Counselor Quick Reference
Key terms, frameworks, and clinical checkpoints from Ault & Emelianchik-Key, Ch. 6 (Test Selection, Scoring, and Statistics), plus the MCPY 622 M5 lecture on selecting assessments — pulled from the full study guide for fast review before class or clinical work.
Test Selection, Scoring, and Statistics
Glossary
- Quantitative variable
- Counted or measured and expressed numerically — e.g., number of therapy sessions attended.
- Qualitative variable
- Classified into categories rather than numbers — e.g., diagnosis or coping-skill type.
- Discrete variable
- Distinct, countable values, like number of sessions — as opposed to a continuous measurement.
- Continuous variable
- Takes on any value within a range, obtained by measuring — e.g., length of time in therapy.
- Descriptive statistics
- Describe the properties of a sample or population's data — what a score of 20 out of 20 actually means.
- Inferential statistics
- Use sample properties to test hypotheses and draw conclusions about a broader population.
- Raw score
- An unadjusted score with no context about the larger scale — meaningless on its own, without comparison data.
- Norm-referenced test
- Compares an individual's performance to a reference group or population — e.g., the WAIS.
- Criterion-referenced test
- Compares performance to a predetermined standard or skill level, not to other people — e.g., a driver's license exam.
- Social desirability bias
- A test taker answers in a way that looks socially acceptable rather than giving their real opinion.
5-Step Assessment Selection Process
1. Gather information & understand needs
The client is the primary source, always. Also pull from records, court documents, school records, and reports from medical/education professionals (Drummond & Jones, 2010).
2. Determine methods & access information
Narrow down formal vs. informal methods — e.g., a standardized instrument like the SASSI-4 vs. an unstructured interview. Interviews build rapport but risk social desirability bias.
3. Examine available instruments
Counselors rarely know every existing test. Use dedicated resources — MMY, TIP, publisher catalogs, professional associations — to research options systematically.
4. Narrow down
Weigh psychometric properties (reliability, validity, limitations), the training/competency required to administer and score it, available formats, and cost.
5. Select & implement
Consult a supervisor when in doubt. Weigh time cost against quality of result — a longer, pricier assessment may be worth it for a far superior outcome.
Measurement Scales
Nominal
- No order, no true zero
- Categories with no inherent numerical value
- Example: theoretical orientation — CBT, Gestalt, solution-focused
Ordinal
- Ordered, unequal spacing
- Distance between points isn't consistent/known
- Example: "Never / Sometimes / Often / Always" ratings
Interval
- Equal spacing, no true zero
- Zero doesn't mean "absence of the quality"
- Example: Big Five Personality subscale scores
Ratio
- Equal spacing + true zero
- Scores are proportionally meaningful
- Example: age, years of education, session counts
Standard Score Conversions
Z-score & T-score
- Z-score: distance from the mean in SD units. Z = (X−μ)/σ
- T-score rescales Z to mean 50, SD 10: T = 10Z + 50 — avoids negatives/decimals
Stanines & Stens
- Stanines ("standard nines"): 9 categorical bands — Stanine 5 = group mean, 7–9 above average, 1–3 below
- Stens ("standard tens"): same idea, chunked into 10 intervals
Deviation IQ
- Replaced the old mental-age ÷ chronological-age ratio, which broke down past adolescence
- Modern IQ tests (e.g., Stanford-Binet) set mean 100, SD 15: IQ = 15Z + 100
NCE & College Entrance Scores
- NCE (normal curve equivalent): mean 50, SD 21.06 — anchored so 1st/50th/99th percentile line up with NCE 1/50/99
- SAT-style section scores: mean 500, SD 100 (200–800 scale)
Correlation Coefficients
Pearson's r
Two continuous variables, linear relationship (parametric). Example: counseling participation vs. anxiety level — expect a moderately strong negative r.
Spearman's rho (ρ)
Rank-ordered / Likert-type variables (nonparametric). Example: NICHQ Vanderbilt ADHD scales — never / occasionally / often / very often.
Phi coefficient (φ)
Two nominal/categorical variables. Example: blood type & country of origin, or STEM vs. non-STEM major & graduating in 4 years vs. more.
Scoring Methods — Pros & Cons
Hand scoring
- Pro: nuanced judgment on open-ended items; adapts to varied assessment types
- Con: time-consuming, prone to scorer bias/variability, may lack interrater reliability
Computer-based scoring
- Pro: rapid, consistent, reduces human bias, immediate feedback
- Con: limited for complex/creative responses; ignores testing irregularities
Publisher scoring
- Pro: accurate/consistent, less time-consuming, requires less training
- Con: turnaround time to get scores back; added cost
Self-scoring
- Pro: saves administrator time, fast results, promotes test-taker autonomy
- Con: anxiety/distress around scoring; higher error rate; risk of bias or cheating
Clinical checkpoints
- Start test selection with the client's own account, not the instrument catalog — records and referral information come second (Drummond & Jones, 2010)
- Weigh a norm group's cultural/demographic fit alongside reliability and validity before handing a client a norm-referenced test
- Document every testing irregularity (e.g., a fire-alarm interruption) and offer a free retake — never penalize the score
- Match the scoring method to the task: hand-score nuanced/open-ended responses; use computer scoring for speed and consistency on structured items
- Never share a raw score without context — pair it with a norm- or criterion-referenced comparison before delivering it to a client
- Keep percentile rank and percentage distinct: a percentile compares a client to other test-takers, a percentage describes performance on the test itself
- Pick the correlation coefficient by variable type: continuous → Pearson's r, ranked/Likert → Spearman's rho, nominal/categorical → phi
- Don't over-interpret a grade-equivalent score as literal grade-level mastery — it means same-test performance, not curriculum mastery