Test selection, administration, interpretation and fairness

AI-narrated overview made with Google NotebookLM. The written lesson below is by Dr. Andrew Wichterman.

Section E of the ACA Code governs assessment. The exam tests competence to use tests, informed consent for testing, release of results, and fairness across groups.

Competence and test user qualifications

Counselors use only assessments they are trained and competent to administer and interpret, and they take responsibility for the use of computer-scored and interpreted results. Publishers classify tests by qualification level: Level A requires minimal training, such as achievement tests; Level B requires graduate coursework in testing, such as interest inventories and many personality inventories; Level C requires advanced training and supervised experience, such as the Wechsler scales, the MMPI and projective tests. The Responsibilities of Users of Standardized Tests, RUST, from AACE, and the Standards for Educational and Psychological Testing from AERA, APA and NCME are the reference documents.

Selection and informed consent

Choose instruments with evidence of reliability and validity for the purpose and the population, considering the client's language, culture, age and disability. Before testing, explain the nature and purpose of the assessment and the intended use of results in language the client understands, and obtain consent. Clients have the right to know results and to have them explained. When a third party will receive results, the client is told in advance. Assessment for forensic purposes has additional rules: the counselor does not evaluate people they counsel, and the examinee is told the purpose and the absence of a counseling relationship.

Administration and scoring

Administer under standardized conditions and note any departures. Do not use obsolete tests or outdated data. Maintain test security and do not reproduce copyrighted materials. Accommodations for disability should preserve the construct measured and be documented. Computer-adaptive testing tailors item difficulty to the examinee's responses using item response theory, reaching comparable precision with fewer items. Scoring errors are a common source of invalid results; check them.

Interpretation and release

Interpret in light of the client's characteristics, the norms used, the standard error of measurement and the situation; never on the basis of a single score. Release results only to persons qualified to interpret them or with the client's consent, and provide accurate interpretations with any release. Explain results to clients in understandable terms and consider their cultural context. Diagnosis is interpretation too: the Code allows counselors to refrain from diagnosing when it would harm the client and requires attention to cultural bias in diagnosis. Item analysis terms may appear: item difficulty, the proportion answering correctly, ideally near .50; item discrimination, how well an item separates high and low scorers; ceiling and floor effects.

Fairness, bias and multicultural assessment

Test bias is systematic error that disadvantages a group for reasons unrelated to the construct: language load, cultural content, differential item functioning. Fairness is the broader question of equitable use and interpretation. Adverse impact is a legal concept about group differences in selection rates. Culture-fair tests reduce but do not eliminate cultural loading. Counselors consider whether norms include people like the client, whether translation preserves meaning, whether the construct means the same thing in the client's culture, and whether stereotype threat or test anxiety affected performance. The Larry P. v. Riles case restricted IQ testing for placing Black children in special education in California and is the standard citation for test bias in the law.