Questions tagged [selection-bias]
Bias introduced by non-random selection of observations, such that the sample is not representative of the underlying population.
83
questions
0
votes
1
answer
26
views
How to understand random assignment eliminates selection bias in the potential outcomes framework
In Angrist & Pischke's book mostly harmless econometrics, they explain that if the treatment in an RCT $D_i$ is randomly assigned, then $D_i$ is independent of potential outcomes and the following ...
- 155
0
votes
0
answers
9
views
Is it possible to use Heckman's correction to correct prevalences and means for attrition bias?
I am interested in comparing the characteristics of participants of a longitudinal cohort study at baseline (Sample A) and follow up (Sample B). Some participants were lost to follow up and did not ...
- 137
1
vote
0
answers
11
views
Selection bias in postmortem data and creating an artificial earlier study endpoint
I want to analyze postmortem (neuropathology) data from dementia patients who are part of a larger ongoing observational study. At the time of the data freeze (i.e. the time at which I access the data)...
- 11
0
votes
0
answers
22
views
Adjusting for selection bias by controling covariate in linear regression
Is it possible to correct for a known selection bias in my sample of cases, when I know the source of the bias (and I can correctly assume that it is the only bias)?
I want to get an unbiased estimate ...
- 101
0
votes
1
answer
28
views
Questions Regarding Sampling Bias
I'm taking a course in R: "Data Analysis in R" on Coursera, and I came across this question during the lecture:
A retail store considering updates to their credit card policies randomly ...
0
votes
0
answers
12
views
Extrapolation problem when only part of space is observed
Let's say I want to open a vendor selling widget $W$ and I consider some fixed list of places to do so. Now, I have data regarding $W$ sales produced by competitors $C_1, C_2, C_3$ (who - combined - ...
- 101
1
vote
1
answer
25
views
How to correct for sampling bias in one population when comparing against another
I have two populations that I'd like to compare across certain metrics. However, most members of population A did not respond to our request for data, and those respondents that did are not ...
- 177
0
votes
0
answers
25
views
Overcome selection bias
I'm new to stata and I have a question regarding selection bias. For example,
I am testing the impact of risk management on firm total risk. So if firms use risk management, they should have a lower ...
- 1
1
vote
1
answer
44
views
Maximum likelihood of Normal density under selection
Consider the density function given by
$$
\left[\dfrac{\gamma_{\leq0} \mathbb{1}(t \leq 0) + \gamma_{>0} \mathbb{1}(t > 0)}{\gamma_{\leq0}\Phi\left(- \mu / \sigma\right) + \gamma_{>0}\Phi\...
- 63
0
votes
0
answers
25
views
Heckman selection model using multinominal probit
I am trying to run a Heckman two-stage selection model. However, the selection variables in the first stage have four categories respondents can select into. Therefore, I need to use a multinomial ...
- 177
1
vote
0
answers
22
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$ ...
- 1,315
0
votes
0
answers
37
views
Inverse Mills Ratio Interpretation
What is the interpretation of inverse mills ratio in Heckman Selection Model ? Why we are including it as an explanatory variable in the OLS estimator?
1
vote
1
answer
21
views
Control Group Selection Bias
I found a study that compared minor physical anomalies(MPA) between certain group of patients with the control group to determine if MPAs occur more frequently among these patients compared to the ...
- 11
1
vote
1
answer
19
views
Estimating interactions from non-interacting features
Suppose I have a sample $\mathcal{D}=\{(\mathbf{x}^{i}, y^{i})\}_{i=1\dots M}$ of binary variables $\mathbf{X}$ ($N$ of them) and a continuous variable $Y$ that I want to predict based on a linear ...
- 306
1
vote
0
answers
81
views
Nested Cross-Validation with Small dataset
I am currently working with a small dataset (only 175 samples, 45 features) and have been reading on the proper way to cross-validate my model. I had started with a basic cross-validation using a grid ...
- 11
2
votes
0
answers
90
views
Sampling weights in Cox proportional hazards models
I'd like to use sampling weights in a Cox proportional hazards regression model to address selection due to different response probabilities. I'm calculating the weights as Inverse Probability of ...
- 31
0
votes
0
answers
12
views
Estimation of the equivalence between 2 populations by using a Fisher's exact test
I am currently working on an article about a genetic disease caused by 3 groups of genetic mutation. A group of genetic mutation is particularly much more serious than the 2 others. In this article ...
3
votes
0
answers
68
views
When can we get unbiased estimate given biased data?
There was a recent "hot take" tweet by Andrej Karpathy (without any comment or clarification from the author):
real-world data distribution is ~N(0,1)
good dataset is ~U(-2,2)
It provoked ...
Tim♦
- 125k
2
votes
1
answer
123
views
Selection Models of Publication Bias for Multilevel Meta-analyses?
Are there any suitable selection models of publication bias for multilevel meta-analyses?
I am currently conducting a 3-level meta-analysis and trying to incorporate selection models to assess ...
- 63
1
vote
1
answer
19
views
Sample Bias in Study
I have following Study statement:
A council wishes to study the digital awareness of its resident senior population (over 65 years), so it questioned in person 50 residents randomly chosen from a ...
- 23
1
vote
1
answer
57
views
How to show how biased a statistic is in a non-random sample, knowing the parameter in the general population?
I have a convenience sample, and want to show readers how biased it is relative to the population it's taken from. It's absolutely certain that the sample is biased, and I want to give readers as many ...
- 442
1
vote
1
answer
60
views
Is it true that a larger, representative dataset is always better to use than a smaller, representative dataset?
By "representative" I mean that the data in the dataset faithfully reflects the "underlying signal" a model is trying to tap in to. Is it always true that, as long as increasing ...
- 121
0
votes
0
answers
17
views
Selection bias / Stratification - How to create a sample that represents the full population
I have built and deployed multiple models, initially starting with a heuristics model, eventually moving to various iterations of machine learning models where the target is binary.
For each model, ...
- 121
0
votes
0
answers
74
views
Heckman model: satisfying exclusion criterion with control-variables
I have a situation where I need to use a Heckman selection model to correct for selection bias. Till now, I have specified my instrument as relevant for the selection equation, and that it has no ...
- 1
0
votes
1
answer
26
views
Can I ignore these individuals without introducing bias?
I have a population that falls under 10 classes. Each individual may or may not come with a location - 83% overall have locations and a breakdown by class is:
Class
# individuals
# with location
# ...
- 113
1
vote
0
answers
40
views
What are the statistical fallacies of illusion of control?
Illusion of control* appears in gambling and events involving randomness. For example, choosing a lottery ticket which has an additional information that participant has a control of choosing, such as ...
- 2,412
0
votes
0
answers
39
views
Sample selection bias as internal bias under heterogeneous effects
I have been reading a lot about the extent to which sample selection biases that come from truncated samples (e.g., Heckman selection situations) are not only externally biased, which is more commonly ...
- 971
1
vote
1
answer
25
views
Estimating population means after a selection procedure
The problem is somewhat related to sequential procedures (e.g. Paulson; Gupta and Miescke)
I have a hundred engines generating normal random numbers with true (unknown) means $E_1,E_2,E_3...E_{100}$, ...
- 13
0
votes
0
answers
14
views
Causal inference when selection bias affects response but not treatment
I have a problem where I need to estimate how variable X (risk indicator, a 0/1 "treatment" variable) affects variable Y (customer's risky behaviour). I have full knowledge of X for all ...
- 31
1
vote
0
answers
20
views
Looking for a type of bias between "Selection bias" and "Framing"
I am looking for a word to describe a certain type of bias, something between Framing and Selection Bias.
I am analyizing a dataset with measurements and their location in Berlin. The dataset includes ...
- 11
2
votes
0
answers
60
views
effect of vaccination against covid on hospitalization - help building a causal DAG [closed]
During the covid crisis, we have seen explanations of why vaccinated people end up less hospitalized than unvaccinated people under the paradox that there were more vaccinated people hospitalized than ...
- 385
2
votes
0
answers
63
views
How to accommodate endogeneity after matching?
I am working on a field experiment where assignment to treatment vs. comparison was random, but participation uptake was not. The design is pre-post, and attrition is certainly not MCAR. This is a ...
- 394
1
vote
1
answer
184
views
What is the difference between selection bias and composition effect?
Groups we seek to compare (e.g., a treatment group and control group) may differ in ways that constrain our ability to do so. Often, potential outcomes may differ systematically across groups, such as ...
- 1,202
10
votes
3
answers
2k
views
What is it called when an experimenter discards results that are too unexpected?
There is a type of scientific error where an experimenter gets a result significantly different from prior researchers, assumes they made a mistake, and redoes the experiment until they get a more ...
3
votes
2
answers
85
views
How to avoid selection bias while updating lead scoring (predictive) model with new data
We developed a standard lead scoring model using logistic regression on couple of months worth data. The model has been working and we have been pushing only top 1/3 leads to sales team basis that. ...
- 341
0
votes
1
answer
32
views
Treatment Selection based on Treatment Effects
In many observational studies, researchers assume that the treatment decision is based on observable covariates (that is, the ignorability assumption).
In this case, there are diverse methods that can ...
- 755
1
vote
0
answers
38
views
Why does the best fit model (lower AIC) yield higher p values than models with higher AIC? [closed]
Background: I am running a model selection in R that includes 1, 2, and 3-covariate models. Each model aims to determine the effect of environmental covariates in the occupancy of different species in ...
- 11
0
votes
0
answers
18
views
How to handle retraining after model introduces bias
Here is my problem.
I am retraining a machine learning model to detect fraudulent purchases.
The training data is based on purchases and the target is whether or not they the purchase was considered ...
1
vote
0
answers
181
views
Causal Inference: Selection Bias and Endogeneity [closed]
I've encountered lots of causal inference terms and jargons (under the Neyman-Rubin potential outcome framework), and I had questions regarding their relationships:
I know that exogeneity E(e|X) = 0 ...
- 173
1
vote
0
answers
65
views
When we up-sample the training set, don't we introduce selection bias?
When doing supervised machine learning in the health or medical domains, we often have a target class that is relatively rare (e.g., prevalence 1-10% of cases). There are a few techniques we can do to ...
- 11
0
votes
2
answers
98
views
Regression with endogenous dummy variables
Suppose I have 1,000 observations on variable Y, and I regress Y on just a dummy variable (d), in two different regressions. In the first regression, I set d equal to 0 for the first 500 observations, ...
- 35
2
votes
0
answers
176
views
Which DAG would explain the lack of correlation between height and performance in NBA players?
A classic example of "selection bias" involves looking at the performance of professional basketball players. The example goes, among NBA players there is no correlation between height and ...
- 1,378
2
votes
0
answers
42
views
Using a sample of paid survey respondents to bias correct lower response rate among larger non-paid sample
I'm running a tracker survey on a website that has a low response rate of about 2%. The survey is not incentivized but the website traffic is large enough in volume I can always meet my sample target ...
- 121
1
vote
0
answers
203
views
Logistic regression for case-control studies
If I have designed a study where participants from 3 disease groups of fixed size were being sampled and suppose the three groups A, B and C are of sizes n_A=50, n_B=50 and n_C=100. Group A is a ...
- 321
2
votes
0
answers
44
views
Intended selection bias
Sampling or selection bias is often presented as something that has to be overcome, avoided, or at least appropriately considered because it's a problem otherwise. I wonder how often situations arise (...
2
votes
1
answer
35
views
Implicit Selection Effects in Mixed Models
I'm struggling with a question that I haven't seen talked about anywhere. Potentially it just illustrates some ignorance I have about how effects are estimated in mixed models.
Let's set up the ...
3
votes
1
answer
79
views
Is self-selection for a treatment a problem after all?
I have devised the following R code.
In this simulated dataset, we know that a certain treatment (e.g. a degree, an education) increases income by $ 1,000. Income is also caused by age.
In the first ...
- 1,684
2
votes
1
answer
696
views
Is it useful to implement clustered SE in the probit-type models?
For my research, I am implementing a two-stage Heckman procedure. I am working with panel data, so I was wondering if it is common and actually needed to use clustered standard errors for the first ...
- 47
0
votes
0
answers
247
views
Lagged dependent variables: Is it safe to ignore bias over a long time horizon?
Suppose I have a Poisson model which investigates the effect of a county level policy on robbery counts. Here is the basic specification:
$$
\text{log}(y_{it}) = \theta y_{i,t-1} + \sum_i\text{County}...
- 5,242
4
votes
2
answers
73
views
How would a statistician describe the problem with the figure in this publication? the solution?
I pointed out a problem with averaging values over time here https://www.researchgate.net/publication/344137839_SARS-CoV-2_binds_platelet_ACE2_to_enhance_thrombosis_in_COVID-19/comments in the ...
