Factorial validity and invariance of a survey measuring psychosocial correlates of colorectal cancer screening in Ontario, Canada - A replication study

Paul Ritvo, Ronald Myers, M. L. Del Giudice, Larry Pazsat, Peter T. Campbell, R. I. Howlett, Verna Mai, Terry Sullivan, Jasmin Tiro, Linda Rabeneck

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

Psychosocial constructs have been used to predict colorectal cancer screening and are frequently targeted as intermediate outcomes in behavioral intervention studies. Few studies have conducted analyses to adequately test construct validity. The psychometric analyses undertaken with U.S. populations of 16 theory-based, colorectal cancer screening items designed to measure five factors (salience-coherence, cancer worries, perceived susceptibility, response efficacy, and social influence) are an exception. The current investigation replicates previous work by examining factor validity and invariance in a random sample of Ontario, Canada residents. A survey instrument was administered to 1,013 Ontario male (49%) and female (51%) residents randomly selected by the Canada Survey Sample. Single-group confirmatory factor analyses (CFA) assessed data fit to the proposed five-factor model for males and females separately, and then a multigroup CFA evaluated if the factor structure was invariant for men and women. The five-factor model provided good fit for both males and females. Tests for factorial invariance between sexes, however, found mixed results. χ2 difference test was significant (P = 0.025); however, ΔRMSEA = 0.0001. Factor loadings were similar by sex except for two social influence items, with item frequency distributions suggesting an extreme response style, in females, on these items. Overall, the single-group and multigroup CFA results support factorial validity and partial invariance of the five-factor model first identified in the U.S. populations. The items can be used to evaluate and compare psychosocial correlates across U.S. and Canadian samples. Additional research is needed to show invariance for other ethnocultural and national subgroups.

Original languageEnglish (US)
Pages (from-to)3279-3283
Number of pages5
JournalCancer Epidemiology Biomarkers and Prevention
Volume17
Issue number11
DOIs
StatePublished - Nov 2008

Fingerprint

Ontario
Early Detection of Cancer
Canada
Colorectal Neoplasms
Statistical Factor Analysis
Population Dynamics
Psychometrics
Surveys and Questionnaires
Research
Population
Neoplasms

ASJC Scopus subject areas

  • Epidemiology
  • Oncology

Cite this

Factorial validity and invariance of a survey measuring psychosocial correlates of colorectal cancer screening in Ontario, Canada - A replication study. / Ritvo, Paul; Myers, Ronald; Del Giudice, M. L.; Pazsat, Larry; Campbell, Peter T.; Howlett, R. I.; Mai, Verna; Sullivan, Terry; Tiro, Jasmin; Rabeneck, Linda.

In: Cancer Epidemiology Biomarkers and Prevention, Vol. 17, No. 11, 11.2008, p. 3279-3283.

Research output: Contribution to journalArticle

Ritvo, Paul ; Myers, Ronald ; Del Giudice, M. L. ; Pazsat, Larry ; Campbell, Peter T. ; Howlett, R. I. ; Mai, Verna ; Sullivan, Terry ; Tiro, Jasmin ; Rabeneck, Linda. / Factorial validity and invariance of a survey measuring psychosocial correlates of colorectal cancer screening in Ontario, Canada - A replication study. In: Cancer Epidemiology Biomarkers and Prevention. 2008 ; Vol. 17, No. 11. pp. 3279-3283.
@article{ba4c942f9eca4c7181d5a3c0fc7ea3f4,
title = "Factorial validity and invariance of a survey measuring psychosocial correlates of colorectal cancer screening in Ontario, Canada - A replication study",
abstract = "Psychosocial constructs have been used to predict colorectal cancer screening and are frequently targeted as intermediate outcomes in behavioral intervention studies. Few studies have conducted analyses to adequately test construct validity. The psychometric analyses undertaken with U.S. populations of 16 theory-based, colorectal cancer screening items designed to measure five factors (salience-coherence, cancer worries, perceived susceptibility, response efficacy, and social influence) are an exception. The current investigation replicates previous work by examining factor validity and invariance in a random sample of Ontario, Canada residents. A survey instrument was administered to 1,013 Ontario male (49{\%}) and female (51{\%}) residents randomly selected by the Canada Survey Sample. Single-group confirmatory factor analyses (CFA) assessed data fit to the proposed five-factor model for males and females separately, and then a multigroup CFA evaluated if the factor structure was invariant for men and women. The five-factor model provided good fit for both males and females. Tests for factorial invariance between sexes, however, found mixed results. χ2 difference test was significant (P = 0.025); however, ΔRMSEA = 0.0001. Factor loadings were similar by sex except for two social influence items, with item frequency distributions suggesting an extreme response style, in females, on these items. Overall, the single-group and multigroup CFA results support factorial validity and partial invariance of the five-factor model first identified in the U.S. populations. The items can be used to evaluate and compare psychosocial correlates across U.S. and Canadian samples. Additional research is needed to show invariance for other ethnocultural and national subgroups.",
author = "Paul Ritvo and Ronald Myers and {Del Giudice}, {M. L.} and Larry Pazsat and Campbell, {Peter T.} and Howlett, {R. I.} and Verna Mai and Terry Sullivan and Jasmin Tiro and Linda Rabeneck",
year = "2008",
month = "11",
doi = "10.1158/1055-9965.EPI-08-0241",
language = "English (US)",
volume = "17",
pages = "3279--3283",
journal = "Cancer Epidemiology Biomarkers and Prevention",
issn = "1055-9965",
publisher = "American Association for Cancer Research Inc.",
number = "11",

}

TY - JOUR

T1 - Factorial validity and invariance of a survey measuring psychosocial correlates of colorectal cancer screening in Ontario, Canada - A replication study

AU - Ritvo, Paul

AU - Myers, Ronald

AU - Del Giudice, M. L.

AU - Pazsat, Larry

AU - Campbell, Peter T.

AU - Howlett, R. I.

AU - Mai, Verna

AU - Sullivan, Terry

AU - Tiro, Jasmin

AU - Rabeneck, Linda

PY - 2008/11

Y1 - 2008/11

N2 - Psychosocial constructs have been used to predict colorectal cancer screening and are frequently targeted as intermediate outcomes in behavioral intervention studies. Few studies have conducted analyses to adequately test construct validity. The psychometric analyses undertaken with U.S. populations of 16 theory-based, colorectal cancer screening items designed to measure five factors (salience-coherence, cancer worries, perceived susceptibility, response efficacy, and social influence) are an exception. The current investigation replicates previous work by examining factor validity and invariance in a random sample of Ontario, Canada residents. A survey instrument was administered to 1,013 Ontario male (49%) and female (51%) residents randomly selected by the Canada Survey Sample. Single-group confirmatory factor analyses (CFA) assessed data fit to the proposed five-factor model for males and females separately, and then a multigroup CFA evaluated if the factor structure was invariant for men and women. The five-factor model provided good fit for both males and females. Tests for factorial invariance between sexes, however, found mixed results. χ2 difference test was significant (P = 0.025); however, ΔRMSEA = 0.0001. Factor loadings were similar by sex except for two social influence items, with item frequency distributions suggesting an extreme response style, in females, on these items. Overall, the single-group and multigroup CFA results support factorial validity and partial invariance of the five-factor model first identified in the U.S. populations. The items can be used to evaluate and compare psychosocial correlates across U.S. and Canadian samples. Additional research is needed to show invariance for other ethnocultural and national subgroups.

AB - Psychosocial constructs have been used to predict colorectal cancer screening and are frequently targeted as intermediate outcomes in behavioral intervention studies. Few studies have conducted analyses to adequately test construct validity. The psychometric analyses undertaken with U.S. populations of 16 theory-based, colorectal cancer screening items designed to measure five factors (salience-coherence, cancer worries, perceived susceptibility, response efficacy, and social influence) are an exception. The current investigation replicates previous work by examining factor validity and invariance in a random sample of Ontario, Canada residents. A survey instrument was administered to 1,013 Ontario male (49%) and female (51%) residents randomly selected by the Canada Survey Sample. Single-group confirmatory factor analyses (CFA) assessed data fit to the proposed five-factor model for males and females separately, and then a multigroup CFA evaluated if the factor structure was invariant for men and women. The five-factor model provided good fit for both males and females. Tests for factorial invariance between sexes, however, found mixed results. χ2 difference test was significant (P = 0.025); however, ΔRMSEA = 0.0001. Factor loadings were similar by sex except for two social influence items, with item frequency distributions suggesting an extreme response style, in females, on these items. Overall, the single-group and multigroup CFA results support factorial validity and partial invariance of the five-factor model first identified in the U.S. populations. The items can be used to evaluate and compare psychosocial correlates across U.S. and Canadian samples. Additional research is needed to show invariance for other ethnocultural and national subgroups.

UR - http://www.scopus.com/inward/record.url?scp=55849092517&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=55849092517&partnerID=8YFLogxK

U2 - 10.1158/1055-9965.EPI-08-0241

DO - 10.1158/1055-9965.EPI-08-0241

M3 - Article

C2 - 18990772

AN - SCOPUS:55849092517

VL - 17

SP - 3279

EP - 3283

JO - Cancer Epidemiology Biomarkers and Prevention

JF - Cancer Epidemiology Biomarkers and Prevention

SN - 1055-9965

IS - 11

ER -