DSAT SCORING AND MODULE ASSIGNMENT

A student’s DSAT score cannot be calculated by looking at the number of correct answers and the total number of questions as it is an adaptive test where the difficulty of the questions answered correctly has a direct impact on the score because it determines which of the second adaptive modules the student receives and thus impacts the potential overall score a student can attain.  The number of medium level difficulty questions is the specific driving factor in the determination of the second module. 

All students take the same module 1 (M1), or baseline, which is composed of a majority of medium-level questions.  The DSAT adapts based on how the student performs in the first part of each section (Reading/Writing or Math).  M1 in each section consists of a mix of easy, medium, and hard questions. Based on how a student performs in that first module, they are routed to either a harder second module or an easier second module. The student's final scaled score depends not only on how many questions the student gets right overall but also the difficulty level of the correct answers and the second Module (2A or 2B) they received.  

The more difficult questions a student answers correctly, the higher they can score. Even though each question is scored simply as correct or incorrect, not all correct answers can contribute equally in terms of how high your scaled score can go. If a student gets only the easy questions right in the baseline module, they are directed to the easier adaptive module. Even if all the questions are answered correctly in the easier module, the score potential is limited because harder questions carry more weight to push your score higher.

 

There is a probability that two students might answer the same number of questions correctly and get the same raw score but still end up with different scaled section scores. This happens because the Digital SAT's second module is adaptive.

 

For example, if student A performs well in M1, they get a harder M2. Correct answers on harder questions give access to higher scoring potential. But if student B has the same raw score but is directed to the easier M2, the overall potential score gets limited. The second module is given more weight because it reflects that the student performed well enough in M1 to be routed to a harder set of questions, which carries higher scoring potential. 

 

There are ceilings and floors within the adaptive scaled score charts that will not allow a student to get a perfect score if they perform poorly on M1 and then well on the easier M2 so that students cannot purposefully tank the baseline section to receive the easier adaptive module.  

 

When looking at this specific student, they received  M2A in the fall for math. (This can be found in the new report suite data table or on the Assessment Score and Review Report in the old suite).  They did not do well when given primarily medium and hard questions, only answering 3 of 22 correctly.  The "score floor" catches that student and the score is scaled due to the difficulty of each question. Conversely, if you look at the winter scores, they received M2B (easier module) because their score on the baseline/M1 was 11/22, which is at or below 50% correct- but specifically the number of correct answers to medium-level questions drives the determination of the module.  You cannot just look at 50% correct as a hard cutoff for module determination.  The difficulty of the question will always impact module determination.  When receiving primarily easy and medium questions, this student was able to answer more questions correctly, however, the majority of the questions they answered correctly (28/44) were easy or medium level in difficulty, thus she could not score as high.  

 

 

DSAT uses Item Response Theory (IRT) scoring model which is a function of three main factors: (1) whether a student answers questions correctly or incorrectly, (2) attributes of the questions, such as difficulty level, that a student answers correctly or incorrectly, and (3) whether a student’s response pattern across questions suggests guessing.  Therefore, two students may obtain the same number of correct responses on two different test forms but still have different reported scale scores.