Why It Matters:

• Confounders can create false associations or hide true associations between variables.

• Properly addressing confounding improves the credibility of your study.

• Helps researchers design better studies and interpret results accurately.

How to Identify and Handle Confounding:

• Use stratification or matching to control confounders in study design.

• Apply multivariable regression models to adjust for confounders in analysis.

• Always discuss potential confounders in the limitations of your research.

Examples:

1. Studying the link between coffee consumption and heart disease without accounting for smoking habits.

2. Examining exercise and blood pressure levels without considering age as a factor.

3. Investigating medication effects on diabetes outcomes without adjusting for diet.

Statistical bias isn’t just a technical error—it’s the silent thief that sneaks into your research and distorts your findings. It occurs when the data collected, analyzed, or interpreted leads to conclusions that consistently lean in one direction, away from the truth. It's not about random mistakes—it’s a systematic problem.

⚠️ Why Should Researchers Care?

Bias can make even the most sophisticated analysis completely unreliable. Imagine putting hours into a study, only to realize your sample was flawed or your method favored one outcome. That’s how bias quietly undermines credibility, trust, and real-world impact.

🧠 Types You Might Not Notice

Selection bias creeps in when your sample doesn’t represent the population. Measurement bias happens when your tools or methods aren’t accurate. And then there’s publication bias—where only “positive” results get published, skewing the entire scientific conversation. These aren’t rare—they’re common traps.

🛡️ Can You Eliminate Bias?

Not completely. But you can detect it, reduce it, and account for it. Techniques like randomization, blinding, and sensitivity analysis aren’t just good practice—they’re your best defense. Awareness is the first step, action is the next.

💬 Let’s Talk

Have you ever encountered bias in your research? How did you deal with it—or did you notice it only after the results came in? Share your experience, tips, or questions below. Let’s learn from each other.

2. Installing and Setting Up Stata
Start by purchasing or accessing an institutional license for Stata. After installation, familiarize yourself with the interface: the Command window, Results window, Variables window, and Data Editor. This setup helps streamline your workflow.

3. Importing Data
Stata supports importing data from various formats such as Excel, CSV, and text files. Use commands like import excel or insheet to bring your data into Stata. Always check for missing values or errors after importing to ensure data integrity.

4. Basic Data Management
Learn essential commands like list, browse, summarize, describe, and generate for exploring and managing your dataset. Understanding how to clean and manipulate data is crucial for producing accurate and meaningful results.

5. Descriptive Statistics
Start with descriptive statistics to summarize your data. Commands like tabulate, summarize, and mean provide an overview of your variables and distributions. These steps lay the foundation for more advanced analyses.

6. Performing Statistical Tests
Stata offers a wide range of statistical tests: t-tests, chi-square tests, ANOVA, regression models, and more. For example, use ttest for comparing means and regress for linear regression. Always verify assumptions before running tests.

7. Graphing and Visualization
Visualizing data helps in understanding patterns and communicating findings. Use commands like graph bar, scatter, and twoway to create informative charts. Stata also allows customization to suit publication standards.

8. Saving and Exporting Results
Save your work frequently with save and export results using outreg2 or export excel. Proper documentation of your outputs ensures reproducibility and ease of reporting for medical journals.

9. Learning Resources
Enhance your Stata skills with online tutorials, official Stata documentation, and community forums like Statalist. Many universities also offer workshops or courses specifically tailored for medical researchers.

10. Conclusion
Mastering Stata takes practice, but it is a valuable tool for medical research. Start with small projects, gradually explore advanced features, and leverage the supportive Stata community to grow your expertise.

Additionally, the latest version of moremata is required. You can update it by running:

ssc install moremata, replace

The moremata package provides Mata functions essential for ebalfit to operate correctly.

Key Features of ebalfit:

• Model Estimation: Estimates an entropy balancing model, similar to a logit model, and displays coefficients with standard errors.​

• Weight Generation: After estimation, use the predict command to generate balancing weights or the implied propensity scores.​

• Variance Estimation: Utilizes influence functions for variance estimation, which can be stored for adjusting standard errors of statistics computed using the balancing weights.

Why use IHS?

1. Handles Zero and Negative Values: Unlike the natural logarithm (ln(x)), which is undefined for zero and negative numbers, the IHS transformation can be applied to the entire real line.
2. Similar to Log Transformation: For large values of x, the IHS transformation behaves similarly to the log transformation (ln(x)), making it useful for dealing with right-skewed distributions.
3. Reduces Skewness: It helps in normalizing highly skewed data, improving the interpretability and efficiency of regression models.

STATA CODE:

gen ihs_var = ihstrans(variable)

gen ihs_income = ihstrans(income)

This sample code will create a new variable ihs_income that contains the IHS-transformed values of income VARIAable. 

Matching methods are used to reduce selection bias in observational studies by pairing treated and control units based on their propensity scores. The most common matching techniques in Stata include:

1. Nearest Neighbor (NN) Matching
2. Caliper Matching
3. Kernel Matching
4. Radius Matching
5. Stratification/Interval Matching

Each method has its own advantages and trade-offs.

1. Nearest Neighbor (NN) Matching

Concept

• Each treated unit is matched with the control unit that has the closest propensity score.
• Can be done with or without replacement.
• Can specify 1-to-1 or 1-to-many matching.
  mplementation in Stata (psmatch2)
  psmatch2 treatment covariates, outcome(outcome_var) neighbor(1)
  
  
• neighbor(1): Matches each treated unit to the single closest control unit.

2. Caliper Matching

Concept

• Similar to NN matching, but imposes a maximum allowed difference in propensity scores (the "caliper").
• Helps avoid poor matches by ensuring that matched units are sufficiently similar.
  
  psmatch2 treatment covariates, outcome(outcome_var) neighbor(1) caliper(0.05)
  caliper(0.05): Ensures that control units are within 0.05 propensity score of the treated unit.
  
  

  3. Kernel Matching

  Concept

  • Instead of picking one nearest neighbor, Kernel Matching uses multiple control units and assigns them weights based on their closeness.
  • Treated unit outcome is compared to a weighted average of the control group.
    
    psmatch2 treatment covariates, outcome(outcome_var) kernel
    
    
    4. Radius Matching

    Concept :

    • Each treated unit is matched with all control units within a certain distance (radius) in propensity score space.

     psmatch2 treatment covariates, outcome(outcome_var) radius caliper(0.05)

    caliper(0.05): Includes all control units within 0.05 propensity score.
    
    
    

5. Stratification (Interval) Matching

Concept

• The propensity score is divided into intervals (strata), and treated/control units are compared within each stratum.
• Works similarly to coarsened exact matching (CEM).
  
  psmatch2 treatment covariates, outcome(outcome_var) strata(5)
  strata(5): Divides the propensity score into 5 groups.
  
  

  Comparison Table of Matching Methods

  Method	Matching Type	Strengths	Limitations
  Nearest Neighbor (NN)	1-to-1 or 1-to-many	Easy to interpret, real units used	Bad matches possible, may drop many controls
  Caliper Matching	NN with a threshold	Prevents poor matches	May drop treated units
  Kernel Matching	Weighted average of controls	Uses all data, reduces variance	Computationally intensive
  Radius Matching	Multiple matches within a range	More control units per treated	Sample size varies
  Stratification Matching	Groups by propensity score strata	Retains most data, simple to apply	Assumes similarity within strata

  
  
  
  

Which Matching Method Should You Use?

• If you want simple matching → Nearest Neighbor Matching
• If you want to avoid poor matches → Caliper Matching
• If you have a large control group → Kernel Matching
• If you want a balance between NN and Kernel → Radius Matching
• If you prefer a stratified approach → Stratification Matching

Summary & Conclusion

Propensity Score Matching (PSM) methods in Stata help reduce selection bias in observational studies by balancing treated and control groups based on their propensity scores. Nearest Neighbor Matching (NN) is the simplest method, pairing each treated unit with the closest control, but may lead to poor matches. Caliper Matching improves upon NN by restricting matches within a specified range, preventing extreme differences. Kernel Matching and Radius Matching use multiple control units per treated unit, reducing variance but requiring careful selection of bandwidth or caliper. Stratification Matching divides the sample into propensity score bins, ensuring comparability within each group. Choosing the right method depends on the dataset and research goals—NN is intuitive but risky, Caliper reduces bias at the cost of sample size, Kernel and Radius improve precision but are computationally complex, and Stratification offers a structured approach. Regardless of the method, researchers should check balance and common support to validate results. 🚀

1. IPW (Inverse Probability Weighting)

• Concept: Weights each observation by the inverse of the probability of receiving treatment, based on the estimated propensity score.
• Use case: Creates a pseudo-population where treatment assignment is independent of covariates.
• Implementation in Stata:
  Estimate propensity scores using logit or probit 
logit treatment covariates
predict ps, pr
• Strengths
  • Uses the entire dataset (no need to drop unmatched units).
  • Can handle high-dimensional covariates better than matching.
• Limitations:
• Sensitive to extreme weights (requires trimming).

2. psmatch (Official Stata Module)

• Concept: Performs nearest neighbor matching based on estimated propensity scores.
• Use case: Compares treated and control units by selecting similar observations.
• Implementation:
  psmatch treatment covariates, neighbor(1)
  
  

  • Strengths:
    • Provides a straightforward approach to matching.
    • Available in newer Stata versions (Stata 16+).
  • Limitations:
    • Less flexible than psmatch2 in terms of options.

   

3. psmatch2 (User-Written Module)

• Concept: An advanced matching method that extends psmatch with additional features.
• Use case: Provides more flexible matching, including nearest neighbor, kernel, and radius matching.
• Installation:
  ssc install psmatch2
  
  

  • Strengths:
    • More matching options (e.g., multiple neighbors, caliper matching).
    • Generates additional diagnostics.
  • Limitations:
    • Requires installation (not built-in).
      
      

      Comparison Table

      Method	Type	Strengths	Limitations
      IPW	Weighting	Uses full sample, better for high-dimensional data	Sensitive to extreme weights
      psmatch	Matching	Official Stata command, simple implementation	Less flexible than psmatch2
      psmatch2	Matching	More flexible, advanced matching options	Requires installation

      
      
      
      

  Which One to Use?

  • Use IPW if you want to keep all observations and reduce selection bias using weighting.
  • Use psmatch if you prefer a simple, built-in solution.
  • Use psmatch2 if you need more advanced matching techniques and diagnostics.

   

The teffects command in Stata is used to estimate treatment effects in observational studies. It provides various methods to adjust for confounding and selection bias when estimating causal effects.

Example Usage in Stata

Suppose we have the following variables:

• Treatment variable: treat (1 = treated, 0 = control)
• Outcome variable: y
• Covariates: x1, x2, x3



  Types of t-tests

  • One-Sample t-test
    This test compares the mean of a single group to a known value or population mean.
    ttest varname == value

  • Two-Sample t-test
    This compares the means of two independent groups to determine if they differ significantly.
    ttest varname, by(groupvar)

  • Paired t-test
    Used when you have paired data, typically before-and-after measurements on the same subjects.
    ttest var1 == var2
    
    Fixed Effects Model

    A fixed effects model controls for unobserved characteristics that vary across units but are constant over time. This is useful when the differences between units are correlated with the independent variables.
    In Stata, to run a fixed effects model, use the xtreg command with the fe option.
    xtreg y x1 x2 x3, fe
    Here, y is the dependent variable, and x1, x2, and x3 are independent variables. The fe option specifies the fixed effects model.
    
    

    Random Effects Model

    The random effects model assumes that the unobserved differences between units are not correlated with the independent variables. It is more efficient than the fixed effects model if this assumption holds true.
    To estimate a random effects model in Stata, use the xtreg command with the re option.
    xtreg y x1 x2 x3, re

    
    
    Interpreting the Results

    After running a t-test, Stata provides an output that includes:

    • t-value: The test statistic.
    • p-value: The probability that the observed difference is due to random chance.
    • Confidence Interval: The range within which the true population mean difference lies.

    Conclusion

    t-tests are powerful statistical tools for comparing group means. Stata provides a straightforward approach to performing these tests, but it is essential to ensure your data meets the assumptions of the test for accurate results.Using fixed effects or random effects models in Stata can help you analyze panel data and account for time or unit-specific unobserved factors. When you add time dummies, you're accounting for time effects in your model, which could be what you're referring to by t-effects.

     

• Retains all observations (unlike PSM).
•  Less sensitive to poor matches

• PSM pairs individuals in the treatment and control groups based on similar propensity scores.
• After matching, treatment effects are estimated using only matched data.

Strengths of PSM:

•  Ensures high comparability between treated and control groups.
• Does not rely on extrapolation, reducing model dependence.

Key Differences: IPW vs. PSM

SUMMERY: 
Inverse Probability Weighting (IPW) and Propensity Score Matching (PSM) are both methods for addressing confounding in observational studies using propensity scores. IPW assigns weights to individuals based on the inverse of their probability of receiving treatment, ensuring all observations are retained and enabling the estimation of the Average Treatment Effect (ATE). However, it can suffer from instability due to extreme weights. In contrast, PSM matches treated and control units based on similar propensity scores, improving comparability but discarding unmatched observations, which can reduce sample size. While IPW is better for retaining data and handling high-dimensional confounders, PSM is more intuitive and less dependent on model assumptions. The choice depends on study goals—IPW for ATE estimation with full data and PSM for better-matched groups, though a combination of both can enhance robustness.
 
Final Recommendation:

• Use IPW if you want to retain all observations and estimate ATE.
• Use PSM if you want a well-matched treatment/control group with fewer model assumptions.
• Consider combining both for robustness.

1. Multivariate Regression in Stata

Overview

• Multivariate regression (typically logistic or linear regression) adjusts for confounders by including them as covariates in a regression model.
• Used when treatment assignment is not random but confounders can be directly included in the model.

Stata Command for Multivariate Regression

Example: Estimating the effect of treatment (treat) on outcome (y), adjusting for covariates (x1, x2, x3):

Linear Regression (Continuous Outcome)
reg y treat x1 x2 x3, robust

Logistic Regression (Binary Outcome)
logit y treat x1 x2 x3, robust

2. Propensity Score Matching (PSM) in Stata

Overview

• PSM estimates the probability (propensity score) of receiving the treatment, then matches treated and untreated individuals with similar scores.
• Reduces selection bias by creating comparable treatment/control groups.

Steps in PSM

1. Estimate the propensity score (logistic regression predicting treatment).
   logit treat x1 x2 x3
   predict pscore
   
   
2. Match individuals (1:1, 1:N, nearest neighbor, caliper, etc.).
   ssc install psmatch2
   psmatch2 treat x1 x2 x3, out(y) neighbor(1) caliper(0.05)
   
   
3. Check balance (assess covariate distributions between groups).
   pstest x1 x2 x3, graph
   
   
4. Estimate treatment effect on the matched sample.
   teffects psmatch (y) (treat x1 x2 x3), atet

Key Differences: Multivariate Regression vs. PSM

Summary

• Use multivariate regression when sample size is small or when you want to adjust for many covariates.
• Use PSM when selection bias is strong and you want a matched control group.
• Combining PSM with regression provides better adjustment.