Summary / Abstract
Title: Validity of the Beck Depression Inventory, MMPI, and Rorschach in assessing adolescent depression
Synopsis: The purpose of the present study was to assess the validity of the Beck Depression Inventory (BDI), theMMPI Depression scale (MMPI-D), and the Rorschach Depression Index (DEPI) in measuring adolescent depression. Retroactivecharts of 118 hospitalized adolescents were divided into depressed (
=66) and nondepressed (
=52)groups, based on the psychiatric diagnosis. Results indicated significant differences between the means of the two groups on theBDI (
<0.00001) and the MMPI–D scale (
<0.0001), establishing concurrent validity for these measuresof depression. Correlation coefficients among the three measures of depression indicated a significant relationship between the BDIand the MMPI–D scale (
<0.01), again supporting the concurrent validity of these instruments. Predictive utilitywas determined for each of the three instruments individually and in combination, with results indicating that the BDI andMMPI–D scales alone and in combination significantly discriminated depressed from nondepressed adolescents. The MMPI–Dscale alone was the most accurate measure in classifying the participants into their respective groups. The DEPI was notestablished as a valid predictor of adolescent depression.
Title: Validity of the Beck Depression Inventory, MMPI, and Rorschach in assessing adolescent depression
Synopsis: The purpose of the present study was to assess the validity of the Beck Depression Inventory (BDI), theMMPI Depression scale (MMPI-D), and the Rorschach Depression Index (DEPI) in measuring adolescent depression. Retroactivecharts of 118 hospitalized adolescents were divided into depressed ( n =66) and nondepressed ( n =52)groups, based on the psychiatric diagnosis. Results indicated significant differences between the means of the two groups on theBDI ( p <0.00001) and the MMPI–D scale ( p <0.0001), establishing concurrent validity for these measuresof depression. Correlation coefficients among the three measures of depression indicated a significant relationship between the BDIand the MMPI–D scale ( p <0.01), again supporting the concurrent validity of these instruments. Predictive utilitywas determined for each of the three instruments individually and in combination, with results indicating that the BDI andMMPI–D scales alone and in combination significantly discriminated depressed from nondepressed adolescents. The MMPI–Dscale alone was the most accurate measure in classifying the participants into their respective groups. The DEPI was notestablished as a valid predictor of adolescent depression.
Related links for: Validity of the Beck Depression Inventory, MMPI, and Rorschach in assessing adolescent depression
Additional keywords for: Validity of the Beck Depression Inventory, MMPI, and Rorschach in assessing adolescent depression
journal adolescence 1996 19 223 231 validity beck depression inventory mmpi rorschach assessing adolescent depression cynthia l carter christine m dacey purpose present study assess validity beck depression inventory bdi mmpi depression scale mmpi d rorschach depression index depi measuring adolescent depression retroactive charts 118 hospitalized adolescents divided depressed n 66 nondepressed n 52 groups based psychiatric diagnosis results indicated significant differences means two groups bdi p 0 00001 mmpi d scale p 0 0001 establishing concurrent validity measures depression correlation coefficients three measures depression indicated significant relationship bdi mmpi d scale p 0 01 supporting concurrent validity instruments predictive utility determined three instruments individually combination results indicating bdi mmpi d scales combination significantly discriminated depressed nondepressed adolescents mmpi d scale accurate measure classifying participants respective groups depi established valid predictor adolescent depression 1996 association professionals services adolescents introduction depression adolescents received increasing attention recent years due research suggesting adolescent depression predicts future difficulties school delinquency substance abuse criminal behavior marital problems unemployment status chiles et al 1980 kandel davies 1986 newcomb bentler 1988 recent study harrington et al 1990 found episode depression early life substantially increases episodes life prevalence rates depression adolescents range roughly 2 3 9 bird et al 1988 mcgee et al 1990 lewinsohn et al 1993 increasing concern regarding ability diagnostic measures adequately assess depression adolescent populations main focus present study assess concurrent validity predictive utility three instruments 1 beck depression inventory bdi beck et al 1961 2 minnesota multiphasic personality inventory mmpi hathaway mckinley 1942 3 rorschach test rorschach 1921 widely instruments represent self report empiricallybased projective techniques respectively measuring depression bdi mmpi depression scale mmpi d rorschach depression index depi exner 1986 extensively researched years adults validity reprint requests correspondence addressed c m dacey phd department psychology xavier university cincinnati oh 45207 6411 u 0140 1971 96 03022309 18 00 0 1996 association professionals services adolescents 224 c l carter c m dacey scales measuring depression adolescent populations yielded conflicting data archer et al 1985 found children aged 12 18 years diagnosed dysthymic disorder produced higher mmpi d scale values children four diagnostic groups borderline personality disorder conduct disorder personality disorders disorders childhood adolescence comprehensive study archer gordon 1988 mmpi rorschach investigated assess usefulness detection depression schizophrenia adolescent inpatients aged 12 17 mmpi d scores rorschach depi scores found significantly related patients diagnoses contrast lipovsky et al 1989 found scores mmpi d differ significantly depressed n 35 nondepressed n 25 adolescents supporting validity depression d scale strober et al 1981 reported bdi psychometrically sound instrument measuring depressive symptomatology psychiatrically disturbed adolescents reported significant correlation bdi scores global clinical rating depression r 76 0 67 p 0 001 results reported kauth zettle 1990 ambrosini et al 1991 marton et al 1991 barrera garrison jones 1988 reported accuracy bdi identifying depressed nondepressed inpatients reported strober et al 1981 cutoff score 11 resulted sensitivity 81 27 7 false positives finch et al 1990 found 500 children adolescents admitted psychiatric inpatient unit m age 13 17 d 3 14 depi positive two patients suggesting incidence positive depression index rare occurrence population present study concurrent validity bdi mmpi d scale depi measuring adolescent depression assessed comparisons depressed nondepressed groups respect scores three instruments second form concurrent validity determined studying correlations three instruments additionally study evaluated predictive utility three instruments individually combination discriminating adolescent depression method participants participants consisted 118 adolescents hospitalized adolescent unit south western ohio 58 boys 60 girls ranging age 12 17 years m 15 03 d 1 36 divided depressed nondepressed groups based psychiatric diagnoses dsm iii r criteria psychiatric diagnosis assigned 72 h admission based clinical interview direct observation attending psychiatrist input multidisciplinary treatment team participants selected retroactive charts maintained psychology department hospital files covering 2 year period reviewed chronologically examiner participants limited adolescents tested trained clinical psychology graduate students supervised two psychologists ensure standardization testing procedures initial selection participants 225 assessment adolescent depression based supervising psychologists psychiatric diagnosis demographic information collected participant official hospital medical records participants divided depressed n 76 nondepressed n 59 groups based psychiatric diagnosis outlined previously data sheet developed record psychiatric diagnosis test data demographic information participant demographic information collected participant gender age grade level previous hospitilizations previous suicide attempts previous reports suicidal ideation depressed group consisted 40 girls 26 boys diagnosed depressive disorder specified 41 major depression 39 dysthymia disorder 11 adjustment disorder depressed mood 9 depressed group mean age 15 12 years d 1 39 participants ranged grade level 6 12 mean grade level 9 5 fifteen cent previously hospitalized 39 reported suicide attempt 66 history suicidal ideation nondepressed sample consisted 20 girls 32 boys diagnosed conduct disorder 73 attention deficit disorder 9 adjustment disorder disturbance conduct 8 oppositional defiant disorder 6 impulse control disorder 4 adolescents mean age 14 94 years d 1 33 ranged grade level 7 12 mean grade level 9 2 twelve cent previously hospitalized 4 history reported suicide attempt 15 history suicidal ideation measures beck depression inventory bdi bdi consists 21 items cover range affective behavioral cognitive somatic symptoms thought indicative unipolar depression beck et al 1961 item subject select four alternative responses reflect increasing levels severity depressive symptomatology possible scores range 0 63 higher scores represent reported depression mmpi depression scale mmpi d primary means assessing depressive symptomatology mmpi depression d scale consists 60 items subject agrees disagrees allowing range scores 0 60 items reflect clinical symptoms characterize feelings hopelessness despair discouragement basic personality features hyper responsibility high personal standards intrapunitiveness hathaway mckinley 1943 d scale hathaway mckinley 1942 able correctly identify 69 sample depressed individuals criterion t score greater 70 adolescent norms marks et al 1974 present study convert raw scores t scores rorschach depression index depi depi computed structural summary comprehensive system exner 1986 depi composed five variables vista responses fvvfv 0 color shading blends color shading blends 0 egocentricity index 3r 2 r 0 30 lower 1 d appropriate age mean children achromatic color responses fc c fc 2 morbid responses mor 3 exner 1986 reported depi represents sum variables positive affective disturbance responder receive score ranging 0 5 depi 226 c l carter c m dacey procedure psychological test data obtained participant department psychology files entering hospital adolescent administered battery tests comprehensive psychological evaluation evaluation conducted independently psychiatric clinical interview played role assignment psychiatric diagnosis battery administered standardized procedures evaluative session breaks necessary keep participant tiring battery involved intelligence testing personality testing bdi mmpi rorschach test measures rorschach administered scored exner comprehensive system exner 1986 checked accuracy scoring supervising psychologist original 135 participant files met criteria inclusion study 17 deleted based criteria 1 incomplete test 2 invalid mmpi profile fk16 3 invalid rorschach test responses 10 responses 10 12 lambda greater 1 5 children older 7 final sample resulted depressed group consisting 66 participants nondepressed group consisting 52 participants data collected participant concurrent validity instruments established comparisons depressed nondepressed groups respect means f tests standard deviations fmax tests instruments second form concurrent validity assessed computation pearson correlation coefficients study relationships three instruments predictive utility test battery established discriminant function analyses results means standard deviations three instruments depressed nondepressed groups presented table 1 univariate f tests conducted means groups bdi mmpi d scale depi results indicated significant differences groups bdi f 1 116 20 33 p 0 00001 mmpi d scale f 1 116 17 23 p 0 0001 statistically significant differences found means groups depi order explore validity depi analyses conducted examine possible differences means groups individual variables depi statistically significant differences found five variables vista color shading blends egocentricity achromatic color morbid content see table 1 compare groups respect variability fmax tests conducted instrument fmax scores bdi fmax 1 1 p 0 05 mmpi d scale fmax 1 2 p 0 05 depi fmax 1 2 p 0 05 significant indicating homogeneity variance depressed nondepressed groups second form concurrent validity assessed computing pearson correlation coefficients test significant relationships scores achieved bdi mmpi d scale depi results analysis revealed statistically significant relationship bdi mmpi d scale r 0 54 p 0 01 significant correlation depi bdi depi mmpi d scale see table 2 pearson correlation coefficients computed depressed nondepressed 227 assessment adolescent depression groups separately results indicated depressed group bdi mmpi d scale statistically significantly correlated r 0 56 p 0 01 statistically significant relationship bdi depi mmpi d scale depi nondepressed sample statistically significant correlation established bdi mmpi d scale r 0 30 p 0 05 bdi depi mmpi d scale depi see table 2 order study difference strengths correlations bdi mmpi d scale two groups z test independent samples computed significant difference groups strength correlations z 1 703 p 0 05 seven discriminant function analyses conducted assess predictive utility table 1 means standard deviations results f tests bdi mmpi d scale depi individual variables depi depressed nondepressed groups depressed m d nondepressed m d f value p bdi 19 95 11 32 20 33 0 00001 10 63 10 04 mmpi d 70 50 60 17 17 23 0 0001 12 08 13 08 depi 1 18 1 07 0 24 0 6238 1 18 1 09 v 0 33 0 32 0 20 0 9641 0 66 0 87 c 0 37 0 40 0 33 0 8560 0 71 0 77 ego 0 36 0 39 0 96 0 3303 0 16 0 16 achr 1 34 1 21 0 20 0 6592 1 69 1 63 mor 1 10 1 01 0 66 0 6938 1 22 1 22 bdi beck depression inventory mmpi d minnesota multiphasic personality inventory depression scale depi rorschach depression index v vista responses c color shading blends ego egocentricity index achr achromatic color mor morbid responses p 0 001 table 2 pearson correlation coefficients bdi mmpi d scale depi total sample depressed group nondepressed group total group n 118 depressed n 66 nondepressed n 52 bdi mmpi d depi bdi mmpi d depi bdi mmpi d depi bdi 1 0 1 0 1 0 mmpi d 0 54 1 0 0 56 1 0 0 30 1 0 depi 0 08 0 11 1 0 0 07 0 09 1 0 0 07 0 15 1 0 bdi beck depression inventory mmpi d scale minnesota multiphasic personality inventory depression scale depi rorschach depression index p 0 05 p 0 01 228 c l carter c m dacey instruments individually combination approach assess value added instrument test battery predicting depression hospitalized adolescents bdi predictor variable discriminant function statistically significant discriminating depressed nondepressed samples c 2 1 n 118 18 65 p 0 00001 function produced classification accuracy rate 62 1 depressed group 73 1 nondepressed group 67 total sample mmpi d scale predictor variable significant function derived c 2 1 n 118 15 99 p 0 0001 produced classification accuracy rate 71 2 depressed group 75 nondepressed group 72 9 total sample depi sole predictor variable yield significant discriminant function considering three instruments predictor variables varying combinations combination increased classification accuracy rates produced mmpi d scale see tables 3 4 table 3 results discriminant function analyses bdi mmpi d scale depi predictor variables predictor variable wilks lambda chi square df significance bdi 0 85 18 65 1 0 00001 mmpi d 0 87 16 00 1 0 0001 depi 0 997 0 24 1 0 6238 bdi mmpi d 0 82 23 30 2 0 00001 bdi depi 0 85 18 60 2 0 0001 mmpi d depi 0 87 15 92 2 0 0003 bdi mmpi d depi 0 82 23 21 3 0 00001 bdi beck depression inventory mmpi d minnesota multiphasic personality inventory depression scale depi rorschach depression index p 0 001 table 4 classification accuracy rates discriminant function analyses bdi mmpi d scale depi predictor variables predictor variable correctly classified depressed group nondepressed group total sample n 66 n 52 n 118 bdi 62 1 73 1 67 0 mmpi d 71 2 75 0 72 9 depi 33 3 69 2 49 2 bdi mmpi d 68 2 75 0 71 2 bdi depi 60 6 73 1 66 1 mmpi d depi 71 2 75 0 72 9 bdi mmpi d depi 68 2 75 0 71 2 bdi beck depression inventory mmpi d minnesota multiphasic personality inventory depression scale depi rorschach depression index 229 assessment adolescent depression discussion purpose study investigate concurrent validity predictive utility beck depression inventory mmpi d scale rorschach depression index measuring depression hospitalized adolescents determine usefulness test instruments discriminating depression make comparisons depressed nondepressed groups respect means standard deviations instruments results analyses indicated significant difference means groups bdi mmpi d scale homogeneity variance established two groups instrument concurrent validity measures depression supported second assess concurrent validity determine correlations three instruments measuring depression statistically significant pearson correlation coefficients indicated bdi mmpi d scale entire sample depressed nondepressed groups separately findings consistent previous research indicated bdi mmpi d scale psychometrically sound instruments measuring adolescent depression strober et al 1981 lipovsky et al 1989 marton et al 1991 mean scores depi individual variables depi differ significantly depressed nondepressed groups depi significantly correlated two measures depression results consistent archer gordon 1988 findings lipovsky et al 1989 findings question validity index discriminating depression adolescents looking mean scores depi subscales low incidence positive indicators depression suggesting low sensitivity instrument assessing adolescent depression findings stand sharp contrast exner 1986 reports positive hit rates depression ranging 70 90 depi individual variables possible adolescents hospitalized adolescents typically distinctive trait emotional constriction contributed inability depi discriminating narrow range possible scores 0 5 depi restrict sensitivity index detecting adolescent depression recently exner 1993 revised depi adding two variables index improve accuracy time study revised version published turning predictive utility three measures results discriminant function analyses suggested combination bdi mmpi d scales statistically effective discriminating depressed vs nondepressed groups hospitalized adolescents bdi correctly classified participants depressed nondepressed accuracy rate 67 mmpi d scale correctly classified participants accuracy rate 72 9 consequently bdi able statistically discriminate depressed nondepressed groups demonstrated value predicting adolescent depression accurate mmpi d scale sole predictor fact combination tests increased accuracy mmpi d scale correctly classifying depressed nondepressed adolescents despite statistical significance established study clinical significance findings needs addressed classification accuracy rates 67 73 allow 230 c l carter c m dacey clinician make correct diagnosis degree certainty like archer et al 1985 results support mmpi d scale bdi initial screening measure background information relevant data considered formulation final diagnosis future research attempt delineate relevant data sources e g number suicide attempts determine value added source accuracy classification depressed adolescents noted 1992 university minnesota press released minnesota multiphasic personality inventory adolescent mmpi butcher et al 1992 specifically address issues involving adolescent development forms psychopathology fifty seven original 60 items mmpi d scale retained mmpi depression scale preliminary data available mmpi depressed adolescents consequently research needed evaluate efficacy instrument assessing adolescent depression summary current study supports validity bdi mmpi d scale measuring predicting adolescent depression depi established valid measure depression population despite indicated utility bdi mmpi d scales combination mmpi d scale shown greatest discriminative ability depression sample hospitalized adolescents bdi requires time easier administer score mmpi added discriminative ability approximately 6 mmpi d scale justify added effort necessary instrument clinical settings consequently screening depression sample hospitalized adolescents clinicians need consider potential usefulness mmpi d scale bdi select clinically appropriate instrument references ambrosini p j metz c bianchi m d rabinovich h 1991 concurrent validity psychometric properties beck depression inventory outpatient adolescents journal american academy child adolescent psychiatry 30 51 57 archer r p ball j d hunter j 1985 mmpi characteristics borderline psychopathology adolescent inpatients journal personality assessment 49 47 55 archer r p gordon r 1988 mmpi rorschach indices schizophrenic depressive diagnoses adolescent inpatients journal personality assessment 52 276 287 barrera m jr garrison jones c v 1988 properties beck depression inventory screening instrument adolescent depression journal abnormal child psychology 16 263 273 beck ward c mendelson m mock j erbaugh j 1961 inventory measuring depression archives general psychiatry 4 53 63 bird h r canino g rubio stipec m gould m ribera j sesman m woodbury m huertas goldman pagan sanchez lacay moscoso m 1988 estimates prevalence childhood maladjustment community survey puerto rico combined measures archives general psychiatry 45 1120 1126 butcher j n williams c l graham j r archer r p tellegen ben porath y kaemmer b 1992 mmpi minnesota multiphasic personality inventory adolescent manual administration scoring interpretation minneapolis university minnesota press chiles c miller m l cox g b 1980 epidemiology depressive symptomatology adolescence journal american academy child psychiatry 23 91 98 exner j e 1986 rorschach comprehensive system vol 1 2nd edn new york wiley pp 59 183 424 425 231 assessment adolescent depression exner j e 1993 rorschach comprehensive system vol 1 3rd edn new york wiley p 189 finch j jr imm p belter r w 1990 brief rorschach records children adolescents journal personality assessment 55 640 646 harrington r fudge h rutter m pickles hill j 1990 adult outcomes childhood adolescent depression archives general psychiatry 47 465 473 hathaway r mckinley j c 1942 multiphasic personality schedule minnesota measurement symptomatic depression journal psychology 14 73 83 hathaway r mckinley j c 1943 manual minnesota multiphasic personality inventory new york psychological corporation kandel d b davies m 1986 adult sequelae adolescent depressive symptoms archives general psychiatry 43 255 262 kauth m r zettle r d 1990 validation depression measures adolescent populations journal clinical psychology 46 291 295 lewinsohn p m hops h roberts r e seeley j r andrews j 1993 adolescent pathology i prevalence incidence depression dsm iii r disorders high school students journal abnormal psychology 102 133 144 lipovsky j finch j jr belter r w 1989 assessment depression adolescents objective projective measures journal personality 53 449 458 marks p seeman e haller d l 1974 actuarial mmpi adolescents adults baltimore md williams wilkins marton p churchard m kutcher korenblum m 1991 diagnostic utility beck depression inventory adolescent psychiatric outpatients inpatients canadian journal psychiatry 36 428 431 mcgee r feehan m williams partridge f silva p kelly j 1990 dsm iii disorders large sample adolescents journal american academy child adolescent psychiatry 29 611 619 newcomb m d bentler p m 1988 impact adolescent drug social support problems young adults longitudinal study journal abnormal psychology 97 64 75 rorschach h 1921 psychodiagnostik bern bircher transl hans huber verlag 1942 strober m green j carlson g 1981 utility beck depression inventory psychiatrically hospitalized adolescents journal consulting clinical psychology 49 482 483, com_apnet_jado_jado_1996_0021, KnowledgeStor, Knowledge-Stor,