Watson–Glaser-Style Critical Thinking Practice

Five response rules in one original edition

Each part has six original questions. Inference uses True / Probably true / Insufficient data / Probably false / False under explicit facts or sampling models. The four binary parts ask separately about assumptions, necessary conclusions, supported interpretations and material arguments. Each prompt repeats its own rule.

This is a fixed learning bank, not a reconstruction of an official form. Inference probability conventions are explicitly bounded for checkable original keys; they do not claim to reproduce historical or current publisher scoring. All questions stay English in both interfaces; instructions, explanations and full guides are bilingual.

One continuous whole-attempt clock

Timed mode starts once at 30:00. Moving between parts, reading instructions, returning or using a background tab does not pause or restart it. You may explicitly leave blanks and revise any question until the whole attempt ends. At expiry it locks with unanswered items retained. Untimed study has no deadline.

A device-clock inconsistency clears the attempt instead of creating a timed result. This is a local practice clock, not proctoring. Thirty-five minutes including instructions is an editorial planning estimate, not observed completion data. Your actual invitation controls its timing, navigation, equipment and permitted help.

Raw counts without a performance label

A correct offered selection adds one correct count; another offered selection adds one wrong count. Missing and skipped responses add one unanswered count. The three counts sum to six per part and thirty overall. No time bonus, guessing adjustment, norm, percentile, performance band, weighted composite or employer cutoff is supplied.

Binary and five-option items have different response sets; difficulty has not been equated. A count cannot diagnose guessing, general critical-thinking ability or a weakest part. After finishing, every question, key and explanation is free, even with all blanks. Five complete public guides are available before answering and identical for everyone.

Review the first unsupported reasoning move

For each item, name the first missing premise, reversed implication, population change or irrelevant reason in your working. Make one fresh example before returning to the fixed bank. Recognition and changed conditions can improve repeat counts as well as learning; this exercise does not separate those causes or guarantee transfer.

Use the public provider material as background and your invitation for the exact assessment you will take. This practice does not predict selection, qualification or legal/professional competence. Shared URLs lead only to complete public guides, without answers, counts or attempt identifiers; refreshing or leaving clears the private on-page record.

Is this an official Watson–Glaser form?

No. All 30 questions and five full guides are independently written. The name identifies a preparation topic, with no publisher item bank, endorsement, score conversion or claim of form equivalence.

What are the five answer rules?

Inference has five certainty/likelihood choices under the prompt’s explicit model. Assumptions, deduction, interpretation and arguments each use their own two-choice rule. Every question repeats its instruction so you can check which decision it asks for.

Does switching parts pause or reset the clock?

No. Timed mode uses one site-selected thirty-minute clock, including reading and navigation. You may revise any item before the whole attempt ends. Untimed study uses the same bank without deadlines. Actual tests may have different rules.

Do blanks count as wrong or show a weak ability?

Blanks remain unanswered, separate from selected wrong answers. Parts are not equated, and their option counts differ. Raw counts do not establish a weakest ability, guessing, a percentile, a pass band or an employer prediction.

Are all explanations and complete guides free?

Yes. All 30 explanations appear when the whole attempt closes, including with zero correct. All five full public guides can be read before answering, without registration, a payment or a score threshold.

Can shared links reveal my responses?

Shared links contain only public guide addresses, never responses, counts or private attempt identifiers. The private record stays in page memory and clears on restart, refresh, leaving or language change. Optional anonymous activity counts have separate consent.