Propensity Score by Rubin), the treatment effect estimand is referred to as the "Average Treatment Effect" (ATE). However, in ...
- 4,260
0
votes
0
answers
49
views
Does G-computation have any advantages over propensity score based methods for very small sample sizes (e.g., <40)
I am looking into the use of g-computation methods as an alternative for causal inference analysis to propensity score based methods (e.g., IPTW, matching). Does anyone have any examples of using this ...
0
votes
1
answer
39
views
If I use entire data, the IPW is effective?
When it comes to causal inference, if I use the entire population data, is Inverse Probability Weighting (IPW) still effective?
I have access to the entire population data and need to conduct some ...
- 61
1
vote
0
answers
108
views
Assessing whether the probability of being assigned treatment is equal (or reasonably close) between two individuals/groups [closed]
I'm currently studying the textbook Design of Observational Studies, second edition, by Rosenbaum. Chapter 3 Two Simple Models for Observational Studies says the following:
3.1 The Population Before ...
- 2,204
3
votes
1
answer
165
views
What are the main pros and cons of the most commonly used weighting methods?
There are many methods to generate balancing weights in observational studies (see, for example, the many methods implemented in the amazing WeighIt package).
I have seen some great discussions about ...
- 79
0
votes
0
answers
60
views
Propensity Score Matching and Weighted Regression Analysis
I have a dataset of ~N=1000 and I want to estimate the average causal/treatment effect of an exposure on an outcome.
I've used propensity score matching to balance baseline covariates, and my matched ...
- 1
1
vote
0
answers
37
views
How to use propensity scores in real examples
I am trying to understand how to use propensity score matching in a real world example (e.g. case control study).
Step 1: Based on what I understand, I think a Logistic Regression is first used to ...
user409304
0
votes
1
answer
78
views
Causal inference and Propensity score
I am trying to understand Rubin's causal model but I can not make the connection between certain notions.
The problem of causal inference lies in calculating the counterfactual, i.e. knowing what the ...
- 63
5
votes
1
answer
417
views
How to derive the GMM estimator for the Covariate Balancing Propensity Score?
The covariate balancing propensity score (CBPS) described by Imai and Ratkovic (2014) involves fitting a logistic regression for the propensity score $\pi_\beta(\mathbf{X}) = P(T = 1\vert\mathbf{X})$ ...
- 36.8k
1
vote
0
answers
20
views
Is a MSM with no lagged values equivalent to simply using IPW to balance a data set?
I am working on a project employing a panel data set with a large N but a fairly small T (5 time periods). MSMs seem like a good strategy, but I am wary of the incorporation of lagged values since ...
- 825
2
votes
0
answers
56
views
IPW Weights for Marginal Structural Models for Different Estimands
As Blackwell and Glynn 2018 note, an interesting property of marginal structural models is their ability to estimate treatment effects that account for the dynamic properties of panel data. For ...
- 825
2
votes
1
answer
111
views
How to obtain the (weighted or unweighted) L1 imbalance measure for "raw" data
I have two questions regarding the JASA paper, Multivariate Matching Methods That Are Monotonic Imbalance Bounding, by Iacus et al. (2011), the authors produce Figure 2 (left panel) where they compare ...
- 43
1
vote
0
answers
38
views
How to estimate interpretable treatment effects using a marginal structural model?
Say that I estimate a marginal structural model with weights obtained by inverse probability weighting. Imagine that my model looks something like: $Y_t = X_t + X_{t-1}$ again, with observations ...
- 825
1
vote
0
answers
126
views
Variance of Influence Functions, Cross-fitting, and the Propensity Score
Following example 2 in this paper, suppose I wanted to estimate $\psi = E[E[Y|X,A=a]] $ and I had an influence function follows:
$$
IF(\psi) = \frac{A}{\pi(X)}\{Y-\mu(X)\} - \psi
$$
where $\pi(X)$ is ...
- 43
2
votes
1
answer
77
views
Treatment Effect Insignificant with OLS but Significant After IPTW
I'm working with some observational data and wanted to assess the effect of variable (D) on outcome (Y) after controlling for a vector of covariates (X). As the data is observational, I wanted account ...
1
vote
0
answers
26
views
Is there relationship between propensity score based causal inference and sampling weights?
Consider observational study with single outcome $Y$, single covariate $X$ and treatment assignment variable $W$. Under unconfounded treatment assignment assumption, $E_{sp}[Y(1)]=E[\frac{Y_i^{obs}W_i}...
- 1,465
1
vote
0
answers
303
views
Do the weights of IPTW have to sum up to the population size?
I'm new to Inverse Probability of Treatment Weighting and trying to understand the mechanism. From several practical guides I understood, that the weights should add up to the population size for ...
- 281
3
votes
0
answers
343
views
Estimation of Propensity Score using Random Forests
Suppose that one has a binary treatment $Z$, and assume that $Z=1|X=x \sim Bern\left(e(x)\right)$.
Furthermore, suppose I want to estimate the propensity score by a random forest. Are there ...
- 111
2
votes
0
answers
63
views
can someone explain the application of row.weights and sampling.weights in lavaan? [closed]
I am trying to include the weight derived from Inverse Probability of Treatment Weighting (IPTW) into a structural equation model (SEM). However, it is unclear how it should be incorporated into the ...
- 21
1
vote
1
answer
482
views
Difference between normalized difference and standardized mean difference in cobalt?
In Imbens & Wooldridge (2009, p. 19), they define the normalized difference as:
whereas the cobalt's package standardized mean difference uses by default (for the ATE) "the 'pooled' standard ...
- 187
1
vote
1
answer
168
views
Propensity Score Matching at the district level and selecting variables
I am trying to use PSM for program evaluation. My data is at the individual level and I would like to do the PSM at the district level (match districts with each other rather than individuals).
Based ...
- 13
0
votes
1
answer
759
views
Formula for standardized mean difference in Cobalt package for categorical variables
I am having troubles in understanding the formula in cobalt package used for standardized mean difference calculation in BINARY variables
...
- 159
0
votes
0
answers
31
views
Confounding variables are nested with treatment, not able to be measured, how to address the influence from confounding factors?
We gathered driving data from two cohorts of drivers belonging to the same age group. The first cohort, Group A, utilized System A (treatment group), whereas Group B drove vehicles without this system ...
- 51
1
vote
1
answer
79
views
How can you rewrite the estimand in terms of propensity scores? Dowhy question
I am going through the backdoor criterion and how we get from an expression involving do to one which doesn't as below.
What i don't quite get is how to rewrite this estimand in terms of propensity ...
- 579
1
vote
1
answer
67
views
Invalid use of Propensity Score Matching?
I wonder if using a propensity score in the following situation is wrong. Imagine I have the next causal model
$$X = N_x$$
$$Y = f(X, N_y)$$
$$ Z = g(X, Y, N_z)$$
Where $N_z, N_y, N_x$ are ...
1
vote
1
answer
236
views
How to Evaluate and Visualize the Positivity Assumption for the ATT
I am trying to formally evaluate and visualize the satisfaction of the positivity assumption when estimating the ATT using R and I am having a tricky time figuring out how to do so. As Greifer and ...
- 825
2
votes
1
answer
321
views
How do matching/weighting outperform regression adjustment for making causal inferences? [duplicate]
In reviewing my notes about making causal inferences under the selection on the observables identification strategy, I reviewed some pieces that make critiques against contemporary strategies in ...
- 825
1
vote
1
answer
166
views
Whether covariates that are balanced at baseline should be put into propensity score matching
I am performing a logistic regression under propensity score matching.
Before matching, the baseline characteristics of intervention group A and intervention group B are statistically described, where ...
- 75
2
votes
0
answers
116
views
Using propensity score methods with multilevel time series data
I wanted to understand whether it would be feasible to use propensity score matching/weighting/stratification on my data. I'm investigating a region of the world where a number of countries joined an ...
- 21
3
votes
1
answer
53
views
Does Propensity Score theorem invalidate linear regression?
The propensity score theorem in Causal Inference implies that $E[T|X]=E[T|p(X)]$, i.e. it is sufficient to control for the propensity score to retrieve the average causal effect. Then, we can use a ...
- 51
2
votes
0
answers
148
views
Propensity score non parametric estimation
In several papers, in the 'double machine learning' literature, the propensity score (a nuisance parameter) is estimated non parametrically. It is a bit unclear how this estimation is performed, as ...
- 111
2
votes
0
answers
27
views
What to make of differeing results from standardized means and density when balancing a data set?
I am attempting to estimate the effect of a treatment (pko) on a specific outcome and I am using IPW to do so, along with nearest neighbor matching and coarsened ...
- 825
7
votes
1
answer
265
views
Are only confounders used to generate propensity scores for propensity score matching/IPW?
If the purpose of a propensity score is to generate a "propensity" of a unit receiving treatment, then is it correct to say that the predictors of the treatment need not be confounders (...
- 825
3
votes
0
answers
242
views
IPW Weights Are "1" For Treated Units
I am in the process of setting up my analysis using IPW with the R package WeightIt. My issue is that my weights seem off from the examples and reading materials I ...
- 825
1
vote
0
answers
35
views
Experiment Design with Known, Targeted Probability of Receiving Treatment
Is there a name in the literature for the following experimental design?
Suppose we have $i=1,...,n$ individuals in our sample. Denote their baseline covariates, binary treatment assignment indicator, ...
- 185
2
votes
1
answer
233
views
Merits of different matching & weighting methods for multiple treatments
I have three land use classes (natural farming, chemical farming, and forest) and would like to compare the densities of different bird species between them. I would like to get the ATE (I think).
I ...
- 21
1
vote
1
answer
129
views
Selecting optimal iteration for GBM purposed to obtain propensity score estimation
I am interested in propensity score estimation by GBM. I am reading Propensity Score Analysis by Bai and Pan. The book suggested several metric to evaluate balancing of propensity score by GBM. In GBM ...
- 1,465
1
vote
1
answer
124
views
Positivity, Causal Inference and Machine Learning methods
I ran into a problem where the predicted probabilities of being treatment/control from a machine learning model (the XGB model) matches the actual outcome almost exactly and have area under the curve (...
- 53
2
votes
0
answers
149
views
Marginal structural model - help with some concepts
I'm trying to gain some (deeper) understanding of MSM's - what exactly they are and when they might be appropriate to use.
Are my thoughts on the following correct (please feel free to correct any ...
- 857
1
vote
0
answers
205
views
Power calculation for observational study vs using propensity score matching estimating treatment effect
For case-control study study, one can calculate sample size to determine power of the study with respect to particular estimator of interest. Suppose I am interested in treatment effect in case ...
- 1,465
1
vote
0
answers
49
views
Why does normalizing difference score>0.25 indicates selection bias which cannot be corrected by regression?
I am reading Propensity Score Analysis(2014) by Guo and Fraser chapter 1 section 4. Denote $\Delta_X$ normalizing difference score of covariate $X$.
"Following Imbens and Wooldridge, a $\Delta_X$ ...
- 1,465
3
votes
1
answer
421
views
Entropy balancing---how do we assess overlap?
I am familiar with propensity score weighting. I set up the propensity score model, and then generally check for balance and overlap in propensity score to ensure that assumptions are met. However, I'...
- 57
0
votes
0
answers
45
views
Is propensity score matching out of favor? [duplicate]
I came across this post, which was largely nonsensical, but a respondent suggested the original poster follow up with two articles:
Ho, D. E., Imai, K., King, G., & Stuart, E. A. (2007). Matching ...
- 399
1
vote
0
answers
320
views
Regression discontinuity vs propensity score matching
I have recently read some pieces suggesting that regression discontinuity designs could be the best statistical approach for causal inference stemming from non-randomized studies (eg 1 and 2).
However,...
- 2,622
4
votes
1
answer
563
views
What is the positivity assumption required for matching and ATT estimand?
Does ATT estimand require a less stringent positivity assumption in matching?
For example, if a small treated group is matched to a large control group, most of the control subjects will be discarded ...
- 773