Validity of MeasuresRefers to whether or not the measurement is accurateTypes:“face”: does it look like it measures what its supposed tohas nothing to do with whether a measure is accuratenot important in determining accuracyContent ValidityMeasure provides a representative sampling of what is being measuredSampling across the full range of what you’re trying to measureDoes it measure what its supposed to measureCriterion –Related ValidityMeasure score is related to a criterion that represents the measurePredictive: related to the futureTaking the SAT, predicts if you’re going to be a good college studentTaking a test to see if you’ll be a good copNFL and predictive screenings of playersConcurrent (sensitivity/specificity): related to the presentResults given the present, is it related to a criterion that exists right nowAntecedent: related to the pastAssessing the criteria of a criminal at the time of the crimeAssessing cognitive function before an accidentConstruct ValidityAre we confident the measure a good indicator of the construct?Intelligence is an example of constructConvergent : is measure related to other things that are trying to be measured?Bad Convergent: you don’t get relationships you should be gettingDiscriminant: are we certain the measure isn’t measuring something else?It isn’t measuring something you didn’t want it to measureBad Discriminant: Get relationships you shouldn’t be gettingAssessment StrategiesSelf reportWhat someone says about themselvesProsIts easy to measureAdministration is simple (can be given in person or over the Web)Scoring is easy, b/c there is a finite number of options to choose from & reliableWho would know more about yourself than you?ConsJust b/c someone says they are a certain way doesn’t mean they areThere are things that you aren’t aware about yourselfb/c the number of responses are finite, it restricts responses b/c people can’t elaborateProjective MethodsProvide someone with an ambiguous situation and allow them to respond in any wayRorshach Inkblot Test: most popularIn the presence of ambiguity, people project their personality onto the ambiguityEx: viewing a indie art film with friends; and you all have different interpretations of what was going on, the differences represent the different personalitiesProsNumber of responses are infiniteNot easy to distort responsesConsThe reliability is an issue b/c responses are subjectiveEven if its reliable, it may not be validBehavioral MethodsSample behaviors relative to the construct; watching what people do and not what they sayProsLess vulnerable for people to distort their responsesCons:Reactivity occurs, behavior might not be representative of actual personality by being followed aroundThe same behavior may represent 2 different thingsThe scorer-reliability isn’t perfect, surveys more reliable that behavioral assessmentsSociometric MethodsPeer-reports; asking other people about youCons:If you’re motivated to distort self perception, what will stop a peer from distorting reportPsychophysiological MethodsLooking at things in the body to determine personalityEx: polygraph (lie-detector test)ConsDiscriminant validity; these measure arousalQuestionable validityNot admissible in courtLecture 4Outline of Last Lecture I. Sampling Strategies II. Interaction Effect III. Reliability and Correlationa. Methods of estimating reliabilityi. Test-retestii. Internal consistency/split-half Outline of Current Lecture IV. Validitya. Faceb. Content c. Criterion-relatedi. Predictiveii. Concurrentiii. Antecedent d. Constructi. Convergent ii. Discriminant V. Assessment Strategies a. Self- reporti. Prosii. Consb. Projective Methodsi. Pros ii. Consc. Behavioral Methodsi. Prosii. Consd. Sociometric Methodsi. Prosii. Conse. Psychophysiological Methodsi. Prosii. Cons PSYC 330 1st EditionValidity of Measures - Refers to whether or not the measurement is accurateo Types: “face”: does it look like it measures what its supposed to has nothing to do with whether a measure is accurate not important in determining accuracy Content Validity Measure provides a representative sampling of what is being measured- Sampling across the full range of what you’re trying to measure - Does it measure what its supposed to measure Criterion –Related Validity Measure score is related to a criterion that represents the measure - Predictive: related to the futureo Taking the SAT, predicts if you’re going to be a good college student o Taking a test to see if you’ll be a good copo NFL and predictive screenings of players - Concurrent (sensitivity/specificity): related to the present o Results given the present, is it related to a criterion that exists right now- Antecedent: related to the past o Assessing the criteria of a criminal at the time of the crimeo Assessing cognitive function before an accident Construct Validity Are we confident the measure a good indicator of the construct?- Intelligence is an example of construct o Convergent : is measure related to other things that are trying to be measured? Bad Convergent: you don’t get relationships you shouldbe getting o Discriminant: are we certain the measure isn’t measuring something else? It isn’t measuring something you didn’t want it to measure Bad Discriminant: Get relationships you shouldn’t be getting Assessment Strategies - Self reporto What someone says about themselves Pros Its easy to measure Administration is simple (can be given in person or over the Web) Scoring is easy, b/c there is a finite number of options to choose from & reliable Who would know more about yourself than you? Cons Just b/c someone says they are a certain way doesn’t mean they are There are things that you aren’t aware about yourself b/c the number of responses are finite, it restricts responses b/c people can’t elaborate - Projective Methodso Provide someone with an ambiguous situation and allow them to respond in any way Rorshach Inkblot Test: most popular In the presence of ambiguity, people project their personality onto the ambiguity Ex: viewing a indie art film with friends; and you all have different interpretations of what was going on, the differences represent the different personalities Pros Number of responses are infinite Not easy to distort responses Cons The reliability is an issue b/c responses are
View Full Document