INXVALID

INXVALID is an R package for validating composite indices. It provides a structured workflow for assessing redundancy, dimensionality, theoretical structure, and contribution balance among variables.

---

🚀 Key Features

• Bivariate screening for redundancy (bi_var)
• Dimensionality assessment using PCA (factor_var)
• Confirmatory factor analysis for theoretical validation (cfa_var)
• Contribution balance diagnostics (contrib_balance)

---

📦 Installation

# install.packages("devtools")
devtools::install_github("S-E-SCROGGINS/INXVALID")

---

🔄 Example Workflow

library(INXVALID)

# Example dataset
set.seed(123)
df <- data.frame(
  var1 = rnorm(100),
  var2 = rnorm(100),
  var3 = rnorm(100),
  var4 = rnorm(100)
)

# 1. Check for redundancy
bi_var(df, threshold = 0.7)

# 2. Assess dimensionality
factor_var(df, var_target = 0.9)

# 3. Confirm theoretical structure
model <- "
  f1 =~ var1 + var2
  f2 =~ var3 + var4
"

cfa_var(df, model = model)

# 4. Evaluate contribution balance
contrib_balance(df)

---

🧠 Conceptual Framework

INXVALID follows a four-step validation pipeline:

1. Redundancy Detection
   Identify highly correlated variables that may distort the index.

2. Dimensionality Assessment
   Determine the number of latent components required to explain variance.

3. Theoretical Validation
   Confirm whether variables load according to expected conceptual groupings.

4. Contribution Balance
   Evaluate whether certain variables dominate the index.

---

📊 Output

All functions in INXVALID:

• Print concise, interpretable summaries
• Return structured objects for further analysis
• Handle missing data appropriately for each method

---

⚠️ Notes

• Input data should consist of numeric variables
• Missing data handling varies by function:
  • factor_var() and cfa_var() use complete cases
  • contrib_balance() uses row-wise weight normalization

---

📌 Future Development

Planned extensions include:

• Sensitivity analysis (leave-one-out)
• Rank stability diagnostics
• Outcome validation tools

---

👤 Author

Stephen Scroggins, PhD