Item Types That Ask You to Commit to a Set
Four miniature interface skeletons: a bowtie template with five empty targets over three option columns, a trend table of three timestamp columns by four finding rows, a matrix grid with single-select circles, and the identical matrix grid with checkboxes, each carrying a corner chip reading `0/1` or `+/−`.
Strip 1 · Bowtie Mechanic: one central condition, two actions to take on the left, two parameters to monitor on the right, all drawn from three separate option columns. Hook: five placements, five points, zero-one scored. Every cell gets filled, including the one you are guessing at, because an empty cell scores exactly the same as a wrong one and you have given away the chance.
Strip 2 · Trend Mechanic: the same client's data at three or more timestamps, asking what changed and what the change means. Hook: the answer is a direction, never a single value. The distractors are the values that stayed flat, offered to you because they are abnormal and therefore tempting.
Strip 3 · Matrix grid, one per row Mechanic: findings down the rows, categories across the columns, exactly one selection permitted in each row. Hook: zero-one scored per row, so an unanswered row is a donated point. Work top to bottom and answer every row, even the two you dislike.
Strip 4 · Matrix grid, many per row Mechanic: the same grid, the same visual weight, one or more selections permitted per row. Hook: plus-minus scored. The discipline reverses completely inside a layout that looks identical, which is why the instruction line above the grid is read before the grid itself.
Strip 5 · Extended multiple response Mechanic: select all that apply, drawn from a longer option list than a traditional item, with the required number stated in the stem. Hook: that stated number is a constraint the item scores you against, not a hint. Count your selections before you move on.
Strip 6 · The container Mechanic: bowtie and trend appear both inside case sets and as stand-alone items; matrix grids sit inside case sets. Hook: the container changes the scoring, not the reasoning. The same cue-to-hypothesis work happens either way.
Margin notesame grid, opposite rules.
☐ I can name the bowtie rule unprompted. ☐ I can tell the two matrix grids apart. ☐ I can state what the required number does on an extended multiple response.
Scoring arithmetic → BOOK 02 · P6. Reading a case at orientation depth → BOOK 01 · P60.
TakeawayIdentical-looking grids can carry opposite scoring rules, so the rule is read before the options.