Psych Ass: Test

Question 1

Classical Test Score Theory

Answer 1

Assumes every person has a true score with no errors. Observed = True + Error

Question 2

Domain Sampling Theory

Answer 2

Items selected from an infinite domain of choices

Question 3

Domain Sampling

Answer 3

Choosing items appropriate for the domain

Question 4

Domain Sampling Model

Answer 4

Considers using limited items for a larger construct

Question 5

Item Response Theory

Answer 5

Focuses on difficulty, moving from easy to harder questions

Question 6

Reliability

Answer 6

Scores are stable and consistent

Question 7

Reliability Coefficient

Answer 7

Ratio of variance between true and observed score

Question 8

Test-Retest Method

Answer 8

Considers consistency, administers the test in different occasions

Question 9

Parallel Forms

Answer 9

Compares two equal forms of tests for the same attribute (different items, same difficulty)

Question 10

Internal Consistency

Answer 10

Consistency with the same form & subset but different approach based on variability

Question 11

Carry Over Effect

Answer 11

First session influences the score for the second session if tests are taken close together

Question 12

Systematic Carryover

Answer 12

Affects all test takers equally, no net error

Question 13

Split Half Method

Answer 13

Divided into halves scored separately

Question 14

Split Half Reliability

Answer 14

Dividing sets of related items into two and comparing the results

Question 15

Coefficient/Cronbach’s Alpha

Answer 15

Internal consistency of a scale (closeness of items to a group)

Question 16

KR-20

Answer 16

Reliability of binary measurements (yes/no)

Question 17

Kappa Statistics

Answer 17

Consistency between the agreement of different raters (observed - expected / 1 - expected)

Question 18

Spearman-Brown Formula

Answer 18

Estimates the correlation between two halves

Question 19

Low Alpha Level

Answer 19

Less information

Question 20

Alpha Coefficient

Answer 20

Marks only the lower bound in reliability

Question 21

KR-20

Answer 21

Considers all possible ways of splitting items

Question 22

Kappa

Answer 22

Best used for assessing levels of agreement

Question 23

Discriminability Analysis

Answer 23

Examines the correlation of individual items and total score

Question 24

Validity

Answer 24

Agreement between a score and what is believed to measure

Question 25

Construct Evidence

Answer 25

Criterion, Content

Question 26

Face Validity

Answer 26

Mere appearance, related to perceived purpose

Question 27

Content Validity

Answer 27

Logical rather than statistical

Question 28

Construct Underrepresentation

Answer 28

Failure to capture important components of a construct (e.g., math test without geometry)

Question 29

Construct-Irrelevant Variance

Answer 29

Unrelated factors affecting scores

Question 30

Criterion Validity Evidence

Answer 30

Shows how well a test aligns with a specific standard

Question 31

Predictive Validity Evidence

Answer 31

Indicates how well the test can forecast future performance

Question 32

Concurrent Validity Evidence

Answer 32

Shows how well a test aligns with performance measured at the same time

Question 33

Validity Coefficient

Answer 33

Expression of the relationship between a test and criterion

Question 34

Construct Validity

Answer 34

Built by defining a concept and creating a tool to measure it, showing the relationship between the test and other measures

Question 35

Convergent Evidence

Answer 35

When a measure works well with other tests measuring the same concept

Question 36

Discriminant Evidence

Answer 36

Confirms uniqueness by showing a low correlation of unrelated constructs

Question 37

Norm-Referenced Test

Answer 37

Compares and ranks test takers in relation to one another

Question 38

Criterion-Referenced Test

Answer 38

Evaluating against a set of pre-specified criteria