Coun 224 Session2
- 1. More on Assessment Part 2: Technical and Methodological Principles
- 2. Chapter 3 Norms and the Meaning of Test Scores Raw scores and percentages are useless in isolation Scores have meaning only when compared to Norms Norms come from a sample that tells us what individuals in a “representative group” would do Where does an individual fall in the distribution?
- 3. Norms: Types and Purposes Purposes Compare different areas within a person Place a person within a group Types Developmental level obtained Position within a group
- 4. Important Statistical Concepts Frequency distribution Distribution curve Histogram Frequency Polygon Normal Curve
- 5. Measures of Central Tendency Mean Median Mode
- 6. Measures of Central Tendency 1, 4, 7, 3, 3, 2, 5, 6, 4, 3, 5, 5, 2, 4, 6, 4 Frequency Distributions
- 7. Measures of Central Tendency Mean, Median, and Mode all equal 4 Normal distribution
- 8. Variability Range Deviation From the mean Variance Standard Deviation Square root of variance
- 9. Standard Deviation Allows for easiest comparison of scores Jane is 1.5 standard deviations from the mean John is .5 standard deviations bellow the mean Jim is 1 standard deviation above the mean in math and 2 standard deviations above the mean in verbal ability Figure 3.3 on page 54
- 10. Developmental Norms Mental Age Grade Equivalents Ordinal Scales for child development for babies…less judgment imposed…more descriptive
- 11. Within-Group Norms Percentiles Every 10% of population is a percentile Radiates out from mean—fig. 3.4 on pg. 59 Standard Scores Using formula on pg. 61, Mean=0 and SD=1 May be modified to avoid negative numbers Deviation IQs A standard score Mean of 100 SD of 16
- 12. Within-Group Norms Stanines Divides the normal curve into nine categories Table 3.4 on pg. 63 Interrelationships of Within-Group Scores Figure 3.6 on pg. 67
- 13. Relativity of Norms Normative Sample Population vs Sample Another warning about representation Interest Comparisons Keep in mind that, even if two tests measure the same thing, different norms mean that 118 may equal 112 or 145 National Anchor norms-tables to compare tests Specific Norms—admit limitations and norm test to narrower groups Fixed reference groups—norms never change (SAT)
- 14. Computers and Test Scores Allows for complex methods of testing Allows for narrative interpretations Allows for easier scoring MMPI templates Over reliance on computer interpretations Compatibility of Scores
- 15. Domain Referenced Interpretation What has been mastered instead of how does this individual compare to others.
- 16. Minimum Qualifications and Cutoff Scores Should be based on Criterion Comparison Remember false positives and false negatives Table 3.6 pg. 82
- 17. Chapter 4 Reliability All about consistency Test score = true score + or – error Measured variance= TV + or - EV
- 19. Correlation Coefficients What do they tell us? Statistical Significance Confidence Intervals
- 20. Reliability Coefficients Are Correlation Coefficients Types of Reliability Coefficients Test-Retest Reliability Alternate Form Reliability Split-Half Reliability Spearmand-Brown Formula on pg. 96 Interitem Reliability Scorer Reliability Speed Tests vs. Power Tests
- 21. Reliability and Sample Greater variance in sample allows for higher reliability to be demonstrated Fig. 4.5 on pg. 106
- 22. Chapter 5 Validity: Basic Concepts Content Validity Face Validity Criterion-Prediction Procedures Concurrent Validity (diagnostic) Predictive Validity Construct Validity Theoretically derived
- 23. More on Construct Validity Developmental Changes Correlations with other tests Factor Analysis Internal Consistency Convergent and Discriminant Validation Interventions
- 24. Chapter 6 Validity: Measurement and Interpretation Test selection Test interpretation
- 25. Validity Coefficient Validity Coefficient is a correlation coefficient of assessment and criterion Dependent on the group Valid for 9 th grade does not mean valid for 12 th Acceptable magnitude will depend on purpose Decision Theory—false accepts and false rejects Sequential testing to improve accuracy Moderator Variables Variables that change the validity
- 26. Combining Information from Tests Multiple Regression pg. 157 Provides a single value Allows for compensation Profile Analysis with cutoff scores Requires review of multiple values Does not allow for compensation
- 27. Uses of Tests Selection Placement Classification Test Bias pg. 167 Slope bias Intercept bias
- 28. Chapter 7 Item Analysis Useful both in evaluation tests and in creating tests (including classroom exams and quizzes) Can allow for a shorter test to be more reliable and valid than a longer one by removing useless items
- 29. Item Analysis Item Difficulty Distribution of Test Scores Item Discrimination Use of extreme groups Item Response Theory (IRT) pgs. 189 and 190 Item Analysis of Speeded Tests Cross Validation Differential Item Functioning
- 30. Item Analysis Cross Validation Use one (or more than one) sample for item selection Use different samples to validate the test Keeps error random instead of allowing it to become bias Differential Item Functioning