College Board just released the results from the first administration of the New SAT, administered March 2016, and with it they also released a set of concordance tables to let us know what exactly a good score is. We’ll take a look at this and analyze what exactly these new scores mean for students!
How do New SAT Scores Compare to Old 2400 SAT Scores?
Since the new SAT is scored out of 1600, while the old SAT was scored out of 2400, the new scores are obviously lower than the old. However, if we adjust the old scores using the “sliding scale” method (i.e. subtract 800 across the board from the old SAT scores, normalizing as if it were on the 1600 scale), we can see (Figure 1 above) that overall there is a much larger difference for lower-scoring students, with the New SAT having higher scores than the old pre-2016 SAT. This suggests that the same performance would net a higher numerical score. This means that the “sliding scale” method would not actually work for comparing new and old SAT scores.
Here (Figure 2), we expand the right side of our graph to see how these scores affect high-performing students. The x-axis denotes Old SAT scores, and the y-axis how much higher New SAT scores are, with respect to the “sliding scale” scores. For example, a student who scored a 2230 on the Old SAT would expect a “sliding scale” score of 1430. However, if this student took the New SAT and performed at the exact same level, he or she would expect to receive a score of 1530, roughly 100 points higher.
However, if we look at even older SAT scores, from before 2005, when the exam was also scored on a 1600 scale, we see a much smaller difference. Especially on the high end, where academically excellent students would expect to score, we see that there is a very small difference between the New and the Old (pre-2005) SATs. What this means is in terms of scoring, the New 2016 SAT is more similar to the pre-2005 SAT than the SAT we have become accustomed to.
Finally, in taking a look at the comparison between the New SAT Math Section and the Old (2005-2015) SAT Math Section, we see that there is also not a very large difference, especially on the high end. Note, because the Writing and Reading Tests did not have a 1-to-1 comparison, due to the differences in scoring, we did not chart the differences between the two.
Why is there such a large difference in scores?
When comparing the New SAT composite score to the Old (2005-2015) SAT’s composite scores, you might be wondering why there is such a large difference. No one can say for sure, but we believe a large contributor to this difference is not some desire by the College Board to inflate scores, as some have suggested. Rather, it is likely due to the fact that the Reading and Writing sections are now combined into a single subsection score. If you look at the comparison between the math sections, there is very little difference, especially on the high end. This suggests that the difference stems from something else. Since College Board had the unenviable task of finding a way to condense two sections, Reading and Writing, into a single score, the large difference between the New and Old tests likely stems from that.
Contrary to many speculators, we do not believe College Board purposely inflated scores to make test-takers feel better about themselves. With the publication of the concordance tables, College Board has made the differences incredibly transparent, so the differences do not appear to us to be anything other than a byproduct of the format change.
What does this mean for students?
Though we are certainly surprised at how high students must score to be in the “elite” range (i.e. around a 1560/1600 for HYPMS [Harvard, Yale, Princeton, MIT, Stanford]), and though this is much higher than our previous predictions, this is not necessarily a bad thing for students.
- Scores are compressed on the high end, but this is not unlike the ACT’s scoring system and is arguably more lenient in terms of spread: Consider that the average ACT score for Harvard, Yale, Princeton, MIT, and Stanford regularly float around a 34. There are only two more “units” to reach a perfect score, a 35 and 36. Now consider that the concordance table suggests that an equivalent score is either a 1540 (ACT to SAT concordance) or a 1560 (Old SAT to new SAT concordance). In either case, there are far more stepwise “units” separating the HYPMS average from a perfect score. Though we are used to seeing a much larger spread, the new spread is still more lenient than the ACT’s, so there is little cause for alarm.
- The compression negates minor differences in scores that may previously have led to anxiety: As evidenced by questions to MIT admissions officers, many students worried that 20-30 points would mean the difference between being accepted to HYPMS or being rejected. However, such differences were pretty negligible – a 2300 was fairly indistinguishable from a 2270. Many admissions officers we’ve talked to have said that on the old scale, there was as much as a +/- 50 point spread where scores are viewed with equal weight. Higher scores matter, of course, but after a certain point the differences become negligible. Though it might seem psychologically more frightening to have “less room to mess up” on the new exam, since there are only 5 stepwise “units” separating an HYPMS average score from a perfect score, when compared to the 10-15 stepwise units, the actual difference may be exactly the same. Compressing the scores will not homogenize the scores. Rather, they will likely reduce negligible, erratic spreads due to a single question difference.
- Ten points on the new exam is not the same as ten points on the old exam: We cannot stress this enough. Though both the Old and the New SAT have increments of 10 points, they cannot be compared to each other. On the old exam, missing a single problem might drop you 10-20 points, but on the New 2016 SAT, you may be able to miss 2-3 problems before seeing any score deduction. Because the new scores themselves are grouped more closely together, you will likely have more wiggle-room for error, which further emphasizes point #2.
- Conversely, high scores will not be any more prevalent: The converse of the previous points is that on the new exam, there will likely not be any more “good scores” than there were on the old SAT. Some have feared that the new scoring scale homogenizes test takers and “flattens out” the curve, reducing the ability to differentiate between the quality of students. However, this is likely not the case. Though more students will score in a narrower range, this does not mean high scores will be devalued or more prevalent. Rather, the new scoring system captures a broader range of equally qualified students. Ten points on the new exam may account for 30 on the old, which means more students will achieve the exact same score, but functionally, the same number of students will score well, in terms of what admissions officers consider. In other words, the percentile ranks have collapsed down so as not to give a false sense of superiority or inferiority to students. Consider the Old SAT’s Percentile breakdown. A 2220-2300 all had the percentile rank of 99, which means they should all be considered equally, but in actuality, students would feel much less comfortable with a 2230 than a 2380. These psychological differences are negated with this new scoring system, so the new scores should reflect true differences more accurately.