Interpolating 95Th Percentile Eigenvalues from Random Data: An Empirical Example

摘要: Selecting the "correct" number of components to retain in principal analysis is crucial. Parallel analysis, which requires a comparison eigenvalues from observed and random data, highly promising strategy for making this decision. This paper focuses on linear interpolation, has been shown be an accurate method implementing parallel analysis. Specifically, article contains tables 95th percentile data that can used when sample size between 50 500 variables 5 50. An empirical example provided illustrating direct computation, regression methods obtaining data. The given report will hopefully enable more researchers use because interpolation simple obviating Monte Carlo requirements