1.3 CBT Strategy, AITs, Calculators, and Pacing

Why strategy is part of competency on a once-yearly CBT

The April 14, 2027 sitting is not a year-round PE you can retake in 60 days. A pacing error that burns 25 minutes on one valve-sizing alternative item type (AIT) is not just “a bad item.” It is a bid to wait until the next annual administration. The technical content still matters; the CBT rules decide whether you convert that content into a scaled score.

Most questions are traditional multiple-choice. NCEES also uses AITs. The NCEES computer-based testing page lists these common types:

NCEES’s rule is blunt: all questions, including AITs, are scored as either correct or incorrect. There is no partial credit. Combined with the scoring process—no deductions for wrong answers—the operational rule is: answer everything. A blank fill-in-the-blank is a guaranteed zero. A guessed multiple-correct set can still be right. A guessed traditional item cannot hurt a neighboring item.

Calculators: the 2026 list, not “any scientific”

NCEES reviews the calculator list annually. For 2026 exams, the published models are the only ones allowed into the room:

• Casio: all fx-115 and fx-991 models (the model name must include fx-115 or fx-991)
• Hewlett Packard: HP 33s and HP 35s only
• Texas Instruments: all TI-30X and TI-36X models (the model name must include TI-30X or TI-36X)

A TI-84, TI-Nspire, HP 12C, or HP 48 is a locker item, not a workstation item. The Examinee Guide allows one approved calculator in the room without its cover; spare approved calculators and covers go with personal belongings. A TI-30XS is also available on screen. If your handheld is rejected at check-in, that on-screen TI-30XS is the remaining legal machine—not a graphing calculator in your backpack.

Confirm the list again as you approach April 2027, because NCEES restates that the list is reviewed each year. Do not discover a policy change at 7:40 a.m. in the Pearson lobby.

Pacing 85 items in 8.5 hours

Exam time is 510 minutes for 85 items:

Worked example. You finish item 40 at 5.0 hours (300 minutes) into the exam clock. You should have been near 40 × 6 = 240 minutes on an even split, or near 40 × 5.5 = 220 minutes on a first-pass plan. You are 60–80 minutes behind. Remaining time is 210 minutes for 45 items. That is 4.7 minutes each if you never review. The recovery is not “try harder on the next SIS calculation.” The recovery is: guess, flag, move on any item that has already eaten eight minutes without a path, then use the leftover block only on flagged items that still have a handbook lookup.

The 50-minute scheduled break is for food, restroom, and resetting your eyes. It is not a secret sixth hour of exam time. If you skip it, you do not gain 50 minutes of questions.

Flag/guess rules that match all-or-nothing scoring

1. Never submit a blank. A wrong multiple-choice or a wrong AIT is the same 0 as an empty fill-in-the-blank, and empty cannot accidentally be right.
2. On multiple-correct, satisfy the instruction exactly. If the stem says select two, selecting three because “they all look true” is an incorrect item—not 75% credit.
3. On fill-in-the-blank, obey the requested format. If the stem wants a numeric Cv to two decimal places, 12 is not the same response as 12.00 when the key is format-strict. Read the unit line; do not type gpm into a blank that already includes units.
4. On point-and-click, zoom the graphic. A click on the adjacent balloon is a zero. The Pearson markers are for scratch work, not for changing the keyed hotspot.
5. When ISA-5.1 or 61511-1 is in play, search first. Deriving a 2009 balloon letter or a 61508 definition from memory is how competent engineers miss credited years. The left-pane search is slower than recall and faster than a year-long retake.

Exam-day scenario: the eight-minute valve AIT

Marcus hits a drag-and-drop on a control-valve failure mode: four tiles (air-to-open, air-to-close, fail-last, solenoid de-energize) and a sketched air-failure path. He is 70% sure of three tiles and stuck on the fourth. At minute six he is still rearranging. Expected value on an all-or-nothing AIT is ugly: extra four minutes might save one item and will steal time from later Measurement or SIS items that he can actually search. He should place the best arrangement now, flag it, and leave. After the first pass, if 20 minutes remain, he returns only if a handbook actuator table can settle the fourth tile.

Same sitting, a fill-in-the-blank orifice-plate beta ratio. Marcus can compute 0.50 but the stem asked for d/D to two decimals. He types 0.5 and moves on. If the key is 0.50, the AIT is incorrect with no partial credit. He should have mirrored the requested precision.

Bring the handheld you actually practice on (an fx-115ES Plus or TI-36X Pro is a typical legal pair), verify it is still on the current-year list, and know the on-screen TI-30XS well enough to finish a Reynolds number if the handheld is taken. Then spend the 8.5 hours like an engineer with one annual attempt: search the supplied files, timebox at ~6 minutes, and refuse to leave zeros on the table.