Find Range of Cronbach Alpha with a Dataset Including Missing Data
Compute Maximum Possible Alpha Value
Compute Minimum Possible Alpha Value
Compute Exact Bounds of Cronbach's Alpha via Enumeration
Compute Rough Approximation of Cronbach's Alpha Bounds
Compute Lower and Upper Bound of Cronbach's Alpha
Display All Possible Parameter Combinations for Cronbach's Alpha
Check Feasibility of Alpha Bound for Optimization Problem
Generate Bernoulli Distributed Scores Matrix with Missing Values
Solve Quadratic Programming Problem using DEoptim
Solve Quadratic Programming Problem using GA
Solve Quadratic Programming Problem using nloptr
General Solver for Quadratic Programming Problems
Provides functions to calculate the minimum and maximum possible values of Cronbach's alpha when item-level missing data are present. Cronbach's alpha (Cronbach, 1951 <doi:10.1007/BF02310555>) is one of the most widely used measures of internal consistency in the social, behavioral, and medical sciences (Bland & Altman, 1997 <doi:10.1136/bmj.314.7080.572>; Tavakol & Dennick, 2011 <doi:10.5116/ijme.4dfb.8dfd>). However, conventional implementations assume complete data, and listwise deletion is often applied when missingness occurs, which can lead to biased or overly optimistic reliability estimates (Enders, 2003 <doi:10.1037/1082-989X.8.3.322>). This package implements computational strategies including enumeration, Monte Carlo sampling, and optimization algorithms (e.g., Genetic Algorithm, Differential Evolution, Sequential Least Squares Programming) to obtain sharp lower and upper bounds of Cronbach's alpha under arbitrary missing data patterns. The approach is motivated by Manski's partial identification framework and pessimistic bounding ideas from optimization literature.