Skip to the page

Same number right, different SAT score

Two students get 30 of 44 right. Route them through a toy adaptive test and see why their scores differ.

Both got 30 right, yet the scores differ by 100: A got those answers right on harder questions.

Student A

Routed to the harder module 2easierharderModule 1 above, module 2 below

650toy score, 30 of 44 right

Student B

Routed to the easier module 2easierharderModule 1 above, module 2 below

550toy score, 30 of 44 right

Toy rule: 14 or more right in module 1 opens the harder module 2. Made-up difficulties, one simple model; not College Board's questions or scoring.

Keep exploring

How this works

Method

  1. A toy two-stage test: 22 mixed questions, then 22 easier or harder ones depending on module 1. All difficulties are our own.
  2. The estimate is the ability that expects the student's number right on the questions they actually saw (a Rasch model), shown on a made-up 200 to 800 scale.
  3. College Board says real section scores depend on the characteristics of the questions answered right, including difficulty. Its own model and questions are not public; this toy shows only the idea.