Parallel analysis with unidimensional binary data

Li Jen Weng, Chung Ping Cheng

Research output: Contribution to journalReview article

41 Citations (Scopus)

Abstract

The present simulation investigated the performance of parallel analysis for unidimensional binary data. Single-factor models with 8 and 20 indicators were examined, and sample size (50, 100, 200, 500, and 1,000), factor loading (.45, .70, and .90), response ratio on two categories (50/50, 60/40, 70/30, 80/20, and 90/10), and types of correlation coefficients (phi and tetrachoric correlations) were manipulated. The results indicated that parallel analysis performed well in identifying the number of factors. The performance improved as factor loading and sample size increased and as the percentages of responses on two categories became close. Using the 95th and 99th percentiles of the random data eigenvalues as the criteria for comparison in parallel analysis yielded higher correct rate than using mean eigenvalues.

Original languageEnglish
Pages (from-to)791-810
Number of pages20
JournalEducational and Psychological Measurement
Volume65
Issue number5
DOIs
Publication statusPublished - 2005 Oct

Fingerprint

Binary Data
Sample Size
Eigenvalue
Percentile
Factor Models
Correlation coefficient
performance
Percentage
simulation
present
Simulation

All Science Journal Classification (ASJC) codes

  • Education
  • Developmental and Educational Psychology
  • Applied Psychology
  • Applied Mathematics

Cite this

@article{7a0fbb6a18174fbbb5585c8fe9001a5f,
title = "Parallel analysis with unidimensional binary data",
abstract = "The present simulation investigated the performance of parallel analysis for unidimensional binary data. Single-factor models with 8 and 20 indicators were examined, and sample size (50, 100, 200, 500, and 1,000), factor loading (.45, .70, and .90), response ratio on two categories (50/50, 60/40, 70/30, 80/20, and 90/10), and types of correlation coefficients (phi and tetrachoric correlations) were manipulated. The results indicated that parallel analysis performed well in identifying the number of factors. The performance improved as factor loading and sample size increased and as the percentages of responses on two categories became close. Using the 95th and 99th percentiles of the random data eigenvalues as the criteria for comparison in parallel analysis yielded higher correct rate than using mean eigenvalues.",
author = "Weng, {Li Jen} and Cheng, {Chung Ping}",
year = "2005",
month = "10",
doi = "10.1177/0013164404273941",
language = "English",
volume = "65",
pages = "791--810",
journal = "Educational and Psychological Measurement",
issn = "0013-1644",
publisher = "SAGE Publications Inc.",
number = "5",

}

Parallel analysis with unidimensional binary data. / Weng, Li Jen; Cheng, Chung Ping.

In: Educational and Psychological Measurement, Vol. 65, No. 5, 10.2005, p. 791-810.

Research output: Contribution to journalReview article

TY - JOUR

T1 - Parallel analysis with unidimensional binary data

AU - Weng, Li Jen

AU - Cheng, Chung Ping

PY - 2005/10

Y1 - 2005/10

N2 - The present simulation investigated the performance of parallel analysis for unidimensional binary data. Single-factor models with 8 and 20 indicators were examined, and sample size (50, 100, 200, 500, and 1,000), factor loading (.45, .70, and .90), response ratio on two categories (50/50, 60/40, 70/30, 80/20, and 90/10), and types of correlation coefficients (phi and tetrachoric correlations) were manipulated. The results indicated that parallel analysis performed well in identifying the number of factors. The performance improved as factor loading and sample size increased and as the percentages of responses on two categories became close. Using the 95th and 99th percentiles of the random data eigenvalues as the criteria for comparison in parallel analysis yielded higher correct rate than using mean eigenvalues.

AB - The present simulation investigated the performance of parallel analysis for unidimensional binary data. Single-factor models with 8 and 20 indicators were examined, and sample size (50, 100, 200, 500, and 1,000), factor loading (.45, .70, and .90), response ratio on two categories (50/50, 60/40, 70/30, 80/20, and 90/10), and types of correlation coefficients (phi and tetrachoric correlations) were manipulated. The results indicated that parallel analysis performed well in identifying the number of factors. The performance improved as factor loading and sample size increased and as the percentages of responses on two categories became close. Using the 95th and 99th percentiles of the random data eigenvalues as the criteria for comparison in parallel analysis yielded higher correct rate than using mean eigenvalues.

UR - http://www.scopus.com/inward/record.url?scp=27144446688&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=27144446688&partnerID=8YFLogxK

U2 - 10.1177/0013164404273941

DO - 10.1177/0013164404273941

M3 - Review article

AN - SCOPUS:27144446688

VL - 65

SP - 791

EP - 810

JO - Educational and Psychological Measurement

JF - Educational and Psychological Measurement

SN - 0013-1644

IS - 5

ER -