MedModr X M Y W
Mediation, Moderation & Conditional Process Tool
Initializing modules...
Path Diagram
Simple Slopes

Categorical Variable Encoding

Loading categorical variables...
Statistical Software

Mediation, Moderation &
Conditional Process Analysis
done right.

A free, user-friendly, open-source application for mediation, moderation, and conditional process analysis. Runs entirely in your web browser with no installation required and no internet connection needed. Use it offline on any device, anywhere.

1 - 92
Analysis Types (Models)
OLS + Logistic
Regression
b + β
Both Coefficients
Bootstrap
Percentile + BC
Key Features
Automatic Logistic Regression
When the outcome variable (Y) is binary (0/1), the system automatically switches to logistic regression with log-odds coefficients, model fit statistics (McFadden R², Cox-Snell, Nagelkerke), and χ² test.
Data Screening & Diagnostics
Outlier detection (Z-score, IQR, MAD, Percentile), normality tests (skewness, kurtosis), correlation matrix (Pearson, Spearman, Kendall), sample size recommendations, histograms, and Q-Q plots.
Missing Data Handling
Multiple Imputation (PMM), Mean/Median/Mode imputation, Hot Deck, Regression imputation, EM Algorithm, and Listwise deletion.
Categorical Variable Encoding
Automatic detection and dummy coding of categorical variables with user-selectable reference categories.
Data Import & Export
CSV/Excel import, multiple sheet support for Excel, data preview with summary statistics, export cleaned data as CSV, export results as Word, PDF, and HTML.
Data Editor with Recode
Edit data directly in the browser. Reverse code variables (e.g., 1↔5, 2↔4) and recode values (text or numeric) into new columns or replace existing data.
Bootstrap Confidence Intervals
Percentile and Bias-Corrected (BC) bootstrap methods for indirect effects. Adjustable number of bootstrap samples (1,000, 5000, and 10,000).
Path Diagrams
Interactive SVG path diagrams for analysis types. Customizable node shapes, line thickness, colors, and export as PNG/SVG.
Simple Slopes Plots
Simple slopes plots for moderation and moderated mediation. Customizable colors, labels, legend, and high-resolution download.
Publication-Ready Tables
Regression tables include unstandardized (b) and standardized (β) coefficients, SE, t/z, p-values, confidence intervals, and VIF for multicollinearity diagnostics.
Mean Centering Options
Choose between No centering, Continuous-only centering (preserves binary variables), or All variables centered.
Moderator Values
Switch between two probing strategies for moderation and moderated mediation models. Use Mean ± 1 SD for standard simple slopes, or Percentile (16/50/84%) for robust probing when the moderator is skewed or non-normally distributed.
Multiple Mediators Support
Parallel mediation supports multiple mediators simultaneously with individual and total indirect effects.
Offline & Mobile Ready
Works completely offline. Responsive design works on desktop, tablet, and mobile devices.
Available Analysis Types (Models)
W moderates X→Y (Model 1)
Tests if the X→Y relationship depends on moderator W (X×W interaction). Includes simple slopes plot and conditional effects.
W & Z moderate X→Y (Model 2)
Tests if two moderators (W and Z) moderate the X→Y relationship (X×W + X×Z interactions).
W & Z jointly moderate X→Y (Model 3)
Tests if the X→Y relationship is moderated by a three-way interaction X × W × Z → Y. Includes all two-way interactions.
M mediates X→Y (Model 4)
Tests if X → M → Y transmits effect through a single mediator. Includes Sobel test and bootstrap CIs.
M₁, M₂,... mediate X→Y in parallel (Model 4)
Tests if multiple independent mediators X → {M₁...Mₖ} → Y simultaneously transmit the effect of X on Y. Supports up to 10 mediators.
W moderates X→Y (Model 5)
Tests if moderator W affects only the direct X→Y path (c'-path) in a mediation model. Uses one mediator.
M₁→M₂→... mediate X→Y in serial (Model 6)
Tests if X affects Y through a causal chain X → M₁ → M₂ → Y with three indirect paths. Supports up to 10 mediators.
W moderates X→M (Model 7)
Tests if the indirect effect depends on W moderating the X→M path (a-path) in a mediation model. Uses one mediator.
W moderates X→M & X→Y (Model 8)
Tests if the indirect effect depends on W moderating both the X→M path and the direct X→Y path. Uses one mediator.
W & Z moderate X→M (Model 9)
Tests if W and Z moderate the X→M path simultaneously with separate moderators. Uses one mediator.
W & Z moderate X→M & X→Y (Model 10)
Tests if W and Z moderate both the X→M path and the direct X→Y path. Uses one mediator.
W & Z jointly moderate X→M (Model 11)
Tests if W and Z jointly moderate the X→M path with a three-way interaction; Y is predicted by X and M. Uses one mediator.
W & Z jointly moderate X→M & X→Y (Model 12)
Tests if W and Z jointly moderate both the X→M path and the direct X→Y path with three-way interactions. Uses one mediator.
W & Z jointly moderate X→M; W moderates X→Y (Model 13)
Tests if W and Z jointly moderate X→M while W moderates X→Y. Uses one mediator.
W moderates M→Y (Model 14)
Tests if the indirect effect depends on W moderating the M→Y path (b-path) in a mediation model. Uses one mediator.
W moderates M→Y & X→Y (Model 15)
Tests if W moderates both the M→Y path and the direct X→Y path. Uses one mediator.
W & Z moderate M→Y (Model 16)
Tests if W and Z moderate the M→Y path; the direct X→Y path is unmoderated. Uses one mediator.
W & Z moderate M→Y & X→Y (Model 17)
Tests if W and Z moderate both the M→Y path and the direct X→Y path. Uses one mediator.
W & Z jointly moderate M→Y (Model 18)
Tests if W and Z jointly moderate the M→Y path with a three-way interaction; X→Y is unmoderated. Uses one mediator.
W & Z jointly moderate M→Y & X→Y (Model 19)
Tests if W and Z jointly moderate both the M→Y path and the direct X→Y path with three-way interactions. Uses one mediator.
W & Z jointly moderate M→Y; W moderates X→Y (Model 20)
Tests if W and Z jointly moderate M→Y while W moderates X→Y. Uses one mediator.
W moderates X→M; Z moderates M→Y (Model 21)
Tests if W moderates the X→M path while Z moderates the M→Y path. Uses one mediator.
W moderates X→M & X→Y; Z moderates M→Y (Model 22)
Tests if W moderates X→M and X→Y while Z moderates M→Y. Uses one mediator.
W moderates X→M; Z moderates M→Y & X→Y (Model 28)
Tests if W moderates X→M while Z moderates both M→Y and X→Y. Uses one mediator.
W moderates X→M & X→Y; Z moderates X→M & X→Y (Model 29)
Tests if both W and Z moderate the X→M and X→Y paths. Uses one mediator.
W moderates X→M & M→Y (Model 58)
Tests if the indirect effect depends on W moderating both the X→M and M→Y paths. Uses one mediator.
W moderates X→M, M→Y & X→Y (Model 59)
Tests if W moderates all three paths (X→M, M→Y, and X→Y) in a full moderated mediation model. Uses one mediator.
W moderates X→M & M→Y; Z moderates X→M (Model 60)
Tests if W moderates X→M and M→Y while Z moderates X→M. Uses one mediator.
W moderates X→M, M→Y & X→Y; Z moderates X→M (Model 61)
Tests if W moderates X→M, M→Y, and X→Y while Z moderates X→M. Uses one mediator.
W moderates X→M & M→Y; Z moderates X→M & X→Y (Model 62)
Tests if W moderates X→M and M→Y while Z moderates X→M and X→Y. Uses one mediator.
W moderates X→M, M→Y & X→Y; Z moderates X→M & X→Y (Model 63)
Tests if W moderates X→M, M→Y, and X→Y while Z moderates X→M and X→Y. Uses one mediator.
W moderates X→M & M→Y; Z moderates M→Y (Model 64)
Tests if W moderates X→M and M→Y while Z moderates M→Y. Uses one mediator.
W moderates X→M, M→Y & X→Y; Z moderates M→Y (Model 65)
Tests if W moderates X→M, M→Y, and X→Y while Z moderates M→Y. Uses one mediator.
W moderates X→M & M→Y; Z moderates M→Y & X→Y (Model 66)
Tests if W moderates X→M and M→Y while Z moderates M→Y and X→Y. Uses one mediator.
W moderates X→M, M→Y & X→Y; Z moderates M→Y & X→Y (Model 67)
Tests if W moderates X→M, M→Y, and X→Y while Z moderates M→Y and X→Y. Uses one mediator.
W & Z jointly moderate X→M; W moderates M→Y (Model 68)
Tests if W and Z jointly moderate X→M while W moderates M→Y. Uses one mediator.
W & Z jointly moderate X→M & X→Y; W moderates M→Y (Model 69)
Tests if W and Z jointly moderate X→M and X→Y while W moderates M→Y. Uses one mediator.
W moderates X→M; W & Z jointly moderate M→Y (Model 70)
Tests if W moderates X→M while W and Z jointly moderate M→Y. Uses one mediator.
W moderates X→M; W & Z jointly moderate M→Y & X→Y (Model 71)
Tests if W moderates X→M while W and Z jointly moderate M→Y and X→Y. Uses one mediator.
W & Z jointly moderate X→M & M→Y (Model 72)
Tests if W and Z jointly moderate both the X→M and M→Y paths. Uses one mediator.
W & Z jointly moderate X→M, M→Y & X→Y (Model 73)
Tests if W and Z jointly moderate all three paths: X→M, M→Y, and X→Y. Uses one mediator.
W & Z moderate X→M & M→Y (Model 75)
Tests if W and Z moderate both the X→M path and the M→Y path. Uses one mediator.
W & Z moderate X→M, M→Y & X→Y (Model 76)
Tests if W and Z moderate all three paths: X→M, M→Y, and X→Y. Uses one mediator.
M₁, M₂, M₃,... mediate X→Y in parallel & serial (Model 80)
Tests if multiple mediators transmit the effect of X on Y through both parallel and serial paths. Supports up to 6 mediators. Includes all indirect effects and bootstrap CIs.
M₁, M₂, M₃,... mediate X→Y in parallel & serial (Model 81)
Tests a combined parallel-serial mediation where M₁ is the first-stage mediator predicted by X, and every subsequent mediator is predicted by both X and M₁. Supports up to 6 mediators with all indirect effects and bootstrap CIs.
M₁, M₂, M₃, & M₄ mediate X→Y in parallel & serial (Model 82)
Tests a four-mediator serial-parallel model where M₁ and M₃ are parallel mediators predicted by X, M₂ is predicted by X and M₁, and M₄ is predicted by X and M₃. Requires exactly 4 mediators. Includes all six indirect effects and bootstrap CIs.
W moderates X→M₁ in serial mediation (Model 83)
Tests if W moderates the first stage of serial mediation (X→M₁→M₂→Y). Supports 2–3 mediators.
W moderates X→M₁ and X→M₂ in serial mediation (Model 84)
Tests if W moderates the first stage (X→M₁) and the second stage (M₁→M₂) of serial mediation. Supports 2–3 mediators.
W moderates X→M₁, X→M₂, and X→Y in serial mediation (Model 85)
Tests if W moderates the first stage, second stage, and direct path in a serial mediation model. Supports 2–3 mediators.
W moderates X→M₁ and X→Y in serial mediation (Model 86)
Tests if W moderates the first stage (X→M₁) and the direct X→Y path in serial mediation. Supports 2–3 mediators.
W moderates M₂→Y in serial mediation (Model 87)
Tests if W moderates the final mediator→outcome path (M₂→Y) in serial mediation. Supports 2–3 mediators.
W moderates M₁→Y and M₂→Y in serial mediation (Model 88)
Tests if W moderates all mediator→outcome paths (M₁→Y and M₂→Y) in serial mediation. Supports 2–3 mediators.
W moderates M₁→Y, M₂→Y, and X→Y in serial mediation (Model 89)
Tests if W moderates all mediator→outcome paths and the direct X→Y path in serial mediation. Supports 2–3 mediators.
W moderates M₂→Y and X→Y in serial mediation (Model 90)
Tests if W moderates the final mediator→outcome path (M₂→Y) and the direct X→Y path in serial mediation. Supports 2–3 mediators.
W moderates M₁→M₂ in serial mediation (Model 91)
Tests if W moderates the intermediary stage (M₁→M₂) in a serial mediation chain. Supports 2–3 mediators.
W moderates all paths in serial mediation (Model 92)
Tests if W moderates every path in the serial mediation chain: X→M₁, M₁→M₂, M₂→Y, and X→Y. Supports 2–3 mediators.

Data Processing

Upload your data, handle missing values, and prepare for analysis.

1
Upload
2
Select Vars
3
Preview
4
Missing Data
5
Impute
6
Screening
7
Finalize
Step 1: Upload Data File
Drag & Drop or Click to Browse
Supports CSV and Excel files (.csv, .xlsx, .xls)
CSV works offline · Excel requires internet
Missing Values Configuration (optional)
Define custom missing values

Statistical Analysis

Select variables, assign to roles, and run analysis.

Variable Assignment
Available Variables Click to select, then use buttons below
X
Independent variable (s)
M
Mediator variable (s)
Y
Dependent variable (s)
W/Z
Moderator variable (s)
Cov
Covariate variable (s)
Analysis Type: Confidence level: Bootstrap samples: Bootstrap method: Decimal places:
Categorical Vars: No categorical variables detected

Author

Connect with the creator of MedModr.

👨‍💻
Mudasir Mohammed Ibrahim

Mudasir Mohammed Ibrahim is a Ghanaian nurse, researcher, data analyst, and open-source advocate 🇬🇭. He is the creator of MedModr, a free tool designed to make commonly used mediation, moderation, and conditional process analysis methods accessible to researchers, students, educators, and practitioners worldwide. His work promotes open science, reproducible research, and equitable access to high-quality statistical tools, especially in underfunded and resource-limited settings.

Connect with me