Rajasthan SET cheat sheet

Teaching aptitude levels follow Bloom's revised taxonomy: Remember, Understand, Apply, Analyse, Evaluate, Create; verbs like 'list' map to Remember, 'design' to Create.

Norm-referenced tests compare a candidate to a group (percentile rank, z-score, stanine); criterion-referenced tests compare performance to a fixed standard or learning objective.

Reliability = consistency of scores; split-half reliability is corrected by the Spearman-Brown formula r_tt = 2r/(1+r). Validity = whether the test measures what it claims.

A z-score of +1.0 equals one standard deviation above the mean; T-score = 50 + 10z; stanine = 5 + 2z (rounded), with stanines 1-9.

Item difficulty index P = proportion answering correctly; item discrimination D = P_upper − P_lower (upper/lower 27% groups); D above 0.30 is generally acceptable.

Levels of teaching: memory level (recall), understanding level (insight and application), reflective level (problem-solving and critical thinking) - reflective is the highest.

Research designs: experimental uses random assignment and manipulation; quasi-experimental lacks random assignment; ex post facto infers causes from pre-existing differences.

Sampling: simple random, stratified (heterogeneous strata), cluster (intact groups), systematic (every kth unit); increasing sample size narrows standard error by √n.

Types of variables: independent (manipulated), dependent (measured), intervening (mediating), moderator (affects strength of relationship); control variables are held constant.

For a subject-specific paper, verify every factual claim against the syllabus unit it belongs to; an answer must be traceable to a named concept, author, date, or formula.

In reasoning items, distinguish deductive validity (conclusion follows necessarily from premises) from inductive strength (conclusion is probable, not certain).

Syllogism rule: a valid categorical syllogism has exactly three terms, the middle term distributed at least once, and no term distributed in the conclusion that was undistributed in the premises.

For numerical items, check units and significant figures before selecting an option; a correct calculation with wrong units is a wrong answer.

When two options seem correct, prefer the one that is more specific and directly answers the stem; overly broad or absolute statements (always, never) are usually distractors.

For date- and author-based questions, anchor to the primary source: textbook edition, original publication year, and attributed author rather than secondary summaries.

In data-interpretation items, compute percentage change as (new − old)/old × 100 and distinguish percentage points from percent change.

For diagram- or passage-based items without a figure, reconstruct the relationship in words first, then match to the option that preserves the stated causal or logical order.