click below
click below
Normal Size Small Size show me how
Psychometric Methods
Vanderbilt University Final Exam PSY 2530
Question | Answer |
---|---|
Reliability/Validity | Reliability-consistency/repeatability of a measure; three types-test retest, parallel forms, internal consistency Validity- accuracy, does the measure measure what it purports to? Three types-content, criterion-related, construct |
Index of reliability | Empirical- upper bound expectation of what you can expect your measure to correlate w/ anything Theoretical- correlation of scale and a perfect measure of the attribute it assesses |
Correction for attenuation | A way to estimate the correlation between the underlying TRUE scores because there is always some amount of measurement error (unreliability) so this is a way to correct for that error |
Standard error of measurement | SD of observed scores around the true score, increasing reliability shrinks CI’s |
Internal consistency reliability | Function of intercorrelation of items and how many Higher correlation&More tests give higher intercorrelation ratio Degree to which all items measure a common characteristic of the person and are free from measurement error. |
Coefficient alpha | A type of internal consistency reliability, estimates the reliability of a single administration of a single test. Mean of all possible reliabilities test can display given by 1/2ing the test |
KR20 | Special case of coefficient alpha test of internal consistency, average of all possible split-half correlations with dichotomous (T/F) items |
Spearman-Brown Prophecy formula | Demonstrates how reliability of a test changes a function of test length, makes it possible to compute reliability estimate from a single test |
Split-half reliability | Special case of SBP, estimates internal consistency, splits test in half (even/odd etc) and makes it easier to compute the correlation when n=2 |
Test-retest reliability | same test administered multiple times is reliable if the correlation between scores is high |
Parallel forms reliability | Extent to which two tests measure the same concepts at the same difficulty level |
Cattell’s L-Data, Q-Data, T-Data | In order to fully understand a person, test needs to contain Questionnaires (subjective), Test data (objective) and Life-record data (how many kids?) |
Methods of scale construction | Internal-theoretical, items in this scale should be highly correlated External-based on a criterion referenced group’s performance Intuitive- comes from your head, two types Rational-from you head (informal) and Theoretical (from formal theory) |
Ceiling effect | Has a – skew, happens when more people get a high (ceiling) score |
Floor effect | Has a + skew, happens when more people get a low (floor) score |
Ecological validity | Methods, materials, and setting of study approximate the real-world which is being examined, observed behaviors reflect what occurs in natural settings |
Test bias | Systematic over/under prediction that compromises accuracy of inferences |
Trait by treatment interaction | Determining which treatment is best by first determining what level you are at on trait |
Radex | Made up of simplexes and circumplexes map different tests by complexity and content. Items next to one another in same circle =similar complexity different content while items follow the same line out from the middle have similar content but diff complex |
Circumplex | The circle part of the radix, items in same circumplex have different content @ same complexity level |
Simplex | Lines radiating out from radix, items in same simplex, have same content @ different complexity levels, complexity increases as you move into center |
Holland’s hexagon | Six themes of vocational interests that represent 6 different types of individuals and 6 types of work environments, Realistic, Investigative, Artistic, Social, Enterprising, Conventional (RIASEC) |
Self-directed search | The test to determine which 3 themes of Holland’s hexagon one scores highest on, which gives them a summary code (RIA) that corresponds with a list of occupations. Designed to be used for career exploration w/o a career counselor |
Positive and Negative Affect | A dichotomous scale. Positive affect ranges from Joy (high) to depression (low) negative affect from anxiety (high) to relief (low) +/- reinforcement/punish moves you from one quadrant to another. People who have high – and low + are seen for depression |
Satisfaction/Satisfactoriness | both necessary for determining tenure Satisfaction-corresponds with interests, needs, and reinforcement in environment Satisfactoriness- corresponds between ability and actual job requirements |
Personality styles (4) | basic ways individuals respond Celerity-quickness of response interacting with environment Pace-level of effort typically expended in interaction w/ environment Rhythm-pattern of pace Endurance-likelihood of maintaining interaction w/ environment |
Meta analysis | Combining Lots of findings from different samples and populations on same construct to get more precise CI on effect size or correlation, allows to generalize to larger phenomenon |
Validity generalization | Capacity of an instrument to be generalized to other populations |
Formative evaluation | Assessment techniques that guide the course of instruction |
Summative evaluation | Formal assessments are used after instruction is completed to summarize what students have learned |
Halo errors | categorize individual on a small number of salient observations and apply that general impression to all traits, positively or negatively |
Generosity errors | Human tendency to rate someone higher than they deserve because of unwillingness to mark someone as average or below average even |
Response set | Source of error, A tendency on part of test taker to approach test in manner that distorts test (saying yes too much because it might sometimes apply etc) |
Social desirability | Source of error, category of response sets with Tendency to choose responses that they believe are more socially accepted |
Projective hypothesis | Assumption that core concerns/conflicts color every perception and action for projective assessment when give an ambiguous stimulus to a person, they will project what is happening with them, a way to see what people are thinking about (ink blots etc) |
Validity scales (including VRIN and TRIN from Katie’s lectures) | A subscale of a test designed to assist in judgments regarding how honestly the test taker responded and whether or not observed responses were products of response style, carelessness, deliberate efforts to deceive, or unintentional misunderstanding. |
Index scores | Allow for comparisons w/ same aged peers, Mean=100, SD=15 |
Wechsler scales of intelligence | 4 index score=full scaleIQ VCI-verbal comprehension PRI- perceptual reasoning removes processing General ability index-reasoning/comprehension WMI- working memory PSI- processingspeed Cognitive Processing Index-relies on speed efficiency |
Confidence interval | If we repeat the test over and over, their score will be included _% of the time |
Flynn effect | Observed IQ increase of 3 points/decade during 20th century, greatest effect in lower portion of IQ distribution (better nutrition, parental literacy, test sophistication) |
What are two Validity scales from MPQ? | VRINvariable response inconsistency SIMILAR items endorsed opposite (feel good-T, feel great-F) Highscore=random TRINTrue response inconsistency OPPOSITE items endorsed same(feel good-T,feel bad-T) High score=all true Low=all false (<30, >70 can’t use) |
(VCI, PRI, WMI, PSI, FSIQ, GAI, CPI)? | Verbal comprehension+Perceptual reasoning=General ability indes Working memory+Processing speed=cognitive processing indes =Full scale IQ |
Discuss the psychological significance of positive and negative affect. | Individual can be placed on graph at any point in time depending on their mood, individuals with high – affect and low + affect are seen for depression typically. Positive and negative reinforcement and punishment can move you around the graph |
How does the model of talent development discussed in class functions to predict job/career tenure? | Matches interests and abilities to determine how satisfactory an individual would be in in a position and how high his satisfaction would be, if matched correctly, should predict job tenure. Puts = stress on individual and the environment. |
Discuss two ways in which ability tests can be psychologically close (or covary). | In reference to the Radex, they can be highly correlated (physically close ) with similar levels of complexity (on same circumplex) or similar content (same simplex) Complexity increases as you move towards the center. |