Chapter 12: Testing and Assessment Chapter 13: Research and Evaluation © 2007 Thomson Brooks/Cole, a division of Thomson Learning Testing and Assessment © 2007 Thomson Brooks/Cole, a division of Thomson Learning Testing: a subset of assessment Assessment includes: Informal Assessment Personality Testing Ability Testing The Clinical Interview See Figure 12.1, p 396 © 2007 Thomson Brooks/Cole, a division of Thomson Learning You will be administering and interpreting assessment instruments You may consult with others on their proper use You may use them in program evaluation and research You will read about them in the professional literature School counselors: Sometimes the only expert on assessment in the schools Other counselors: Will likely be using them in your setting and consulting with others who use them Why testing? Why not testing? Testing is an additional method of gaining information about your client © 2007 Thomson Brooks/Cole, a division of Thomson Learning 2200 BCE: Chinese developed essay type test for civil service employees Darwin, set the stage for modern science and the examination of differences Wundt, Fechner: 1st experimental labs to examine differences in people Binet: Hired by Ministry of Public Education in France to develop intelligence test Binet test, later became “Stanford Binet”—revised by Terman © 2007 Thomson Brooks/Cole, a division of Thomson Learning Spread of testing at beginning of 20th century: Psychoanalysis spurred on development of objective and projective personality tests Industrial Revolution and need for vocational assessment WWI: Ability and personality tests used to determine placements of recruits 1940s and 1950s: advances in statistics led to better test construction 1980s and on: Personal computers make tests easier to develop, analyze, use, administer, and interpret © 2007 Thomson Brooks/Cole, a division of Thomson Learning Ability Testing (Testing in the Cognitive Domain) (see Figure 12.2, p 399) Two types ▪ Achievement Testing (What one has learned) ▪ Aptitude Testing (What one is capable of learning) Achievement Testing ▪ Survey Battery Tests ▪ Diagnostic Tests (see Box 12.1, p 400: PL 94-142) ▪ Readiness Tests © 2007 Thomson Brooks/Cole, a division of Thomson Learning Ability Testing (Testing in the Cognitive Domain) (see Figure 12.2, p 399) (Cont’d) Aptitude Tests (What one is capable of learning) ▪ Intellectual and Cognitive Functioning Testing ▪ Intelligence Tests ▪ Neuropsychological Assessment ▪ Cognitive Ability Tests ▪ Special Aptitude Tests ▪ Multiple Aptitude Tests © 2007 Thomson Brooks/Cole, a division of Thomson Learning Personality Assessment (Testing in the Affective Domain; see Figure 12.3, p 399) Objective Tests Projective Tests Interest Inventories Informal Assessment (see Figure 12.4, p 399) Observation Rating Scales (see Box 12.2, p 404) Classification Systems (see Box 12.3) Environmental Assessment Records and Personal Documents Performance-Based Assessment © 2007 Thomson Brooks/Cole, a division of Thomson Learning The Clinical Interview Sets a tone for the types of information that will be covered during the assessment process Allows client to become desensitized to information that can be very intimate and personal Allows examiner to assess nonverbals of client while he or she is talking about sensitive information Allows examiner to learn problem areas firsthand Gives client and examiner opportunity to study other’s personality style to assure they can work together © 2007 Thomson Brooks/Cole, a division of Thomson Learning 10 Four Types Validity: Is the test measuring what it’s supposed to measure? Reliability: Is the test accurate (consistent) in its measurement? Practicality: Is this a practical test to use? Cross-Cultural Fairness: Has the test been shown to be fair across different cultures? © 2007 Thomson Brooks/Cole, a division of Thomson Learning 16 Three types Content Criterion-Related ▪ Concurrent ▪ Predictive Construct ▪ ▪ ▪ ▪ Experimental Convergent Discriminant Factor Analysis © 2007 Thomson Brooks/Cole, a division of Thomson Learning Face validity Not a “real” type of validity Does the test, on the surface, seem to measure what it’s supposed to measure Some tests may be valid, but may not seem to be measuring what it’s supposed to measure 17 Is bias removed—as best as possible? Does it predict well for all cultural groups? Griggs v Duke Power Company: Tests must show that they can predict for job performance A number of ethical and legal issues have been addressed (see later under “Ethical, Professional, and Legal Issues”) See Table 12.2, p.417: Summary of Types of Validity and Reliability © 2007 Thomson Brooks/Cole, a division of Thomson Learning 18 Four Types: Test-Retest Alternate (Parallel; Equivalent) Forms Split-Half (Odd-Even) Internal Consistency ▪ Cronbach’s Coefficient Alpha ▪ Kuder-Richardson © 2007 Thomson Brooks/Cole, a division of Thomson Learning 19 Is this a realistic test to give? Based on: Cost Time to administer Ease of administration Format of test Readability of test Ease of interpretation © 2007 Thomson Brooks/Cole, a division of Thomson Learning 20 Over 4000 assessment procedures How you find them: Publisher resource catalogs Journals Source Books and On-Line Source “Book” Information ▪ Buros Mental Measurement Yearbook ▪ Tests in Print Books on Testing and Assessment Experts The Internet © 2007 Thomson Brooks/Cole, a division of Thomson Learning 21 Info usually included: Demographic information Reason for referral Family background Other relevant information (e.g., legal, medical, vocational) © 2007 Thomson Brooks/Cole, a division of Thomson Learning Behavioral observations Mental status Test results Diagnosis Recommendations 10.Summary 22 Usually a few pages long Problems with: Overuse of jargon Focusing on assessment procedures & downplaying person Focusing on person and downplaying assessment results Poor organization Poor writing skills Failure to take a position Demographics © 2007 Thomson Brooks/Cole, a division of Thomson Learning 23 Caution in Using Assessment Procedures Cultural bias continues to exist in testing Standards and ethical codes have been developed to help us: ▪ Understand the cultural bias inherent in tests ▪ Know when a test should not be used due to bias ▪ Know what to with test results when a test does not predict well for minorities Standards for effective use of assessment instruments Association for Assessment in Counseling’s Standards for Multicultural Assessment Code of Fair Testing in Education ACA Ethics Code © 2007 Thomson Brooks/Cole, a division of Thomson Learning 24 Take A Stand—Do Something! Our duty and moral responsibility to something when ▪ Tests have been administered improperly ▪ Tests are culturally biased and the bias is not addressed ▪ Cheating has taken place ▪ Tests were used with limited validity or reliability © 2007 Thomson Brooks/Cole, a division of Thomson Learning 25 Ethics Guidelines for use of assessment instruments(see bottom p 420) Informed consent Invasion of privacy and confidentiality Competence in the use of Tests ▪ Levels A, B, and C Technology and Assessment ▪ Sometimes, counselor not used with computer-generated reports ▪ Issues of confidentiality and privacy ▪ Knowing laws relative to the impact of on-line technology ▪ Adequate training in technology © 2007 Thomson Brooks/Cole, a division of Thomson Learning 26 ETHICAL ISSUES PROFESSIONAL ISSUES Professional Issues Ethics (Other Issues) Proper release of test Computer-Driven Assessment results Reports Selecting Tests ▪ Can be very good Administering, Scoring ▪ Make sure they reflect “you” and Interpreting Tests Professional Association Keeping Tests Secure ▪ Assoc for Assessment in Picking up-to-date tests Counseling and Education (AACE) Proper Test Construction ▪Adivision of ACA © 2007 Thomson Brooks/Cole, a division of Thomson Learning 27 Legal Issues Americans with Disabilities Act: Accommodations must be made when taking tests for employment (FERPA) Buckley Amendment: Right to access school records, including test records Carl Perkins Act (PL98-524): Right to vocational assessment, counseling, and placement for disadvantaged Civil Rights Act (‘64) & Amendments: Tests must be shown to be valid for the job © 2007 Thomson Brooks/Cole, a division of Thomson Learning 28 Legal Issues (Cont’d) Freedom of Information Act: Right to access federal records, including test records PL94-142 and IDEIA: Right of students to be tested, at school’s expense, for a suspected disability that interferes with learning Section 504 of Rehabilitation Act: Instruments must measure person’s ability, not be a reflection of his or her disability HIPAA: Right of privacy of records, including test records © 2007 Thomson Brooks/Cole, a division of Thomson Learning 29 Assessment of clients is not just giving a test Use multiple methods and be wise Remember, people can and will change over time Don’t view them as “stagnant” and always the same © 2007 Thomson Brooks/Cole, a division of Thomson Learning 30 [...]... takers who took the same test, but are in different norm groups to compare their results ▪ Allow an individual to compare his or her results on two different tests © 2007 Thomson Brooks/Cole, a division of Thomson Learning 12 Some statistics help us make meaning of test scores Measures of Central Tendency ▪ Mean ▪ Median ▪ Mode Measures of Variability ▪ Range ▪ Interquartile Range ▪ Standard Deviation... researched as standardized tests (e.g., teacher made tests) See Table 12. 1, p 407 © 2007 Thomson Brooks/Cole, a division of Thomson Learning 11 Relativity and Meaningfulness of Scores Raw scores don’t hold much meaning unless you do something to them By comparing raw scores to those of an individual’s peer group, you are able to: ▪ See how the individual did in comparison to similar people... Brooks/Cole, a division of Thomson Learning 14 A basic statistic not directly related to interpretation of test but crucial in test construction Ranges from -1.0 to +1.0 The closer to -1.0 and +1.0 the strong the relationship between variables Positive correlation: tendency for two sets of scores to be related in same direction Negative correlation: tendency for two sets of scores to be related in... well for all cultural groups? Griggs v Duke Power Company: Tests must show that they can predict for job performance A number of ethical and legal issues have been addressed (see later under “Ethical, Professional, and Legal Issues”) See Table 12. 2, p.417: Summary of Types of Validity and Reliability © 2007 Thomson Brooks/Cole, a division of Thomson Learning 18 Four Types: 1 Test-Retest 2 Alternate... instruments(see bottom p 420) Informed consent Invasion of privacy and confidentiality Competence in the use of Tests ▪ Levels A, B, and C Technology and Assessment ▪ Sometimes, counselor not used with computer-generated reports ▪ Issues of confidentiality and privacy ▪ Knowing laws relative to the impact of on-line technology ▪ Adequate training in technology © 2007 Thomson Brooks/Cole, a division of Thomson... 26 ETHICAL ISSUES PROFESSIONAL ISSUES Professional Issues Ethics (Other Issues) Proper release of test Computer-Driven Assessment results Reports Selecting Tests ▪ Can be very good Administering, Scoring ▪ Make sure they reflect “you” and Interpreting Tests Professional Association Keeping Tests Secure ▪ Assoc for Assessment in Picking up -to- date tests Counseling and Education (AACE)... be a reflection of his or her disability HIPAA: Right of privacy of records, including test records © 2007 Thomson Brooks/Cole, a division of Thomson Learning 29 Assessment of clients is not just giving a test Use multiple methods and be wise Remember, people can and will change over time Don’t view them as “stagnant” and always the same © 2007 Thomson Brooks/Cole, a division of Thomson Learning... Predictive 1 Construct ▪ ▪ ▪ ▪ Experimental Convergent Discriminant Factor Analysis © 2007 Thomson Brooks/Cole, a division of Thomson Learning Face validity Not a “real” type of validity Does the test, on the surface, seem to measure what it’s supposed to measure Some tests may be valid, but may not seem to be measuring what it’s supposed to measure 17 Is bias removed—as best as possible? Does... ethical codes have been developed to help us: ▪ Understand the cultural bias inherent in tests ▪ Know when a test should not be used due to bias ▪ Know what to do with test results when a test does not predict well for minorities Standards for effective use of assessment instruments Association for Assessment in Counseling s Standards for Multicultural Assessment Code of Fair Testing in Education ... Measures of Variability ▪ Range ▪ Interquartile Range ▪ Standard Deviation ▪ See Figure 12. 5, page 409 ▪ See Figures 12. 6 and 12. 7; page 410 and 411 © 2007 Thomson Brooks/Cole, a division of Thomson Learning 13 TYPES OF DERIVED SCORES TYPES OF DERIVED SCORES Percentile Rank Normal Curve Equivalents (NCEs) T-Scores Stanines Deviation IQ Sten Scores SAT/GRE Type Scores Grade Equivalent Scores ... information Allows examiner to learn problem areas firsthand Gives client and examiner opportunity to study other’s personality style to assure they can work together © 2007 Thomson Brooks/Cole,... consult with others on their proper use You may use them in program evaluation and research You will read about them in the professional literature School counselors: Sometimes the only expert... don’t hold much meaning unless you something to them By comparing raw scores to those of an individual’s peer group, you are able to: ▪ See how the individual did in comparison to similar people