Questions tagged [matching]
Matching refers to a process in experimental design in which observations are sampled in a systematic, non-random fashion to be analyzed more efficiently with special statistical methods.
544
questions
1
vote
1
answer
25
views
Paired or unpaired tests after using propensity score matching [closed]
I used propensity score matching in one of my studies to reduce the effects of confounding.
Although some authors have suggested that methods of inference appropriate for independent samples can be ...
2
votes
1
answer
28
views
Retrospective cohort study with case:control matching 1:5. Is clogit really appropriate?
I study school grades in children with brain tumors (=cases), using databases.
Every case in the database has five healthy controls, matched on sex, birth year and residency. The hypothesis is that ...
- 21
0
votes
1
answer
23
views
Statistical test for 3x2 contingency table, with cases:controls matched 1:5 (i.e. non-independent)
I am examining a cohort of cases against controls, matched at a 1:5 ratio (on sex, birth year, and residency/postal code at time of diagnosis).
Education level
Cases (n=206)
Controls (n=957)
p-value
...
- 21
0
votes
1
answer
24
views
Is it ok to include unbalanced covariates in regression after full matching?
A couple of my covariates remained imbalanced after full matching. Is it acceptable to include those imbalanced covariates in the post-matching regression analysis (in addition to the matching weights)...
0
votes
1
answer
26
views
What should the estimand be in Coarsened Exact Matching?
I am using the MatchIt package in R for Coarsened Exact Matching. I only understand the basic idea of Coarsened Exact Matching. ...
- 29
0
votes
0
answers
22
views
A policy evaluation with a dependent variable which is continuous and in (0,1) interval
I am conducting research in which I want to investigate the effect of tax incentives on research and development intensity in a group of firms. I have access to the data of a survey that:
It's only ...
- 1
1
vote
0
answers
25
views
Matching weights vs 'weighting' weights?
What is the difference between matching weights obtained from full matching (with propensity score or mahalanobis distance, using MatchIt and then match.data in R) and the "weighing" weights ...
- 11
1
vote
1
answer
52
views
Marginal Effect for Poisson Model
I am using package marginaleffects for calculating the AME of an exposure variable on a count dependent variable. I am hence using Poisson (and Negative Binomial as robustness). The dependent variable ...
- 87
0
votes
0
answers
33
views
Propensity Score Matching on Pre-treatment Covariates
I plan to do a propensity score matching and then do a Difference-in-Difference design. I'd like to only match using pre-treatment covariates. However, when I use the MatchIt package in R, I could not ...
- 1
2
votes
0
answers
23
views
How to use Kernel weighted PSM and IPW in Triple Difference estimator (ika Difference in Difference-in-Differences)
I have learned a lot from @dimitriy answer in this post: Propensity Scores Weighted DID
I am currently utilizing the triple difference estimator for estimating treatment effects (please see the paper ...
- 21
0
votes
0
answers
19
views
Interpreting a coefficient for a matched/weighted data set for non-linear outcomes
Prior research has demonstrated that the interpretation of a regression coefficient no longer represents the marginal effect once non-linear term transformations (logs, interactions, etc.) are ...
- 539
1
vote
1
answer
56
views
Propensity scores: Variable K:1 matching with caliper- some technical questions (MatchIt, cobalt)
I have a data set from a cohort comparing two treatmens which I want to balance via propensity score matching.
I read some literature and decided to use a variable K:1 matching because this seems to ...
- 11
1
vote
1
answer
66
views
Log-Rank Test After Optimal Pair Matching
I have clinical data and used optimal matching in the MatchIt package to match cases to controls on several variables. Matching was done in a 1:1 ratio, and balance was achieved.
I then did a Kaplan-...
- 225
2
votes
1
answer
61
views
In g-computation after matching, do we mimic the formula used in matching even if you would have typically added more terms such as polynomial terms?
I have some questions regarding best practices in g-computation post-matching in the standard case of a dichotomous treatment and a continuous outcome.
When defining the regression model with the ...
- 23
1
vote
0
answers
53
views
Creating a case-control cohort in R
I have a variable: hyperglycemia and a bunch of other variables like Age, Gender, BMI, many other variables. I would like to create a matched case control group case being those with hyperglycemia and ...
Daniel
0
votes
0
answers
117
views
How do you calculate descriptive statistics before and after matching multiply imputed datasets?
I'm trying to find the best way to calculate descriptive statistics following matching on a dataset with missing values (handled using multiple imputation). The purpose is to present this data as &...
- 1
2
votes
1
answer
86
views
Bootstrap for nonlinear regression (quasibinomial GLM) using `marginaleffects::inferences` after matching
Background
Looking into the MatchIt articles made me realize that using Bootstrap with BCa is a better practice for assessing uncertainty estimation (since: "For nonlinear models (e.g., logistic ...
- 125
8
votes
3
answers
2k
views
How can this counterintiutive result with the Mahalanobis distance be explained?
I encountered a strange issue when performing Mahalanobis distance matching. Let's say I have one treated unit with the following values on two variables: $T:(17, 4)$. I have two control units with ...
- 30.2k
4
votes
1
answer
81
views
For a time-series cross-sectional (TSCS) data, how does one account for 'future' treatment history while calculating ATE/ATC/ATT?
I refer to this seminal paper (Matching Methods for Causal Inference with Time-Series Cross-Sectional Data) by Kosuke, In Song and Erik.
Context
In the paper, the authors propose the use of a two-...
- 43
0
votes
1
answer
28
views
Comparison of matching and 'normal' regression outcomes
As a form of sensitivity test, I want to compare the results from a regression model where I first conducted matching and the results from the same model without conducting matching prior to that. For ...
- 1
0
votes
1
answer
43
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 ...
- 165
1
vote
0
answers
29
views
Event study estimates between matching groups
I have a dataset with: ID, years, $Y$ (dependent variable), treatment, a matching-group variable and a control variable $X$.
I have created the matching-group variable using $X$, the idea is to ...
- 11
2
votes
1
answer
48
views
Should matching (without discarding units) be attempted before weighting?
In ecology, we are often working with very small samples, i.e. I only have 13 sites of each of the two farming practices I am comparing. Thus, I really don't want to discard any of my sites and thus I ...
- 21
0
votes
1
answer
94
views
PSM Propensity Score Matching Combination of exact and full matching
I would like to compare two groups (one treatment group and one control group). In both groups one can identify three different subgroups (let's call them Index-subgroup). I would first like to ...
user389026
1
vote
0
answers
22
views
Extracting matched sample dataframe from MatchThem
I am using the R package MatchThem to match a dataset that contains 19 control samples and 13 case samples.
After running the ...
- 11
0
votes
0
answers
15
views
Solution when matching on $X_1$ increases imbalance in $X_2$ and vice versa?
My work team has come across a situation in our propensity score matching analysis where matching treatments to controls on one covariate, $X_1$, increases the imbalance in a second covariate, $X_2$, ...
- 5,442
0
votes
0
answers
36
views
How can I get the most balanced groups using matchit in R?
I have an imbalanced treatment group and a control group. I want to use Matchit in R to calculate propensity scores and get balanced groups.
However, there are several parameters I have to set in the ...
- 1
0
votes
0
answers
14
views
How to implement exact matching on some variables and nearest-neighbor or optimal matching on others? [duplicate]
I have a group of patients that I want to match on a number of variables: age, sex, BMI in category (5 categories), year and comorbidity score.
In order to be computationnaly efficient I'd like to do ...
- 745
0
votes
1
answer
96
views
Log-Rank test with propensity score matched data
Let's assume that I have two groups that I would like to compare in terms of death rates. One group (group A) got a specific treatment and the other group (group B) did not get any treatment. I ...
user389026
0
votes
0
answers
36
views
Is fixed-effects ordered logit models (feologit in Stata) suitable for matched case-control data?
Does anyone happen to be familiar with this method: fixed-effects ordered logit models (feologit in Stata)? In this paper: https://journals.sagepub.com/doi/pdf/10.1177/1536867X20930984
In the paper, ...
- 617
3
votes
1
answer
167
views
What are the possible solutions to do matching on very large dataset
Hi I would like to match a group of treated patients with an untreated group. I have about a million patients in the treatment group and ten times that in the control group. Conventional matching ...
- 745
0
votes
1
answer
81
views
Evaluating the quality of matches after Mahalanobis distance matching on stata
I am new to Mahalanobis distance matching and trying to see for information on how I can assess whether the matches I got using Mahalanobis matching are good matches.
...
- 5
0
votes
1
answer
47
views
Can I replace a binary treatment variable with a continuous treatment variable after matching
I have a continuous treatment variable which I dichtomize into a binary treatment dummy. Using MatchIt and the newly created dummy variable, I conducted genetic matching to create a dataset of ...
1
vote
1
answer
184
views
Nearest neighbor matching without replacement with MatchIt
I am doing the nearest neighbour matching from the package MatchIt.
For example
...
- 29
1
vote
1
answer
182
views
how to understand the weights in PSM？
When using propensity score matching or weighting, a column of weights is generated that is used to estimate the effect of interest.
According to a blog I read, there are three types of weights ...
- 33
3
votes
3
answers
86
views
How should control candidates be decided for causal inference?
I’m getting a bit confused about who to include in the control and treatment pools and would appreciate the help. I need to estimate the effect of treatment where 100% of the population was assigned a ...
- 31
3
votes
1
answer
178
views
Is this sample size big enough to analyze with Propensity Score Matching?
Suppose I have a dataset where 9 patients occured with the post-operative complication.
(e.g. information such as height, smoking, weight, age, disease status) and rest of the 150 patients without the ...
- 47
2
votes
1
answer
182
views
Using Propensity Score Matching to Reduce "Class Imbalance" Biases?
Suppose I have a dataset where 100 patients have the disease (e.g. information such as height, smoking, weight, age, disease status) and 10000 patients do not have the disease (i.e. class imbalance).
...
1
vote
0
answers
89
views
matchit: Specified caliper still matches everyone!
Suppose we are using the matchit function from the MatchIt package in R, as in the following example given in the package,
...
- 11
0
votes
0
answers
34
views
How to estimate variance in censored data?
I have the following model. Let $BP$ be a continuous variable s.t. true relationship is $BP=144+0.5age+4sex+3gene+\epsilon$. Suppose for subjects with $BP>160$, there is a probability of 0.5 ...
- 1,365
1
vote
0
answers
30
views
Matching, event study and cohort data
I have merged data to create an event study for the treated population. This treatment happens in 4 batches for some university students in certain cohorts (when they turn 19 in 2005 and are at ...
- 11
2
votes
0
answers
108
views
How to model an interaction in a propensity-score matched dataset
Suppose I am performing a propensity score matched analysis using the MatchIt package in R, following the example reported here: https://kosukeimai.github.io/...
- 646
1
vote
1
answer
125
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 ...
- 539
2
votes
0
answers
36
views
Propensity-Score Matching - what's the best choice when matching?
I'm using matchit package to create a propensity match.
I'm trying to match control and treated with a 2:1 ratio in order to maximize the population and exclude ...
1
vote
1
answer
83
views
In causal inference, can you control for confounders by matching the treatment and control group based on the time series of the outcome variable?
Suppose that Walmart has 100 stores. It has a coupon for cereal, and it wants to know if the coupon increases cereal sales by a significant amount.
Walmart puts the coupon on the cereal shelf in 10 ...
- 111
0
votes
0
answers
11
views
"Reverse" case-control: is there a way to match most DISsimilar "controls" to my cases?
For one of my analyses, I have to find cases that are most different from my cases. It's like case-control matching, but "controls" should be as different as possible. What would be a good ...
2
votes
0
answers
22
views
Understanding the matching propensity score
I am currently using the package PSweight from R to do propensity score with 2 different methods : IPW and Matching.
I first used it to make a propensity score based on IPW method to get the average ...
- 23
2
votes
1
answer
61
views
Evaluating success of propensity score matching with single metric that accounts for both covariate balance and matching rate?
Propensity score matching techniques can be assessed and compared with covariate balance metrics like the standardized mean difference (SMD).
However, SMDs don't account for varying matching rates.
...
- 5,442
1
vote
0
answers
57
views
Propensity score matching with replacement - OK to trim excess control group matches to same treatment subject?
My team is conducting propensity score matching with 1:1 nearest neighbor replacement for a case-control healthcare study.
While we're obtaining match rates of 80-90% with good covariate balance, we ...
- 5,442
1
vote
0
answers
118
views
PSM kernel matching and bandwith: what observations are used
I'm using PSM with an Epanechnikov Kernel and a bandwith of 0.06.
I'm confused about which observations are matched. I thought it was (broadly) like a wheighted radius matching, where every control ...
- 11