Measurement Levels, Interval

Interval Scaling

The scholastic aptitude test (SAT) scores are an example of interval-level data. The SAT test attempts to measure students’ academic preparedness for college. In addition to equal unit size, interval-level scaling assumes equity and nonequity, meaning that two people who have similar levels of preparation for colleges should also have similar scores on the exam. The rule of greater-than-or-less-than is also assumed. As a result, the SAT score functions to rank order students such that more prepared students would have a higher score than less prepared students.

The value of an interval scale becomes clear when seen in use. College admissions officers use all the principles identified above in making decisions using the SAT exam scores. For example, the college admissions officers uses the assumed equity and greater-than-or-less-than principles when asserting that students with higher scores are more prepared, but he or she can make other claims using the assumption of equal unit size to compare student SAT scores. If, for example, a college admissions officer were to compare student A’s score of 1,400 to student B’s score of 1,600, he or she could say that student B is more prepared for college than student A by 200 units. If he or she then compared the same student B to student C who scored 1,800, again it could be said that student C is 200 units more prepared for college than [Page 944]student B. The admissions officer could conclude that students A and B were equally underprepared for college compared to a set reference point, but it would be important to note that the reference point is different for each student. It is important to note that with interval scaling the admissions office could not claim that student C was twice as prepared as student A. Without an absolute zero, the data cannot be directly used for ratios. However, because the distance between 1,400, 1,600, and 1,800 is assumed to equally distant from the other scores, the admissions officer could make additional claims about how prepared each student is in comparison to the others using statistical analysis. It is common to calculate means and standard deviations, among other statistics, using interval-level data.

...