# Questions tagged [causality]

The relationship between cause and effect.

1,841 questions
Filter by
Sorted by
Tagged with
1 vote
8 views

### Techniques for uplift modelling/Conditional Average Treatment Effect(CATE) estimation for observational data

I have very recently started learning CI and was going through this very famous paper:https://proceedings.mlr.press/v67/gutierrez17a.html which mentions that Randomised Control Trials are an essential ...
1 vote
57 views

### How to determine if a directed edge is visible? What does visibility tell us?

Both of the above PAGs (Partial Ancestral Graph) are generated using the FCI (Fast Causal Inference) algorithm under identical conditions, the only difference being that the one on the right is ...
• 4,804
29 views
+50

### Why are these 2 MAGs Markov Equivalent?

Can someone help me think through why these 2 MAGs (Maximal Ancestral Graph) are Markov Equivalent (imply the same constraints by the m-separation criterion)? I ask because I am not yet fully ...
• 4,804
12 views

### When does Matching result in an ATE vs. ATT on observational causal studies?

I have read that matching nearly always yields the $ATT$ effect, but that subclassification matching can yield the $ATE$. I am therefore wondering what is a heuristic for determining what kinds of ...
• 4,626
10 views

### For estimating weights in the Synthetic Control Method, what linear combination of the outcomes and symmetric matrix are used?

I am reading Synthetic Control Methods for Comparative Case Studies Paper, and on pg. 5 it states how to estimate Synthetic Control Weights. The following is a summary of the methods in the paper. I ...
• 4,626
1 vote
39 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 ...
• 4,626
867 views

• 572
1 vote
37 views

### Can Multiple Regression Output (coefficients, ratios, etc.) Be Given a Predictive Interpretation?

Generally speaking, regression output can be given a causal interpretation for typically one variable in the model (that is under the assumption of no unobserved confounding and this is not to speak ...
116 views

### Is it possible to evaluate causal algorithms on real world observational data?

Lot of times I get asked to use causal algorithms (e.g. algorithms estimating intervention results, or in general causal inference algorithms) and to compare them against non-causal prediction ...
• 460
1 vote
59 views

### Granular difference-in-differences with non-repeating unit of observation

I want to analyze changes in characteristics of job postings around an (exogenous) event. However, rather than conducting the analysis at the job poster level (e.g., a company or geographic area), my ...
• 55
21 views

### How to interpret the conditional expectation mean $E[Y(0)\mid Z=1]$ where $Y(0)$ is the potential outcome on control and $Z$ the treatment assignment?

How can I interpret the conditional expectation mean $E[Y(0)\mid Z=1]$ where $Y(0)$ is the potential outcome on control and $Z$ the treatment assignment? I believe it is the "expected outcome if ...
• 4,626
1 vote
17 views

### Why the confusion in Difference-in-Differences with pre-intervention/post-intervention periods and treated/control units?

In Difference-in-Differences, it typically has a pre-intervention and a post-intervention periods. This is coupled with treated and control units. From what I understand, since the comparison for the ...
• 4,626
14 views

### Is the treatment effect in the Synthetic Controls Method the ATE or the ATT?

In the Synthetic Controls Method, the treatment effect for a time $t$ is given to be the difference in outcomes for the treated unit and control unit, in the post-treatment period. In this regard, I ...
• 4,626
1 vote
30 views

### Causal additive intervention operation

I am reading the following paper by Shpitser and Pearl: Effects of Treatment on the Treated: Identification and Generalization. On page 517 they provide a corollary concerning the additive ...
• 697
498 views

### What is it called when two variables causally affect one another?

Suppose two variables X1 and X2 are correlated and we know that X1 causes X2 and X2 causes X1. For example, leg strength and an interest in cycling interest might be correlated. And (suppose) we know ...
• 336
1 vote
25 views

### How to do analysis of correlation from multiple-timepoints measurements?

My case is analyzing the association between the concentration of HIV DNA prior to therapy (time point 0, $t_0$) and the concentrations of biomarkers of HIV infection after therapy, measured in 6 time ...
• 13
1 vote
32 views

### Causal Inference: Meta Learners usage

I have been running causal inference using Econ ML package on my data. I have a dataset containing customers divided into treatment and control and many other features. I run matching on those and ...
23 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 ...
1 vote
69 views

### Event study regression specification: interacting covariates with leads and lags

I want to create an event study regression specification for the following: $$\ln(y_{ijt}) = \gamma \ln (x_{jt}) + \tau \ln(p_{t}) + \lambda \ln(x_{jt}) * \ln(\mbox{p}_{t}) + \epsilon_{ijt}.$$ I am ...
33 views

### Is naive mean estimator uniformly worse than HT (IPW) or Hajek estimators in survey sampling? If not, why is it less discussed in the literature?

Consider a toy example: we are interested in the average height of $n$ students $\bar{\tau}=\frac{1}{n}\sum_{i=1}^n\tau_i$, but for some reason, we can only access a random subset $S$ of it. Every ...
• 113
1 vote
24 views

### How can I derive the bivariate breakdown of a multiple linear regression?

I recently start to learn causal inference from this website. It claims that "efficient of a multivariate regression is the bivariate coefficient of the same regressor after accounting for the ...
• 11
1 vote
37 views

### Is including weights in g-computation not the same as a plug-in doubly robust estimator?

In the R package vignette for WeightIt(), in the section "Modeling the Outcome", it explains that (assuming I'm reading correctly) that the purpose of applying g-computation after creating ...
• 121
39 views

### Missing Data in experiments, MIPO, MIPO|X vs MCAR, MAR, MNAR [closed]

Hello I was reading Field Experiments by Alan Gerber and Donald Green and was introduced to the idea of missingness independent of potential outcomes (MIPO). And MIPO|X which is missingness ...
• 508
1 vote
145 views

### Average treatment effect: counterfactual and graphical derivation

I have some (shameful) doubts about the Average Treatment Effect (ATE), also known as Average Causal Effect (ACE). In this setting, I am interested in a binary exposure/treatment variable ...
• 697
1 vote
40 views

### question about including an independent variable

I'm seeking advice on whether to include the independent variable 'smoking' (Yes/No) in my analysis. The objective is to examine the impact of COVID-19 on female construction workers. The outcome ...
• 567
30 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 ...
1 vote
32 views

### Rubin Causal Model and Selection Bias

In the Rubin Causal Model, with a binary treatment $T \in \{0,1\}$, the selection bias is expressed as: $$E(y_0|T=1) - E(y_0|T=0)$$ where $E(y_0|T=1)$ denotes the ...
• 279
9 views

### longitudinal causal mediation analysis with recall periods

I am new to causal mediation analysis (using longitudinal data), and trying to learn best practices. I have tried to read some methods reviews -- e.g. Preacher2015, O'Laughlin2018, Assaad2022 -- but ...
• 215
1 vote
43 views

### Estimating Average Marginal Effects with Ordered Logistic Regression

I am estimating a series of ordered logistic regression models for a 4-level ordered dependent variable and I am trying to estimate the average marginal effect (using {marginaleffects}) of moving from ...
1 vote
102 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 ...
• 1,942
54 views

### How to decompose the Naive Average Treatment Effect into the ATE, Selection Bias, and Differential Effect Bias

From this online document, it is assumed that we have a treatment and control group, with $N_1$ and $N_0$ being the number of treated and control units, respectively. $D$ is the treatment variable ...
• 4,626
7 views

### Variation in treatment condition in regression discontinuity model

I have a conceptual question regarding the implementation of a regression discontinuity model. I am working on a paper examining the effect of a policy that allows for program funding if a geographic ...
1 vote
40 views

### Is stationarity always an requirement for Spearman correlation between two time-series?

I would like to know if time-series data always (in every case imagineable) has to be stationary before computing the Spearman correlation between two time-series. I subsequently explain my specific ...
• 235
63 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 ...
92 views

### Causal Mediation Analysis with treatment smoothed in the outcome stage and linear in the mediator stage

I am considering a mediation analysis that looks like the following in r: ...
• 65
32 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 vote
22 views

### Causal Loop Diagrams vs. Structured Causal Models

What's the difference between causal loop diagrams (CLDs) and structured causal models? I know that Pearl's Structured Causal Models (SCMs) are directed acyclic graphs (DAGs). However, it seems that ...
146 views

### Stable violation of faithfulness

Faithfulness is often justified by the argument that any violations of it require very specific "fine-tuned" parameters (for some appropriate SCMs/SEMs/SFMs), and that such violations are ...
• 161
18 views

### conditional-on-positives bias

I am reading the Bad COP section on https://matheusfacure.github.io/python-causality-handbook/07-Beyond-Confounders.html#bad-cop. I am confused if  E[Y|T = 1] - E[Y|T = 0] = \\ E[Y|Y > 0, T = 1]...
• 113
32 views

### Process of establishing a causal relationship

I am trying to clarify my understanding of establishing a causal effect, as mentioned in textbooks/online resources that I am reading. Is the following a correct understanding? To identify a causal ...
• 389
1 vote
41 views

### causal forest and conditional unconfoundness

I recently read the paper by Kunzel et al. on meta-learners, I encountered there the following statement: I fail to understand this statement, as from my understanding, under the assumption of strong ...
• 1,355
10 views

### Inducing Paths Definition Ambiguity

I am relatively new to the causal discovery space and I have a clarifying question regarding the definition of an inducing path. In Causation, Prediction and Search (page 173), the authors state that ...
1 vote
12 views

### Including the same shock with different share in Bartik instrument

Suppose I'm interested in the causal effect of $x$ on $y_i$, where $i$ is the unit of analysis that could either be individual, county, or state level. Suppose $x$ is randomly assigned to unit $d$, ...
• 201
12 views

### An interesting setting in statistical causal inference

I'm writing a research paper and considering the following settings: There are three discrete variables which construct a simple causal graph: Treatment T, outcome Y, and confounders U. Here we have ...
33 views

### ACME Significant in Mediation analysis, but not Proportion Mediated and Fitting terminated with step failure warning

I am running a series of mediation analyses in R using the mediation package and the following code: ...
• 65
1 vote
41 views

### Causality - Can adding predictors unblock causal effects?

My question pertains to the following empirical situation. You have a bivariate model (e.g. regression), and you find no effect of $X$ on $Y$. Then you add a covariate in your model, $Z$, and now you ...
• 911
### Definition of $\text{do}$ operator [closed]
I'm looking for a hint in understanding semantics of $\text{do}$ operator. Starting from the original distribution $P$, an intervention $\text{do}(X=x)$ takes us to another distribution $P_x$ - in ...