Complete exam-focused revision notes on test types, construction, item analysis, standardization, intelligence, personality, attitudes, computer-based testing and real-world applications.
किसी test का असली मूल्य उसके questions में नहीं, बल्कि इस बात में है कि उसके scores कितने विश्वसनीय, वैध और न्यायपूर्ण हैं।
Try a broader term such as “validity”, “intelligence” or “attitude”.
What a psychological test is, and the major ways tests are classified
A psychological test is a standardized, objective and systematic procedure for obtaining a sample of behaviour and describing it with scores or categories.
यह पूरे व्यक्ति को सीधे नहीं मापता; यह व्यवहार का एक sample लेकर किसी construct के बारे में inference बनाता है।
| Classification lens | Type A | Type B | Fast distinction |
|---|---|---|---|
| Administration | Individual | Group | Depth vs economy |
| Time/difficulty | Speed test | Power test | Time pressure vs item difficulty |
| Response | Verbal / performance | Paper-pencil / computer | Mode of responding |
| Interpretation | Norm-referenced | Criterion-referenced | Compared with people vs fixed standard |
| Scoring | Objective | Subjective | Fixed key vs trained judgment |
Name → Order → equal Intervals → true zeRo.
ऊपर जाते हुए हर level नीचे वाले की properties को अपने साथ रखता है।
From defining the construct to publishing the manual
Blueprint content validity की पहली सुरक्षा-दीवार है।
MCQ, true/false, matching.
Strength: objective and efficient scoring.
Risk: guessing and cueing.
Short answer, essay, performance task.
Strength: samples organization and production.
Risk: scorer subjectivity; needs a rubric.
Difficulty, discrimination, distractor quality and item bias
p = R / N
Proportion answering correctly. Higher p = easier item.
For dichotomous items, q = 1 − p.
D = pᵤ − pₗ
How strongly the item separates high scorers from low scorers.
Positive high D is desirable; negative D is a warning.
Point-biserial correlation is used when item score is truly dichotomous (0/1) and total score is continuous.
Corrected item-total correlation excludes that item from the total.
At p = .50, a dichotomous item has maximum response variance p(1−p), often useful for discrimination.
| Indicator | Common interpretation | Decision clue |
|---|---|---|
| p > .80 | Very easy | Keep only if essential or for confidence/warm-up. |
| p ≈ .30–.70 | Moderate | Often most useful for norm-referenced discrimination. |
| p < .20 | Very difficult | Check ambiguity, content mismatch or key error. |
| D ≥ .40 | Very good | Usually retain. |
| D .20–.39 | Moderate | Review or revise. |
| D < 0 | Reverse discrimination | Investigate miskey, ambiguity or multidimensionality. |
A functional distractor attracts some lower-performing examinees. A distractor chosen by almost nobody is non-functional and should be revised.
Good distractors are plausible, mutually exclusive and free of obvious clues.
An Item Characteristic Curve plots probability of a correct response against latent ability. IRT parameters commonly include difficulty (b), discrimination (a) and guessing (c).
DIF: equally able groups show different item-response probabilities. It signals possible bias and demands substantive review.
Uniform procedure, representative norms and transparent interpretation
Same instructions, time limits, materials, scoring and environmental conditions.
Large and representative of the population for whom interpretations will be made.
Purpose, population, administration, scoring, norms, reliability, validity, fairness and limitations.
Observed score (X) = True score (T) + Error (E)
Reliability = True-score variance / Observed-score variance
True score is the expected average over infinitely many equivalent measurements—not a perfectly knowable score. Random error makes observed scores fluctuate.
Observed score को “सच्चाई” न मानें; वह true component और measurement error का मिश्रण है।
SEM = SD × √(1 − rₓₓ)
Higher reliability → smaller SEM → narrower confidence interval around a score.
Approximate 95% interval: observed score ± 1.96 SEM.
A test can have norms yet be poorly standardized if administration varies. Standardization is the whole common procedure, while norms are the reference distribution.
Consistency, sources of error and the correct coefficient for each situation
Stable individual differences attributable to the construct.
Fluctuation from time, items, raters, conditions or scoring.
| Reliability type | Core question | Main error source | Useful statistic |
|---|---|---|---|
| Test–retest | Stable across time? | Time sampling | Correlation / ICC |
| Parallel/alternate forms | Equivalent across forms? | Content + time sampling | Correlation |
| Split-half | Equivalent halves? | Content sampling | Half correlation corrected by Spearman–Brown |
| Internal consistency | Items work together? | Item/content sampling | Cronbach’s α; KR-20 for dichotomous items |
| Inter-rater | Scorers agree? | Scorer differences | Cohen’s κ / weighted κ / ICC |
Spearman–Brown: rₙₑw = nrₒₗd / [1 + (n−1)rₒₗd]
SEM = SD√(1−rₓₓ)
The prophecy formula predicts reliability after changing test length by factor n.
Evidence for the meaning and use of test scores
Validity concerns how strongly evidence and theory support the interpretation of scores for a proposed use. Tests are not simply “valid forever”; a use and population matter.
Validity score के अर्थ और उसके उपयोग पर लागू होती है—केवल test के नाम पर नहीं।
| Evidence/type | Key question | How established | Example |
|---|---|---|---|
| Face validity | Does it appear appropriate? | Examinee/stakeholder impression | A depression test looks relevant |
| Content validity | Does it represent the domain? | Blueprint + expert judgment; CVR/CVI | Syllabus topics represented in correct weight |
| Criterion-related | Does it relate to an outcome? | Correlation with criterion | Selection test and job performance |
| Concurrent | Agreement with present criterion? | Both measured at about the same time | New scale vs current clinical diagnosis |
| Predictive | Forecast future criterion? | Test first, criterion later | Aptitude score predicts later training success |
| Construct validity | Does it behave like the theory predicts? | Convergent, discriminant, factor, group-difference evidence | Anxiety measure relates to similar but not unrelated constructs |
Convergent: strong relations with different methods measuring the same/similar construct.
Discriminant: weak relations with measures of different constructs.
MTMM matrix (Campbell & Fiske) jointly examines trait and method patterns.
Factorial: internal structure matches proposed dimensions.
Incremental: test improves prediction beyond existing information.
Known-groups: groups expected to differ show predicted score differences.
Reference groups, transformations and the normal probability curve
Percentage of norm group scoring at or below a score. Ordinal; gaps are unequal.
Linear transformations with known mean and SD; preserve score spacing and correlations.
Age/grade equivalents compare performance with typical levels, but are not equal-interval.
| Score | Mean | SD / range | Formula or note |
|---|---|---|---|
| z | 0 | 1 | z = (X − M)/SD |
| T | 50 | 10 | T = 50 + 10z |
| Stanine | 5 | Approx. 2; range 1–9 | Broad nine-category normalized scale |
| Sten | 5.5 | Approx. 2; range 1–10 | Standard ten |
| Deviation IQ | 100 | Usually 15 | Position relative to age norm group |
Symmetrical, unimodal; mean = median = mode. Total area = 1. Approx. 68.26% lies within ±1 SD, 95.44% within ±2 SD and 99.73% within ±3 SD.
Positive skew: right tail; usually mean > median > mode.
Negative skew: left tail; usually mean < median < mode.
Leptokurtic peaked, mesokurtic normal-like, platykurtic flatter.
Major theories, landmark tests and interpretation cautions
Historical ratio IQ = Mental age / Chronological age × 100
Modern tests mainly use deviation IQ: performance located relative to same-age norms, commonly M = 100, SD = 15.
| Test/family | Population / nature | High-yield idea |
|---|---|---|
| Stanford–Binet | Broad age range; individual | Descended from Binet–Simon; modern versions assess multiple cognitive factors. |
| WAIS | Adults | Wechsler intelligence battery; index and full-scale scores. |
| WISC | School-age children | Wechsler child battery; current editions use multiple index scores. |
| WPPSI | Preschool/young children | Age-appropriate Wechsler format. |
| Raven’s Progressive Matrices | Nonverbal/group or individual | Abstract pattern completion; strong fluid-reasoning component. |
| Culture Fair Intelligence Test | Reduced verbal/cultural loading | Cattell; aims to reduce—not eliminate—cultural influence. |
Current Wechsler batteries organize subtests into index scores. Common domains include verbal comprehension, perceptual/fluid reasoning, working memory and processing speed; exact indexes vary by edition.
Consider confidence intervals, index discrepancies, language, education, sensory/motor conditions, motivation and cultural opportunity. Large scatter may make a single global score less representative.
Divergent production, scoring dimensions and process models
Generates multiple varied responses to an open problem. It is central to creativity assessment, but creativity also requires usefulness, context and evaluation.
Guilford emphasized divergent production within the Structure of Intellect model.
Narrows alternatives to one best/correct answer. Traditional intelligence and achievement items often emphasize it.
Creativity में “बहुत सारे उत्तर” और intelligence item में “सबसे सही उत्तर” का फर्क याद रखें।
TTCT (Torrance Tests of Creative Thinking): verbal and figural tasks; commonly score fluency, originality, elaboration, flexibility/other creative strengths depending on form.
Alternative Uses Task: generate unusual uses for a common object.
Preparation → Incubation → Illumination → Verification.
तैयारी → समस्या से थोड़ी दूरी → “Aha!” → idea की जाँच।
Three questions that career guidance must never confuse
Question: What is the person’s potential to learn or perform with training?
Examples: Differential Aptitude Tests (DAT), General Aptitude Test Battery (GATB), specific mechanical/clerical aptitudes.
Question: What has the person already learned?
Curriculum, knowledge or skill attained after instruction. Includes standardized educational achievement tests.
Question: What activities or occupations does the person prefer?
Interest predicts choice/engagement better than ability. It is not an intelligence test.
| Interest inventory | Main orientation | Exam clue |
|---|---|---|
| Strong Interest Inventory | Interests compared with patterns of people in occupations; modern reports use Holland themes too. | Occupational interest patterns |
| Kuder inventories | Preference across broad interest areas/occupational scales. | Forced-choice tradition and vocational interests |
| Vocational Preference Inventory | Holland’s RIASEC personality–environment types. | Realistic, Investigative, Artistic, Social, Enterprising, Conventional |
Good fit = abilities + interests + values + personality + opportunities + constraints. No single test should dictate a career. Results are hypotheses for exploration, integrated with interview, experience and local opportunity.
Objective inventories, projective techniques and multimethod interpretation
Standardized prompts and scoring; efficient and psychometrically testable.
Risks: response sets, social desirability, acquiescence, faking.
Ambiguous stimuli invite open responses, interpreted for needs, conflicts and themes.
Risks: lower standardization, scorer effects and variable validity.
| Structured measure | What to remember |
|---|---|
| MMPI / MMPI-2 | Empirical keying; clinical and validity scales; psychopathology/personality assessment. |
| 16PF | Raymond Cattell; 16 primary personality factors derived through factor analysis. |
| NEO inventories | Five-Factor Model: Neuroticism, Extraversion, Openness, Agreeableness, Conscientiousness. |
| EPQ | Eysenck: Psychoticism, Extraversion, Neuroticism plus Lie scale. |
| MBTI | Jung-inspired preference typology; popular in development settings, but categorical typing has important psychometric limitations. |
| Projective technique | Stimulus / response | Association |
|---|---|---|
| Rorschach | 10 inkblot cards; “What might this be?” | Perceptual and personality patterns; requires standardized system and training. |
| TAT | Ambiguous pictures; stories about past, present, future | Murray’s needs and environmental press; motivational themes. |
| CAT | Pictures designed for children | Child-oriented apperception themes. |
| Rotter Incomplete Sentences Blank | Complete sentence stems | Adjustment, attitudes, conflicts and concerns. |
| Word association | First response to stimulus words | Associations and emotional complexes. |
| Draw-a-Person / HTP | Human/house-tree-person drawings | Hypothesis-generating; interpretation must be cautious. |
Batteries and focused tests for cognitive functions
Neuropsychological assessment identifies patterns of strengths and weaknesses in attention, memory, language, visuospatial skill, motor function and executive control. It supports diagnosis, rehabilitation planning and monitoring—but is integrated with history, medical data and functional observation.
| Test/battery | Main domain | High-yield clue |
|---|---|---|
| Halstead–Reitan Battery | Broad brain–behaviour functioning | Fixed-battery tradition; includes Category, Tactual Performance, Speech-Sounds Perception and other tests. |
| Luria–Nebraska Battery | Motor, tactile, language, memory, intellectual processes | Based on Luria’s functional approach; standardized battery of many items/scales. |
| Bender–Gestalt | Visual–motor integration | Copy geometric designs; screening, not a stand-alone localization tool. |
| Wisconsin Card Sorting Test | Executive function | Set shifting, abstraction, feedback use; perseverative errors are important. |
| Stroop Test | Inhibitory control / selective attention | Interference between word reading and ink-colour naming. |
| Trail Making Test | Visual scanning, speed, flexibility | Part B adds alternating set demand. |
| Benton Visual Retention Test | Visual perception and memory | Reproduction/recognition of geometric designs. |
| Wechsler Memory Scale | Multiple memory systems | Memory profile, often paired with broader assessment. |
Fixed: same comprehensive battery for all; comparability but long administration.
Flexible: tests selected for referral question; efficient but depends strongly on examiner expertise.
Premorbid ability, education, culture/language, sensory or motor problems, medication, fatigue, pain, mood and effort. Performance validity measures may be needed.
Likert, semantic differential and Stapel scales with live demonstrations
Respondents indicate degree of agreement with several favourable and unfavourable statements. Item scores are summed; negative items are reverse-scored.
Statement: “Psychological tests should be explained to examinees in plain language.”
Choose a response to see the coded score.
A concept is rated between bipolar adjective pairs, commonly on a 7-point continuum. Major meaning dimensions include Evaluation, Potency and Activity.
A single adjective is rated on a unipolar scale, typically from +5 to −5, without a zero category. It is useful when a natural opposite adjective is difficult to find.
Accurate
| Scale | Stimulus format | Typical scoring idea | Memory hook |
|---|---|---|---|
| Likert | Agreement with statements | Summated item ratings; reverse-score negative items | “Level of agreement” |
| Semantic differential | Bipolar adjective pairs | Usually 7 positions | Two opposite poles |
| Stapel | One adjective | +5 to −5, usually no zero | Single adjective |
| Thurstone | Pre-scaled statements | Equal-appearing intervals; judges assign scale values | Judges first |
| Guttman | Cumulative ordered items | Endorsing a stronger item implies weaker ones | Scalogram/cumulative |
CBT, adaptive testing, security, accessibility and equivalence
CAT uses an item bank—often calibrated with IRT—to choose items near the examinee’s current ability estimate. It is not simply a computer-delivered fixed test. It requires a large secure calibrated bank, content-balancing rules and exposure control.
Clinical, organizational, educational, counseling, military and career settings
Diagnostic clarification, severity, risk, treatment planning and outcome monitoring.
Combine: interview + tests + observation + records.
Selection, placement, promotion, training needs, leadership and team development.
Demand: job analysis, predictive validity and adverse-impact monitoring.
Achievement, readiness, learning needs, giftedness, disability support and programme evaluation.
Demand: age/grade norms and appropriate accommodations.
Self-understanding, emotional concerns, strengths, decision-making and progress.
Demand: collaborative feedback, not labels.
Selection, classification, specialist placement, readiness and leadership potential.
Demand: high reliability, security, standardization and fairness.
Integrates aptitude, interests, values, personality, achievement and opportunity.
Demand: exploration of options, not a single deterministic verdict.
Selection: Who should enter?
Placement/classification: Where will the admitted person fit best?
Diagnosis: What pattern/problem is present?
Evaluation: Did an intervention work?
The higher the stakes, the stronger the required evidence. Use multiple sources, report uncertainty and check consequences for different groups.
High-stakes decision में एक score को अकेले “final truth” न बनाएं।
Protect the person behind the score—then test your revision
Clarify referral question, competence, informed consent, purpose, access, accommodations and appropriate instrument.
Standard administration, dignity, accessibility, rapport without coaching, security and observation of relevant behaviour.
Accurate scoring, contextual interpretation, confidential records, understandable feedback and limits of inference.
| Ethical principle | What it requires | Common violation |
|---|---|---|
| Competence | Use tests only with adequate training and within scope. | Untrained interpretation of complex instruments. |
| Informed consent | Explain purpose, procedure, foreseeable use and limits. | Hidden high-stakes use. |
| Confidentiality | Restrict access and disclose only with authority/lawful basis. | Sharing identifiable scores casually. |
| Test security | Protect items, scoring keys and copyrighted materials. | Publishing live secure items. |
| Fairness | Appropriate norms, language, accessibility and bias review. | Using irrelevant barriers or outdated norms. |
| Feedback | Explain results accurately, respectfully and with uncertainty. | Reducing a person to a label or number. |
KR-20 estimates internal consistency for dichotomous items.
The criterion occurs in the future, so the evidence is predictive.
p is the proportion correct. A larger p means an easier item.
Semantic differential uses pairs such as good–bad or strong–weak.
WCST assesses set shifting and cognitive flexibility; perseveration is a key index.
RIASEC is Holland’s vocational interest/person–environment typology.
SEM = SD√(1−r). As r rises, the error term shrinks.
CAT updates ability after responses and chooses the next informative item.
Research, content development and hosting all carry real costs. If this work has supported your preparation, a voluntary contribution helps it keep growing.
Support ClassRoom Thinker →Every contribution is voluntary and directly supports the platform.Found an error, noticed something unclear, or have a useful idea? Educators, writers, designers and developers are always welcome to reach out.
Corrections, suggestions and genuine collaboration are always welcome.