Questions tagged [causality]
The relationship between cause and effect.
331
questions with no upvoted or accepted answers
9
votes
0answers
525 views
Propensity Score Matching – How do the mechanics lead to a different result than unmatched?
The gist of propensity score matching, as I understand it, is as follows:
You want to estimate the average treatment effect (ATE) of a treatment on some outcome. However, if you simply calculate the ...
6
votes
0answers
261 views
Avoiding adjustments for time-varying controls in difference-in-differences (DID)?
In difference-in-differences (DID) analysis, it seems like a "folk theorem" that one should be very wary of adjusting for time-varying controls. The reason, eminently plausible, is that ...
6
votes
1answer
5k views
Difference-in-difference in panel data
Under which conditions should we expect the difference-in-difference estimate to be equal to the equivalent panel data model?
Strictly speaking, whenever we have a experiment that offers a well ...
5
votes
0answers
128 views
Regression where predictors are correlated with past values of y
Setup
We are interested in estimating a model for the following setup:
$Y_t=\beta_0 + \beta_1^{'}X^{'}_t + \epsilon_t$
$COV(X^{'}_t,Y_{t-1,t-2,...,1} | X_{t-1,t-2,...,1}) = 0$
Where $\epsilon_t$ is ...
5
votes
1answer
253 views
Instrumental variables: In which cases would the average treatment effect on the treated (ATT) and local average treatment effect (LATE) be similar?
It seems that if the proportion of always-takers in the control group (to whom eligibility was not assigned) is much smaller than the proportion of compliers in the treatment group (to whom ...
5
votes
1answer
50 views
Treatment interference (causal analysis)
I am doing research on students and their perception of their grades. Specifically, I want to do an experiment where students either (a) see their actual grades in a course (as a percentage) - the ...
5
votes
0answers
46 views
Before using CV-selected Regression model for Inference, shouldn't model performance be evaluated on unused test set?
I just came across a biokinesiology paper that used some Machine Learning methods, but I think there is a flaw in their methodology.
The authors had data on stroke patients and used Lasso regression ...
5
votes
0answers
222 views
Structural equation and causal model in economics
Structural equations is a useful language for causal analysis in economics. In Causality Pearl (2009 cap 5) we can find the best discussion about this.
My question: is possible to use the concept of ...
5
votes
1answer
112 views
Conterfactual estimation in machine learning model
There are various techniques to build counterfactual estimations of certain variables for linear models in observational studies.
Some of those are based on comparing the change in the predicted ...
5
votes
0answers
357 views
Machine Learning for Causal Inference with Panel Data: Possible to combine ML estimators with additive/linear terms to derive diff-in-diff estimator?
My question is motivated by the following. First consider the non-panel case, where we have two groups, the treated group ($g=t$) and the comparison group ($g=c$), and are trying to estimate an ...
5
votes
0answers
768 views
How can I control for time-varying unobserved heterogeneity with panel data?
I'm just beginning to learn econometrics, and I just learned about the fixed effects estimator, and the first-difference estimator. It's quite straightforward that those techniques allow me to control ...
5
votes
0answers
70 views
Ranking of causal models
I remember seeing a paper that ranked causal statistical models, published by some research body, by quality in terms of generalizability or some similar facet(s). This would be a "hierarchy of ...
4
votes
0answers
45 views
Causal Bayesian network, causal diagram, structural causal model and marginal structural model: what do they exactly mean?
In the Book of Why, Judea Pearl gives a comprehensive overview of the causal diagrams (or causal graphs), but to me, the terminology is not clear yet. In the book, he presents Bayesian network in the ...
4
votes
1answer
88 views
Comparing time series data with multiple pairs of time series, or difference-in-difference with continuous treatment conditions?
My dataset contains time-series for two variables ($X$ and $Y$) from 2017 to 2020, for each of many different countries. Each country has its own time series for each variable (X_usa, X_india, Y_usa, ...
4
votes
0answers
92 views
Why is BART so accurate in causal inference?
The famous paper Dorie,2017 shows that BART performs dramatically well in causal inference. In my replication, MSE in BART can be 40% lower than MSE in other machine learning methods.
But all machine ...
4
votes
0answers
49 views
How is EMSE derived for causal trees in Athey and Imbens (PNAS 2016)?
Athey and Imbens build a non-parametric matching procedure to identify and estimate causal effects. To this end, they minimize the expected mean squared error (EMSE) of their procedure, but I don't ...
4
votes
1answer
121 views
Are model diagnostics necessary for linear model run on matched data?
On https://cran.r-project.org/web/packages/cem/vignettes/cem.pdf, it mentions that "Using the output from cem, we can estimate SATT via the att function. The simplest approach requires a weighted ...
4
votes
0answers
210 views
Causality in Time Series
I am reading an article which is trying to justify the need for causal inference in their inferential framework. The thought experiment is as follows:
Suppose a statistician is asked to design a ...
4
votes
0answers
59 views
Adjusting for experimentally-caused panel attrition when evaluating treatment effects
This question involves a questionable hypothetical scenario, but please bear with me.
Suppose I ran an experiment in a coffee stand where the treatment was playing country music instead of the usual ...
3
votes
0answers
80 views
All-subsets regression and parameter shift to estimate or identify omitted variable biases?
I have multiple ($12$) predictors ($X$) for an outcome (spending) where it's likely/possible that:
Some predictors are correlated
Some predictors could (partially) mediate the effect of others
There ...
3
votes
0answers
23 views
Reverse Causality with Additional Period
I have been struggling for a model to estimate related to sequential treatment effect and need a help desperately.
I would greatly appreciate it if you guide me to the resources or advice me on this ...
3
votes
0answers
20 views
How to find the [marginal] effect of X on Y when Y is binary and very rare. Can I make groups of similar X and model counts of Y instead?
tl;dr How to model the causal impact of X on binary Y when Y is very rare. Can I make groups and model count instead?
Background/What I tried
I want to know what the effect of the number of "...
3
votes
0answers
43 views
Failing to fully control for a variable
Lets assume we want to perform a 'reduced-form' causal analysis to evaluate the impact of a program on the dependent variable of interest. (However the question is more universal).
Lets further assume,...
3
votes
0answers
51 views
Extending external validation diagnostics to experiments with continuous treatment
Does the external validation diagnostic methods discussed in Stuart et al. (2011) (i.e., inverse propensity score weighted regressions) also apply to the experimental setting in which the treatment is ...
3
votes
0answers
275 views
Regression: Causation vs Prediction vs Description
In my experience it seems me that the interpretation about regression, its meaning and its scope, are debatable and great confusion exist about those things. It seems me that confusions are not go ...
3
votes
1answer
43 views
Measuring the causal impact of a policy that is not binding
This may be a little tricky because it's difficult to explain but bear with me. Assume a new policy implemented in 2015 which is a new requirement for firms, let's say for instance that the ...
3
votes
0answers
70 views
How come parents of $X$ always satisfy the backdoor criterion relative to $(X,Y)$?
Pearl et al. "Causal Inference in Statistics: A Primer" (2016) p. 61 presents the backdoor criterion:
Definition 3.3.1 (The Backdoor Criterion) Given an ordered pair of variables $(X,Y)$ in a
...
3
votes
0answers
171 views
Mathematical details in the definition of a Structural Causal Model
Pearl defines (see Causality, Judea Pearl, 2nd ed., Definition 7.1.1) a Structural Causal Model (SCM) as a triple $(\mathscr U, \mathscr V, F)$ where $\mathscr U$ is a set of "exogenous variables," $\...
3
votes
0answers
203 views
Theory behind Targeted Maximum Likelihood Estimation (TMLE)
There are many fine how-to articles describing how to implement TMLE but they avoid the details of the underlying theory. I'm currently working my way through Targeted Learning: Causal Inference for ...
3
votes
0answers
41 views
Applying Heterogeneous treatment effects to clinical research (non technical explanation)
I'm trying to understand the hype around this estimation of heterogeneous treatment effects in the machine learning literature lately. It seems super interesting, but alot of it is beyond me. I read ...
3
votes
0answers
29 views
Causality in variance with a BEKK model
I am using a BEKK model in the following form,
$$H_t=C^\ast{C^\ast}^\prime+\sum_{i=1}^{m}{A_i\varepsilon_{t-i}\varepsilon_{t-i}A_i^\prime+\sum_{j=1}^{s}{B_jH_{t-j}B_j^\prime}}$$
I first start with a ...
3
votes
0answers
390 views
How to relate roots of AR and MA to unit circle
I'm working on these problems and think I figured out most of the steps, but am stuck near the end as I don't understand how to relate my roots back to the unit circle in order to determine ...
3
votes
0answers
596 views
Pro and cons between Bayesian structural time series (BSTS) vs difference-in-differences?
Google's paper markets BSTS's benefits over DID such that
"In contrast to classical difference-in-differences schemes,
state-space models make it possible to (i) infer the temporal
evolution of ...
3
votes
1answer
250 views
Invariance of causal prediction
I am reading Causal inference by using invariant prediction: identification and confidence intervals by Jonas Peters (link to the resource is here: https://rss.onlinelibrary.wiley.com/doi/full/10.1111/...
3
votes
0answers
31 views
What simple test-case and corresponding solution, would demonstrate what it is possible to achieve using causation/precedence-analysis techniques
What would be minimal test-case(s) and corresponding solution, demonstrating curent know-how in causation/precedence analysis solving, or more simply
what is it possible to achieve using causation/...
3
votes
0answers
1k views
Marketing/Sales Mix/Response Models: approaches and comparisons
CV/SO Community: I am probably skirting (or crossing) the line of the preference for questions that can be answered vs. those that can (only) be discussed.
That said, I'm trying to wrap my head ...
3
votes
0answers
116 views
When does the Rubin Causal Model fail in practice?
The Rubin Causal Model frames the causal inference question as the problem of inferring missing potential outcomes (what the outcome would have been if a unit had received a different treatment) in ...
3
votes
0answers
500 views
reverse causality when dependent variable and independent variable are observed on different levels
Consider the following:
$$
Y_{ijt} = \gamma \cdot P_{jt} +X_{ijt} \beta+\epsilon_{ijt}
$$
The j dimension is only indicated for clarity, the regression is a panel regression on i-t dimensions.
The ...
3
votes
0answers
483 views
Estimating a confidence interval around the average of an interacted treatment effect
I've got a causal inference situation with heterogeneous treatment effects. In addition to simply estimating coefficients, I'd like to get an average effect of the treatment at the covariate values ...
3
votes
0answers
54 views
Guessing test question answers from scores
My teacher likes to give online quizzes that are about 20-30 questions long. Every student has the same questions in the same order. We are not told after taking the quiz which questions we got wrong, ...
3
votes
0answers
606 views
Heterogeneous Treatment Effects - How to test differences in the ATE?
I want to conduct a simple propensity score estimation where the treatment $D_i$ is a binary variable ($D_i=1$ individual $i$ participates in the labor market program, zero otherwise). I estimate the ...
3
votes
0answers
803 views
Implication / Interpretation of long term equilibrium VECM
I want to test the influence of exchange rates on a price index and struggle with the interpretations. My variables are I(1)
First, I ran an OLS on first differenced variables which indicated a ...
3
votes
0answers
680 views
Inductive vs deductive Inference
I am curious to know exactly, what are the (possible) differences between inductive and deductive statistical inferences in applied statistics.
Suggestions for some good resources to learn their ...
3
votes
0answers
538 views
Matched Analysis with Complex Survey Data
Complex survey data is that typically found produced by the National Center for Health Statistics (NCHS) or the NSLY; it typically contains information on PSU, strata, and weights. To make nationally ...
2
votes
0answers
20 views
Does education level affect salary?
This is an interview question that I came across, and I'm not sure how to answer it.
There are two ways of phrasing a causal question:
If someone has a high salary, was it because they had a high ...
2
votes
0answers
87 views
Size of Complier Group (Instrumental Variables) - Mostly Harmless vs. Abadie Kappa
Focusing solely on the case with a binary endogenous explanatory variable $D$ and binary instrument $Z$ and control variables $X$, Angrist & Pischke 2009 ("Mostly Harmless Econometrics") ...
2
votes
0answers
40 views
Non-causal AR models
There is a theorem which states that for a non-causal AR(MA) process, you can produce another equation for $X_t$, with a different (but related) white noise sequence, which is causal. (See the clipped ...
2
votes
0answers
54 views
complex Difference-in-difference formula
I have 5-year sales information from a grocery store in Canada. I want to check whether an event that happened in 2017, affected the effect of the price of a product on its sales.
For example, imagine ...
2
votes
1answer
52 views
Does convergent cross-mapping require you to control for other variables?
There's a really cool method called convergent cross-mapping (Tsonis et al. (2018)) that's used to see if two time-series are causally linked within a dynamic system. It seems really powerful and like ...
2
votes
0answers
41 views
how to measure the causal impact of interventions which happen at different times on different time series?
My data consists of a bunch of time series of daily clicks on some merchandises on a website, a portion of which have an intervention (only 1 intervention per time series), others don't. The ...
