Table 1

Validity and reliability of self reported occupational histories

Authors, yearStudy populationOccupational history measureComparison measureMethod of assessing validity or reliabilityResults
Samet et al, 1978470 shipyard insulation workersYears of self reported employment as pipefitter in interviewer administered questionnaire in 1976Years of self reported employment, by same method, 1 year later- R2 for regression of two measures of years of employment- R2 = 0.95
Baumgarten et al, 19835274 cancer cases and 23 controls from Montreal, CanadaSelf reported employer, as indicated on a self administered questionnaire with an in-person interview follow up (1979–81)Quebec Pension Plan records from 1966–1978 linked by social insurance number- % agreement; agreement if start or end date within ± 1 year- 82% of person-years in agreement
- 87.9% had at least 6 of 13 years in agreement
- 64.3% of work histories had at least 12 of 13 years in agreement
Koskela et al, 198463000 living current and former employees of foundries, metal product manufacturers, and electrical devices manufacturers hired between 1950 and 1976 in FinlandSelf reported occupational histories as reported on a questionnaireOccupational histories based on employer personnel records- % agreement- 85–86% of person-years in agreement
Rosenberg et al, 19877288 workers employed at a US capacitor manufacturing plant between 1947 and 1976Self reported work history, as reported in an interviewer assisted questionnaire in 1976Company personnel records with jobs classified into 40 categories- % of person months in which job classification agreed, computed for each individual- 76.6% mean agreement for men (median >85%, range 20–99%)
- 74.3% mean agreement for women (median >80%, range 15–100%)
Stewart et al, 19878229 controls from a cohort employed between 1940 and 1979 in 2 US shipyardsSelf reported work at the shipyard, year of hire and termination, and job held longest, as reported in a telephone interview in 1983Company personnel records of all jobs held for at least 1 month- % agreement and kappa- 95% agreement on shipyard employment
- sensitivity and specificity of reporting of specific jobs- 81% on start year within ± 1 year, kappa = 0.78; 73% on termination year within ± 1 year, kappa = 0.71; 67% on job held longest, kappa = 0.65
- specificities for jobs from 0.95 to 1, sensitivities from 0 to 1, median = 0.71
Eskenazi and Pearson, 1988957 women working during pregnancy, recruited in a US prenatal clinicSelf reported primary occupation and industry reported in a self administered questionnaireSelf reported primary occupation and industry reported in a clinical interview an average of 3.3 weeks apart- % agreement- 86% perfect matches on primary occupation, 95% at least close matches
- 84% perfect matches on primary industry, 93% at least close matches
Bond et al, 198810143 males employed for at least 1 year at a US chemical plant on or after 1940Self reported work history at plant, as reported in a telephone interview in 1984Company records of 400 department titles collapsed into 50 similar work areas- % agreement- 76.2% agreement on usual work area assignments
- 50.6% agreement on number of work area assignments
Bourbonnais et al, 198811100 male employees with at least 5 years' experience in a Canadian shipyardSelf reported work history in shipyard, as reported in an in-person interviewCompany records of job titles and starting dates for each job held ≥6 months- % agreement- 89% agreement on job held longest
- 62% agreement on starting date within ± 1 year
Rona and Mosbech, 198912370 cancer cases from 8 European countriesSelf reported work history, as reported in interviewer or self administered questionnairesSelf reported work history by same method, at least one month apart- % agreement and kappa- 84% agreement on most recent job, kappa = 0.82
- 78% on previous job, kappa = 0.76
- 74% on jobs held >10 years, kappa = 0.74
Brisson et al, 199113154 females employed for at least 5 years in 5 garment factories in Montreal, CanadaSelf reported employer from 1955 to 1983, as reported in an in-person interviewPublic pension and union records- % agreement; agreement if start or end date within ± 1 year- 81% of person-years in agreement
- 89% for recent employment vs. 74% for employment ≥12 years earlier
Wärneryd et al, 19911425,586 randomly selected members of the Swedish population aged 25 to 74Self reported jobs held for at least 2 years throughout working life, as reported in an interview in 1977 or 1979–81Self reported job during week of census in 1960, 1970, 1975, and 1980, coded to the Nordic Occupational Classification- % agreement with occupational codes at 1-, 2-, and 3-digit level for each census year- 81–88% agreement at 1-digit level
- 66–78% at 2-digit level
- 52–72% at 3-digit level
Booth-Jones et al, 19981549 US carpenters from a study examining musculoskeletal symptomsSelf reported history of any of 14 specific job dutiesSelf reported history of any of 14 specific job duties, 1 to 3 weeks later- % agreement and kappa- 87.9 % agreement
- kappa = 0.73
Brower and Attfield, 199816480 coal miners enrolled in US National Study of Coal Workers' PneumoconiosisSelf reported work history, as reported in interview in 1969–1971Self reported work history, as reported in interview in 1977–1981- intraclass correlation coefficient for job tenure- ICC=0.95
- mean tenure 20.6 years based on 1st interview, 20.8 years based on 2nd