Chapter 4: of Tests and Testing 12 Assumptions in Psychological Testing and Assessment

Chapter 4: Of Tests and Testing
12 Assumptions in Psychological Testing and Assessment

• Assumption 1: Psychological traits and states exist
• Assumption 2: Psychological traits and states can be quantified and measured
• Assumption 3: Various approaches to measuring aspects of the same thing can be useful
• Assumption 4: Assessment can provide answers to some of life’s most momentous
questions
• 12 Assumptions in Psychological Testing and Assessment
• Assumption 5: Assessment can pinpoint phenomena that require further attention or study.
• Assumption 6: Various sources of data enrich and are part of the assessment process.
• Assumption 7: Various sources of error are part of the assessment process.
• Assumption 8: Tests and other measurement techniques have strengths and weaknesses
• Assumption 9: Test-related behavior predicts non–test-related behavior
• Assumption 10: Present-day behavior sampling predicts future behavior
• Assumption 11: Testing and assessment can be conducted in a fair and unbiased manner
• Assumption 12: Testing and assessment benefit society
Most Controversial Assumption?

• Why?
What are Norms?

• Derived typical test performance of a standardization sample
• The test score distribution that provides the average or typical (normal) score level on a test
Standardization (normative) Sample

• The normative sample is a representative subset drawn from the broader target population
• Typically, a large random sample
• Sample size should be large enough to obtain stable values.
Sampling Techniques
• Randomization
– every case has an equal chance of selection
• Stratified
– representative proportions of groups
– e.g. age, socioeconomic level, ethnicity
• Incidental
– Convenience sampling
– Not a desired procedure
Types of Norms
• Developmental Norms
• Indicates developmental level attained
• Age equivalent norms (Mental Age)
• e.g., a 7 year old who scores the same mean obtained by 10 year old children has a mental
age of 10.
• Grade equivalent norms
• e.g., average score of 4th graders is 23, a child with a raw score of 23 is given a 4th grade
age equivalence.
Norm-Referenced (Within group) Norms

• Individual performance is evaluated in reference to a standardization group
• The same test is used to compare other groups of test-takers
• Deviation IQs
What is Correlation?
• Index of linear association between two variables (X and Y)
• Does not suggest cause and effect
• Computed value is called a coefficient
• Best example is the Pearson product-moment correlation coefficient (r)
Pearson Formula (definitional)

• Co-variation between X and Y
• Ratio of the variability between X and Y
Values of r
• Coefficient values range between -1 and + 1
• What does 0 mean?
– The closer the coefficient value is to 0, the weaker the association between two
variables
• The further a coefficient moves from 0, the stronger the association between two variables
– Coefficients of -1 and +1 have the same magnitude of association
Coefficient of Determination (r2)

• Correlation coefficient squared
• The value indicates the proportion of the variation in Y scores that is a function of the X
scores
• i.e., the variance in X explained by Y
Graphing Correlation
• Correlations between two variables can be displayed in a scatterplot
• Individual scores are plotted on two-dimensional axes
– X scores plotted on horizontal axis (abscissa)
– Y scores plotted on vertical axis (ordinate)
Positively Correlated
• As X increases, Y increases
Negatively Correlated
• As X increases, Y decreases
No Correlation
• No relationship between X and Y
Curvilinear Relationship
• Non-linear relationship between two variables
• The scatterplot has a significant curve
• U-shaped curve
• Umbrella-shaped curve
• S-shaped curve
Other Correlation Coefficients

• Spearman rho
– Used in rank-order correlation
• ordinal scales
• Evaluate the differences (or agreement) between rankings of two variables
– Students’ scores on a mid-term 1 and mid-term 2 are ranked from lowest to highest;
the rankings are correlated
Point Biserial
• Comparison of one continuous variable and one dichotomous variable
• Dichotomous variables include Yes/No or True/False scales
– Correlation between Age (continuous) and Active Class Participation (Yes or No)
Phi
• Correlation between two dichotomous variables
– Correlation between Active Class Participation (Yes or No) and Mid-term results
(Pass or Fail)
Biserial r
• Comparison of one continuous variable and one artificially dichotomized variable
• An artificially dichotomized variable is a continuous variable that is transformed to
dichotomous variable
– e.g., Age in years converted to age groups
• 18-25, 26-30, 31-40, 40-50, 51-60, etc.
Tetrachoric
• Correlation between two artificially dichotomized variables
– Correlation between age groups and mid-term score
Roles of Correlations in Testing

• Test-retest reliability
– Correlation between scores on the same test at two different times
• Correlation of GRE in the Fall and Spring semesters
• Criterion (predictive) validity coefficients
– Correlation between test scores and results of an independent criterion
• Correlation between SAT and College GPA
• Convergent validity coefficients
– Correlation between scores on two conceptually similar tests
• Correlation between self-esteem and self-concept
Regression
• Degree of predictability between two variables
• Extends the concept of correlation to the prediction of a test score (Y) based on a another
test score (X)
Regression Equation
Y’ = a + bX
• X = predictor (test score)
• Y’ = criterion (predicted score)
• a = y-intercept (criterion score if the predictor score is 0)
• b = slope (correlation between the predictor and criterion)
Regression Line
• Line drawn through the scatter of scores
• The regression line represents the Principle of Least Squares
• least squared deviation from the line
• The line demonstrates the best fit for all data points
Slope
• Essentially, the correlation coefficient
Y-Intercept
• Where the regression line crosses the Y-axis
• Criterion score if the predictor score is 0
a = Y – bX
• Y is the mean of the Y scores

• X is the mean of the X scores
Regression Example
Y’ = 2 + 0.67X
• What is the predicted score (Y’) if X is 10?
Regression Line
• Line drawn through the scatter of scores
• The regression line represents the Principle of Least Squares
– least squared deviation from the line
• The line demonstrates the best fit for all data points
• Regression Line Example
Residuals
• Difference between the predicted (Y’) and observed criterion (Y) values
• Y – Y’
• Principle of Least Squares
• Minimize the deviation between Y and Y’
Standard Error of Estimate

• Error in the prediction estimate
– Standard deviation of the residuals
• The square root of the residual variance
• The lower the standard deviation, the lower the degree of error in the regression equation
Inference from Measurement

• Meta-Analysis
– Statistical combination of studies
• Culture and Inference
– Individualists vs. collectivists

Chapter 4: of Tests and Testing 12 Assumptions in Psychological Testing and Assessment

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Chapter 4: of Tests and Testing 12 Assumptions in Psychological Testing and Assessment

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 4: of Tests and Testing 12 Assumptions in Psychological Testing and Assessment

Uploaded by

Copyright:

Available Formats

Chapter 4: Of Tests and Testing

12 Assumptions in Psychological Testing and Assessment

Most Controversial Assumption?

What are Norms?

Standardization (normative) Sample

Norm-Referenced (Within group) Norms

Pearson Formula (definitional)

Coefficient of Determination (r2)

Other Correlation Coefficients

Roles of Correlations in Testing

• Y is the mean of the Y scores

• Regression Line Example

Standard Error of Estimate

Inference from Measurement

You might also like