Personality Assessment, Measurement, and Research Design CHAPTER 2 © 2015 M. Guthrie Yarwood 4 Sources of personality data II. Reliability / Validity in Personality III. A little more on observer reports I. Outline © 2015 M. Guthrie Yarwood I. Sources of Personality Data Self-Report Data (S-Data) Observer-Report Data (O-Data) Test-Data (T-Data) Life-Outcome Data (L-Data) © 2015 M. Guthrie Yarwood Self-Report Data (S-Data) Information provided by a person, such as through a survey or interview Limitations of S-data? © 2015 M. Guthrie Yarwood O-Data Information provided by someone else about another person Professional personality assessors People who actually know the target person Naturalistic vs. Artificial Observation Limitations? © 2015 M. Guthrie Yarwood Test-Data (T-Data) Information provided by standardized tests or testing situations Situation designed to elicit behaviors that serve as indicators of personality © 2015 M. Guthrie Yarwood Test-Data Creativity Example What are unusual uses for common objects – bricks, knives, newspapers? Answers to hypothetical events What would happen if people went blind? What would happen is people shrank to 12 inches tall? (Paul Silvia’s work) © 2015 M. Guthrie Yarwood Test-Data: Other Examples Mechanical recording devices, e.g., “Actometer” used to assess children’s activity Physiological data Projective Tests Ex: Fairy Tale Test Limitations? © 2015 M. Guthrie Yarwood Life-Outcome Data (L-Data) Information that can be gleaned from events, activities, and outcomes in a person’s life that is available for public scrutiny—e.g., marriage, speeding tickets Can serve as important source of “real life” information about personality Ex: implicit egotism: people gravitate toward people, places, things that resemble the self © 2015 M. Guthrie Yarwood (Pelham and colleagues’ work) Issues in Personality Assessment Links among different data sources Fallibility of personality measurement All sources of data have limitations Results that replicate through “triangulation” are most powerful © 2015 M. Guthrie Yarwood You are a personality psychologist and would like to measure the personality trait risk-taking (i.e., sensation seeking). How could you measure risk-taking using each of the four data sources? S-Data O-Data T-Data L-Data © 2015 M. Guthrie Yarwood Evaluation of Personality Measures How do we know whether our scale is a “good scale?” Types of Errors Reliability Validity Threats to Reliability and Validity © 2015 M. Guthrie Yarwood Extraneous vs. Confounding Variables Extraneous variables are variables that may compete with the independent variable in explaining the outcome of a study. Confounding variable: an extraneous variable that does indeed influence the dependent variable. A confounding variable systematically varies or influences the independent variable and also influences the dependent variable. © 2015 M. Guthrie Yarwood Random vs. Constant Errors Random Errors (unsystematic): extraneous variables whose average influence on the outcome is the same in both (or all) conditions (Aronson et al., 1990) Affects reliability AND validity Constant (systematic) Errors: influences all the scores in one condition in the same direction and has no effect or a different effect on the other condition. Affects ONLY validity Confounding variables © 2015 M. Guthrie Yarwood Clarification Check! IV: Type of show Condition 1: Extraverts watch comedy show Condition 2: Extraverts watch neutral show DV: Self-reported emotions Extraverts report more positive emotions in Comedy Condition than in Neutral Condition © 2015 M. Guthrie Yarwood Clarification Check! – Random Errors Extraverts in conditions 1 and 2 are placed in very hot rooms. The hot temperature lowers their self-reported levels of positive emotions. But, we still find that extraverts report more positive emotions when watching the comedy show than the neutral show. Accurate: We find differences in self-reported positive emotions (on 5-point scale): Comedy Cond. = 4.8, Neutral Cond. = 2.5 Random Errors: In presence of hot temperature, we find differences in self-reported positive emotions Comedy Cond. = 3.0, Neutral Cond. = 1.5 © 2015 M. Guthrie Yarwood Clarification Check! – Random Errors Extraverts in conditions 1 and 2 are placed in very hot rooms. The hot temperature lowers their self-reported Hot temperature lowers the levels of positive emotions. scores in both conditions!! But, we still find that extraverts report more positive emotions when watching the comedy show than the neutral show. Accurate: We find differences in self-reported positive emotions (on 5-point scale): Comedy Cond. 1 = 4.8, Neutral Cond. 2 = 2.5 Random Errors: In presence of hot temperature, we find differences in self-reported positive emotions Comedy Cond = 3.0, Neutral Cond.2 = 1.5 © 2015 M. Guthrie Yarwood Clarification Check! – Constant Errors Extraverts in Comedy Cond. are in a 70◦ room. Extraverts in Neutral Cond. are placed in a 78◦ room. In Neutral Cond. only, the hot temperature lowers selfreported levels of positive emotions. We find that the comedy show results in more positive emotions than the neutral show. Accurate: We do not find differences in self-reported positive emotions Comedy Cond. = 4.8, Neutral Cond. = 4.6 Constant Error: We find differences in positive emotions due to the hot temperature, not due to the manipulation! Comedy Cond. = 4.8, Neutral Cond. = 2.5 © 2015 M. Guthrie Yarwood Clarification Check! – Constant Errors Extraverts in Comedy Cond. are in a 70◦ room. Extraverts in Neutral Cond. are placedHot intemperature a 78◦ room. lowersIn the scores in Condition Neutral Cond. only, the hot temperature lowers self- 2, but not Condition 1!! reported levels of positive emotions. We find that the comedy show results in more positive emotions than the neutral show. Accurate: We do not find differences in self-reported positive emotions Comedy Cond. = 4.8, Neutral Cond. = 4.6 Constant Error: We find differences in positive emotions due to the hot temperature, not due to the manipulation! Comedy Cond. = 4.8, Neutral Cond. = 2.5 © 2015 M. Guthrie Yarwood Reliability The ability of a test to measure an attribute consistently Does this extraversion scale measure the true level of each participant’s extraversion (over time)? Are coders following the same method? To achieve reliability we want to reduce random error. © 2015 M. Guthrie Yarwood Threats to Reliability (random error) Participant Characteristics: fatigue, motivation, boredom Testing situation: time of day, room temperature Testing Instrument: instructions, rating scale, items, reading level. Experimenter Characteristics and Errors: interactions with participants; incorrect observations of participants; incorrect coding of behavior © 2015 M. Guthrie Yarwood Estimating Reliability Test-Retest Coefficient Parallel-Forms Coefficient Internal Consistency Coefficient Interrater (interobserver) Reliability © 2015 M. Guthrie Yarwood Validity Validity: The degree to which a test or measurement accurately measures or reflects what it claims to measure. Internal Validity: Did the experimental treatments make a difference in this specific experimental instance? External Validity: Generalizability; To what populations, settings, treatment variables, and measurement variables can the effect be generalized? **Never completely answerable © 2015 M. Guthrie Yarwood Internal Validity Degree to which test measures what it claims to measure 5 types of internal validity Face validity Predictive or criterion validity Convergent validity Discriminant validity Construct validity © 2015 M. Guthrie Yarwood Threats to Internal Validity (errors) Affected by random and constant errors Random (unsystematic) Errors: same errors that affect reliability Constant Errors (systematic): errors that affect measurement of variable; does not affect reliability © 2015 M. Guthrie Yarwood Knowledge Check! You are conducting a study on the personality traits associated with the frequency of exercising. For your study, which of the following poses a threat to validity, but not reliability? A. All the participants are bored. B. The construction outside the laboratory window is very loud. C. You recruited participants from the Rec Hall. D. The experimenter who greets all participants is very rude. © 2015 M. Guthrie Yarwood Knowledge Check! You are conducting a study on the personality traits associated with the frequency of exercising. For your study, which of the following poses a threat to validity, but not reliability? A. All the participants are bored. B. The construction outside the laboratory window is very loud. C. You recruited participants from the Rec Hall. D. The experimenter who greets all participants is very rude. © 2015 M. Guthrie Yarwood Threats to Internal Validity – MRS SMITH Maturation Regression to the Mean Selection of Subjects Selection by Maturation Interaction Mortality Instrumentation Testing History © 2015 M. Guthrie Yarwood Regression to the Mean Example In an experiment involving reading instruction, subjects grouped because of poor pre-test reading scores show considerably greater gain than do the groups who scored average and high on the pre-test. © 2015 M. Guthrie Yarwood Regression to the Mean Example In an experiment involving reading instruction, subjects grouped because of poor pre-test reading scores show considerably greater gain than do the groups who scored average and high on the pre-test. Pre-test Poor © 2015 M. Guthrie Yarwood Average / Mean High Regression to the Mean Example In an experiment involving reading instruction, subjects grouped because of poor pre-test reading scores show considerably greater gain than do the groups who scored average and high on the pre-test. Pre-test Poor After reading Instruction © 2015 M. Guthrie Yarwood Average / Mean High Selection by Maturation Interaction Group Pre-test Head Start Intervention Middle-class children 65/100 Disadvantaged children 65/100 Control Group – No Intervention Middle-class children 65/100 Disadvantaged children 65/100 © 2015 M. Guthrie Yarwood Post-Test (after head start) Selection by Maturation Interaction Group Pretest Head Start Intervention Control Group – No Intervention © 2015 M. Guthrie Yarwood Post-Test: 6 months Post-Test: 12 months Post-Test: 18 months 65/100 70/100 75/100 80/100 65/100 65/100 65/100 65/100 Selection by Maturation Interaction Group Pretest Head Start Intervention Control Group – No Intervention Post-Test: 6 months Post-Test: 12 months Post-Test: 18 months 65/100 70/100 75/100 80/100 65/100 65/100 65/100 65/100 Ss in Intervention are middle class, while Ss in control group are disadvantaged. Over time, Intervention Ss show improvement of posttest due to better health care, greater parental support, greater access to resources, etc. © 2015 M. Guthrie Yarwood Identify the Threat to Validity! In a short experiment designed to investigate the effect of computer-based instruction, Ss missed some instruction because of a power failure at school. A. History B. Mortality C. Testing D. Instrumentation © 2015 M. Guthrie Yarwood Identify the Threat to Validity! In a short experiment designed to investigate the effect of computer-based instruction, Ss missed some instruction because of a power failure at school. A. History B. Mortality C. Testing D. Instrumentation © 2015 M. Guthrie Yarwood Identify the Threat to Validity! In a health experiment designed to determine the effect of various exercises, those Ss who find the exercise most difficult stop participating. A. Selection of Subjects B. Mortality C. Testing D. Maturation © 2015 M. Guthrie Yarwood Identify the Threat to Validity! In a health experiment designed to determine the effect of various exercises, those Ss who find the exercise most difficult stop participating. A. Selection of Subjects B. Mortality C. Testing D. Maturation © 2015 M. Guthrie Yarwood Estimating Internal Validity Content Validity Criterion-Related Validity Concurrent Validity Predictive Validity Construct Validity Convergent Validity Discriminant Validity © 2015 M. Guthrie Yarwood Content Validity Definition: whether the content of a test elicits a range of responses that are representative of the entire domain or universe of skills, understandings, and other behaviors a test is designed to measure. To assess: compare tests’ content with an outline of specifications concerning subject matter to be covered in test. © 2015 M. Guthrie Yarwood ????? ????? Openness to Experience ????? ????? © 2015 M. Guthrie Yarwood Behaviors? Thoughts/Cognitions? Openness to Experience Feelings/Emotions? Perceptions? © 2015 M. Guthrie Yarwood Criterion-Related Validity Definition: procedures in which the test scores of a group of people are compared with ratings, classifications, or other measures of performance. © 2015 M. Guthrie Yarwood Concurrent Validity – A type of criterion-related validity Concurrent Validity: when a test is administered to people in various categories, to determine whether test scores of people in 1 category are significantly different from people in other categories. Clinical vs. non-clinical group Different socioeconomic levels © 2015 M. Guthrie Yarwood Concurrent Validity Group A shows more openness to experience Group B shows less openness to experience © 2015 M. Guthrie Yarwood Concurrent Validity High IQ group shows more openness to experience Low IQ group shows less openness to experience © 2015 M. Guthrie Yarwood Predictive Validity – A type of criterion-related validity Predictive Validity: how accurately test scores predict criterion scores. Indicated by correlation between test score (the predictor) and a criterion of future performance (what the test predicts) © 2015 M. Guthrie Yarwood Predictive Validity Openness to Experience (Predictor) © 2015 M. Guthrie Yarwood ????? (Criterion) Predictive Validity Openness to Experience College Major Openness to Experience Well-being, Psychological Adjustment © 2015 M. Guthrie Yarwood Construct-Related Validity Definition: extent to which scale measures a particular construct or psychological concept To assess: need to determine whether an assessment instrument that presumably measures a certain personality variable is actually doing so © 2015 M. Guthrie Yarwood Convergent Validity – A Type of Construct-Related Validity Convergent Validity: the measure has high correlations with other measures or methods of measuring the same construct © 2015 M. Guthrie Yarwood Convergent Validity Our Openness to Experience Self-Report Scale Our Openness to Experience Self-Report Scale © 2015 M. Guthrie Yarwood Same Construct, Same Measurement Same Construct, Different Measurement Convergent Validity Our Openness to Experience SelfReport Scale Our Openness to Experience SelfReport Scale © 2015 M. Guthrie Yarwood Costa & McCrae’s Self-Report OE Dimension Observer-Report OE Dimension Discriminant Validity – A Type of Construct-Related Validity Discriminant Validity: the measure has low correlations with measures of different constructs © 2015 M. Guthrie Yarwood Discriminant Validity Our Openness to Experience Self-Report Scale Different Construct, Same Measurement OR Our Openness to Experience Self-Report Scale © 2015 M. Guthrie Yarwood Different Construct Different Measurement Discriminant Validity Our Openness to Experience Self-Report Scale Our Openness to Experience Self-Report Scale © 2015 M. Guthrie Yarwood Self-Report Sensation Seeking Scale Clinical Diagnosis of Schizotypal PersonalityDisorder Discriminant Validity Our Openness to Experience Self-Report Scale Self-Report Sensation Seeking Scale These correlations should be significant, but lower than correlations for convergent validity. Our Openness to Experience Self-Report Scale © 2015 M. Guthrie Yarwood Clinical Diagnosis of Schizotypal PersonalityDisorder To assess convergent/discriminant validity The same construct using the same method (c) The same construct using different methods (c) Different constructs using the same method (d) Different constructs using different methods (d) Note: c = convergent; d=discriminant © 2015 M. Guthrie Yarwood Reliability & Validity A test can be reliable, but not valid In other words, to be valid a test must first be reliable. NO VALDITY WITHOUT RELIABILITY! © 2015 M. Guthrie Yarwood Reliability and Validity: Summary Internal Validity Reliability Test-Retest Coefficient Parallel-Forms Coefficient Internal Consistency Coefficient Interrater (interobserver) Reliability © 2015 M. Guthrie Yarwood Content Validity Criterion-Related Validity Concurrent Validity Predictive Validity Construct Validity Convergent Validity Discriminant Validity A LITTLE MORE ON OBSERVER REPORTS (O-DATA) © 2015 M. Guthrie Yarwood Correlation • Self-report • Observer report 1 1 Correlation • Observer report 1 • Observer report 2 2 Correlation • Self-report • Several different observers 3, 4, 5… © 2015 M. Guthrie Yarwood Aggregating Scores Averaging the self-report and the observer report/s provides a clearer picture of personality than the self or observer report alone (Kolar et al., 1996) © 2015 M. Guthrie Yarwood Let’s discuss the purpose of observer reports What is the purpose of an observer report? Compared to a self-report measure, what results would you expect from an observer report? What would you think if you did not obtain these results? Can you think of any threats to validity that observer reports may pose? © 2015 M. Guthrie Yarwood Aggregating Scores A positive correlation between an observer report and a self-report, would provide evidence for which type of validity? A. Convergent B. Predictive C. Concurrent D. Discriminant © 2015 M. Guthrie Yarwood Recall Reflection Post and TIPI! 1 Extraverted, enthusiastic 2 Reserved, quiet r 3 Critical, quarrelsomer 4 Sympathetic, warm 5 Dependable, self- disciplined 6 Disorganized, carelessr 7 Anxious, easily upset r 8 Calm, emotionally stable 9 Open to new experiences, complex 10 Conventional, © 2015 M. Guthrie Yarwood r uncreative Disagree Disagree Disagree a Strongly Moderately little (1) (2) (3) Neither Agree nor Disagree (4) Agree a Agree Agree little Moderately Strongly (5) (6) (7) Self Scores! 7 6 5.71 5.54 5.13 5 4 4.46 3.96 3 2 1 Extraversion Agreeableness Conscientiousness TIPI Dimensions © 2015 M. Guthrie Yarwood Emotionally Stable Openness to Experience Observer Scores! 7 6 4.92 5 5.25 5.13 4.38 4.29 4 3 2 1 Extraversion Agreeableness Conscientiousness TIPI Scores © 2015 M. Guthrie Yarwood Emotionally Stable Openness to Experience Based on the following graphs, which dimension do you think shows the weakest correlation between self and observer scores? A. Extraversion B. Agreeableness C. Conscientiousness D. Emotional Stability E. Openness to Experience © 2015 M. Guthrie Yarwood AGREEABLENEESS 7 6 5 4 3 2 1 1 2 3 4 5 6 7 SELF SCORE OBSERVER SCORE OBSERVER SCORE EXTRAVERSION 7 6 5 4 3 2 1 1 2 OBSERVER SCORE OBSERVER SCORE 7 5 4 3 2 1 3 4 5 6 7 SELF SCORE OBSERVER SCORE 6 7 7 6 5 4 3 2 1 1 2 3 4 SELF SCORE Openness to Experience 7 6 5 4 3 2 1 1 2 3 4 SELF SCORE © 2015 M. Guthrie Yarwood 5 Emotional Stability LOW NEUROTICISM 6 2 4 SELF SCORE Conscientiousness 1 3 5 6 7 5 6 7 Based on the following graphs, which dimension do you think shows the weakest correlation between self and observer scores? A. Extraversion B. Agreeableness C. Conscientiousness D. Emotional Stability E. Openness to Experience © 2015 M. Guthrie Yarwood LET’S CORRELATE THE SCORES! TIPI DIMENSION 1 EXTRAVERSION r Extraverted, enthusiastic .74 Reserved, quiet r 2 3 TIPI TRAIT AGREEABLENESS Critical, quarrelsomer 4 Sympathetic, warm 5 Dependable, self-disciplined CONSCIENTIOUSNESS 6 Disorganized, carelessr 7 Anxious, easily upset r 8 (LOW NEUROTICISM) 9 OPENNESS TO 10 © 2015 M. Guthrie Yarwood EMOTIONAL STABILITY Calm, emotionally stable .48 .52 .07 Open to new experiences, EXPERIENCE complex Conventional, uncreativer .39 AGREEABLENESS 7 7 6 6 SCORE SCORE EXTRAVERSION 5 4 3 5 4 3 2 2 1 1 1 2 3 4 5 6 7 8 9 10 11 1 12 2 4 5 6 7 8 9 10 11 12 STUDENT STUDENT EMOTIONAL STABILITY (LOW NEUROTICISM) CONSCIENTIOUSNESS 7 7 6 6 5 SCORE SCORE 3 4 3 2 5 4 3 2 1 1 1 2 3 4 5 6 7 8 STUDENT 9 10 11 12 1 2 3 4 5 6 OPENNESS TO EXPERIENCE 7 SCORE 6 5 4 3 2 1 1 Note. Red = Self; Yellow = Observer 2 3 4 © 2015 M. Guthrie Yarwood 5 6 7 STUDENT 8 9 10 7 STUDENT 11 12 8 9 10 11 12 AGREEABLENESS 7 7 6 6 SCORE SCORE EXTRAVERSION 5 4 3 5 4 3 2 2 1 1 1 2 3 4 5 6 7 8 9 10 11 1 12 2 4 5 6 7 8 9 10 11 12 STUDENT STUDENT EMOTIONAL STABILITY (LOW NEUROTICISM) CONSCIENTIOUSNESS 7 7 6 6 5 SCORE SCORE 3 4 3 2 5 4 3 2 1 1 1 2 3 4 5 6 7 8 STUDENT 9 10 11 12 1 2 3 4 5 6 OPENNESS TO EXPERIENCE 7 SCORE 6 5 4 3 2 1 1 Note. Red = Self; Yellow = Observer 2 3 4 © 2015 M. Guthrie Yarwood 5 6 7 STUDENT 8 9 10 7 STUDENT 11 12 8 9 10 11 12 Summary and Evaluation Decisions about data source and research design depend on (1) the purpose of study and (2) threats to validity/reliability There is no perfect data source There is no perfect research design Assessing threats to reliability and validity will assist in selecting a data source and research design. Observer reports improve validity and reduce social desirability concerns © 2015 M. Guthrie Yarwood Reminder – Paper Topic!! Find a group member (can do on ANGEL!) Select a topic! Submit Paper topic to drop-box Due: Friday, January 30th at 9 AM ET Paper Topic Outline available on ANGEL. Guidance: Meet with Michelle or Celina Questions??? © 2015 M. Guthrie Yarwood Chapter 3 (Dispositional Domain) Survey HEXACO Follow link to access survey Score your survey according to the instructions on ANGEL Big Five Model Access through Chapter 3 Survey on course website © 2015 M. Guthrie Yarwood