Psychometric Considerations of the MMPI-2Slide 2Original MMPIPrior to MMPI: Logical KeyingFace ValiditySlide 6Slide 7Empirical KeyingControl GroupSlide 10MMPI vs MMPI 2MMPI-2 normative groupMMPI-2 compositionMMPI-2 RestandardizationRequirementsT Score transformationsStandard Scores or Z scoresSlide 18Percent of scores falling belowSlide 20Z-scoreMMPI-2 T scoreMMPI Uniform T scoreMMPI Uniform T-scoresFrequency high points in contemporary settingsT-score cutoffIntercorrelationsHomogeneity of itemsTemporal StabilityTest-retest reliability one weekTest-retest reliabilityInternal consistencyFactor AnalysisResponse sets and stylesMMPI vs MMPI-2Items changed MMPI-2Basic Qualifications for usersSlide 38Lexile Reading Levels in SIRSReading LevelLexile examplesTesting conditionsSlide 43The EndPsychometric Considerations of the MMPI-2William P. Wattles, Ph.D.Francis Marion University•MMPI/MMPI-2 second most widely used test by clincical psychologists (86%)•Translated into more than 50 languagesOriginal MMPI•Published 1943•Paper and pencil improvement on clinical interview and individual psychological testingPrior to MMPI: Logical Keying•Test items generated rationally based on:–Face validity–Subjective judgment•Logically keyed items problematic:–Subject to faking–Not always correctFace Validity•Does the test appear to measure what it is purported to Measure?Face Validity 0 I do not feel sad. 1 I feel sad. 2 I am sad all the time and can't snap out of it. 3 I am so sad or unhappy that I can't stand it.•70. I am easily downed in an argument•89 My hardest values are with myself•267 I have periods in which I feel unusually cheerful without any special reason.•219 I have been disappointed in loveEmpirical Keying•Original items came from many sources.•Pool of 1,000 items reduced to 566•Rewritten to be less formal and allow for some reversal of responses.Control Group•724 Visitors to the hospital in Minnesota.•Representative of Minnesota in the 1930s–16-65–Average age mind 30’s–Rural–8th grade education–WhiteEmpirical Keying•Using groups of diagnosed patients•Contrast and Cross-validationMMPI vs MMPI 2•Improved norms•Score has meaning only when compared to a representative sample.•Original sample Caucasian, 35, married, small town, good job.•New sample large and more representative.•Higher education level than populationMMPI-2 normative group•2600 Participants•Paid $15 ($40 for couples)•Tested in 7 locations–Minnesota, Ohio, North Carolina, Washington, Pennsylvania, Virginia, California•Selected from phone directoriesMMPI-2 composition•2600 Participants (started with 2900) –1138 men–1462 women•Age 18-85 (M=41, SD=15)•61% married•Education 3-20 years (M=14, SD=2)MMPI-2 Restandardization•Caucasian 81%•African-American 12%•Hispanic 3%•Native-American 3%•Asian-American 1%Requirements•Eight Grade reading level required•Satisfactory cooperation and commitment to task•Internal checks for the aboveT Score transformations•Transforming a score makes it easier to interpret.•13 validity and clinical scales converted to T scores•T score is a standard score with a mean of 50 and a standard deviation of 10.•Thus, a 70 is like a Z score of 2Standard Scores or Z scores•Z score: how many standard deviations a score lies above or below the mean.4668%99.7%95%Percent of scores falling below4699.85%84%97.5%33Z-scorezx33MMPI-2 T score)(1050xtMMPI Uniform T score•Original linear T-scores were problematic because the underlying data is somewhat skewed. •Thus a T score on one scale represented a different percentile than one on another scaleMMPI Uniform T-scores•Involve averaging of the T-scores across the scales.Frequency high points in contemporary settingsScalelinearT uniform T linearT uniform T1 5.2 4.7 6.3 4.22 21.6 13.8 15.7 15.73 3.4 3.9 3.7 6.34 10.3 12.9 11.5 13.66 18.5 18.5 23.6 28.37 4.7 8.6 2.6 3.18 29.3 29.3 29.8 19.99 6.9 8.2 6.8 8.9MalesFemalesT-score cutof•Formerly T-scores of 70 were considered clinically significant. Now the MMPI-2 recommends 65. •That puts the score above 93% of those who answer–65-50/10 = 1.5–1.5 = .9332 area under standard normal curve to the left of Z = 1.5Intercorrelations•There is considerable overlap between some scales. 13 of 39 items in scale 6 also appear in scale 8Homogeneity of items•The empirical keying approach did not favor item homogeneity thus internal consistency is not high.Temporal Stability•Ability, interest and aptitude tests should be high in temporal stability•Personality and psychopathology measures less clear.Test-retest reliability one weekTest-retest reliability•Summary. Test is fairly stable and changes when current appear consistent. Significant changes generally correctly reflect behavior change.Internal consistency•Moderate, not a strength for the MMPI-2 due to empirical keying approach.Factor Analysis•Two strong factors identified–General maladjustment and psychotic thought–Neurotic characteristicsResponse sets and styles•Charges that MMPI and MMPI-2 were confounded by response style.•Block modified MMPI to have equal number of true and false items•Test seems to be valid in a variety of settings.MMPI vs MMPI-2•Validity similar•Raw Scores higher on MMPI-2–May be explained by instructions•T-scores compensate for higher raw scoresItems changed MMPI-2Basic Qualifications for users•Graduate-level course in psychological testing.–Standard scores–Transformations–Understand limits of accuracy–Standard error of measurementBasic Qualifications for users•Graduate-level course in psychopathology–Personality structure–Dynamics–Deviance psychodiagnostic concepts–Diagnostic systems –Broad understanding of human personalityLexile Reading Levels in SIRSWhat Are Lexile Reading Levels•Lexile scores match reader ability and text difficulty, allowing individualized monitoring of student progress.•Due to the accountability requirements of NCLB, many states are turning to standardized systems for reading which help to track student progress.How does it Work?•Lexile measures are based on two well-established predictors of how difficult a text is to comprehend: semantic difficulty (word frequency) and syntactic complexity (sentence length).Age School YearTypical Lexile Level789101112131415163456789101112300- 800400-