Research On The Revised Fujian Version Of Minnesota Multiphasic Personality Inventory (MMPI-FJ)

Objective:Main strategies of MMPI revision were item translation equivalence,development of specific in-country norms, adopting rational transformations of the raw scores both at home and abroad in the past. But because of enormous cultural difference between china and American, these methods didn't solve fundamentally the problem that validity of MMPI decreased obviously in china.To raise cross-cultural validity in china, this study followed principle of MMPI construction and selected Chinese criterion group and normal control group to screen out native scoring keys through item discrimination analysis,thus reconstructing new MMPI suitable for Chinese (called MMPI-FJ) and developing norms in Fujian province and examining its psychometric characteristics.Methods: we collected 224 psychiatric MMPI datas and medical histories, comparing sampling ratio of Fujian norm to 2005 Fujian sampling survey results. Twelve basic scales (except F scale) were reconstructed through empirical keying approach. F scale was reconstructed through concluding items in which"Yes"or"No"response rate was lower than 10%.The reliability analysis included test-retest reliability. The study adopted correlation analysis, discriminatory analysis, factor analysis and profile chart comparison to compare construct validity, validate validity and diagnostic validity between MMPI-FJ and MMPI.Results:1.Thirteen basic scales'scoring keys in MMPI-FJ differed mostly with that in MMPI.(The former were screened from 1-566 items and were 269 items in all,among them 115 items were same with the latter.The latter were screened from anterior 399 items and were 397 items in all.2.test-retest reliability: The test-retest coefficients of thirteen basic scales ranged from 0.542(Pd scale)～0.77(Si scale),averge test-rest coefficients was 0.688.3.correlation analysis: If MMPI-FJ was adopted, F scale was positively and remarkably correlated with Pa,Pt,Sc,Pd,D,Ma scales for male and female,Hs scale was positively and remarkably correlated with D,Pt scales for male and female,Hy scale was positively and a little highly correlated with Hs scale for male and female.If MMPI was adopted,F scale was positively and remarkably correlated with Pt,Sc,Pd, Ma scales,being weak and positive correlation with Pa scale for male;F scale was positively and remarkably correlated with Pt,Sc scales, being weak and positive correlation with Pd,Pa,Ma scales for female;Hs scale was positively and remarkably correlated with D,Hy scales.4.factor analysis: six factors were obtained by the way of factor analysis to original raw scores of basic scales in MMPI-FJ,accounting for 91.678% of the total variance.The first factor showed high loadings on F,Pa,Sc,Ma scales;The second factor showed high loadings on D,Pt,Si scales;The third factor showed high loadings on Hs,Hy scales;The fourth factor showed high loadings on L,K scales;The fifth factor showed a high loading on Mf scale;The sixth factor showed a high loading on Pd scale.5.Fisher discriminatory analysis: In two categories discriminant analysis, if MMPI-FJ was adopted,Total discriminant accuracy rates ranged from 98.1% to 100% through retrospective validation and 96.7% to 100% through cross validation between seven psychiatric groups and normal group; if MMPI was adopted, Total discriminant accuracy rates ranged from 94.3% to 100% through retrospective validation and 92.2% to 100% through cross validation between seven psychiatric groups and normal group.In multiple categories discriminant, if MMPI-FJ was adopted, averge discriminant accuracy was 83.8% through retrospective validation and 78.9% through cross validation; if MMPI was adopted, averge discriminant accuracy was 59.4% through retrospective validation and 47.3% through cross validation.6.Low boundary of clinical elevations on the MMPI-FJ were defined as T scores of 65 through receiver operator characteristic curve analysis and 92 tercile method.7.profile comparison:①MMPI-FJ's average T score's gaps between psychiatric groups and norm group were high totally than MMPI.②Highest points of MMPI-FJ profiles of psychiatric groups all were on the corresponding scales and were higher than MMPI in general.③In the MMPI-FJ profile,the group T score of whom was highest in every clinical scale(except Mf,Si scale) was corresponding psychiatric group.In the MMPI profile,the group T score of whom was highest in Hs,D,Pa,Ma scale respectively was corresponding psychiatric group,but the group T score of whom was highest in Hy,Pd,Pt,Sc scale respectively was Hs,Pz,BD,Ps group.④Profiles of ten psychiatric groups didn't overlap on other scales ( except Mf,F scale) and could been used for clinical diagnoses.Conclusion:1.MMPI-FJ was fitter for chineses than MMPI in the study, but needing to be further inspected in clinical work.2.The norms in Fujian province for subjects ranging age from 11 through 78 years have been developed.low boundary of clinical elevations of MMPI-FJ were defined.3.Eastern and Western cultures result in enormous difference of psychonosema's syndrome between China and America.4.MMPI-FJ has generated good reliability and validity.