What to do during m1 to ensure 250+

[sloop]

I've seen 3 studies that showed no statistically significant relationship between preclinical grades and step 1 score. Obviously it's logical to think grades correlate, but the data just isn't there.

Maybe the correlation is there at your school, but the variation in curricula, grading, etc makes it difficult to apply nationally.

http://www.ncbi.nlm.nih.gov/pubmed/26363626
http://www.ncbi.nlm.nih.gov/pubmed/20135567

The first study indicates that different courses have different weights as far as how well they predict and that some courses are not statistically significantly correlated with step 1 score. Still, major courses like pathology, epidemiology and pharm are significantly correlated (p<0.001). Anatomy, biochem, psychiatry, and some others were not significantly correlated. In any case, many courses correlate with step 1 performance. The second article indicates that cumulative preclinical grades predict step 1 failure.

In any case, even if studies were not finding a correlation, I'd think it's more likely that the study is wrong because it makes no sense. The preclinical courses are the subjects that step 1 tests. Findings indicating no correlation have a ton of prima facie evidence to overcome. Those studies are more likely to have something flawed in their methodology or be chance findings than they are to actually reflect that two things that are so obviously related turn out to have absolutely no bearing on one another.

The problem is that these kind of studies can't be carried out across institutions. Grading schemes are different and different schools don't even have all the same courses. The only standardized measures we have are the preclinical nbmes which not all schools use. Besides, studies have been done looking at the correlation between individual nbme exams and step 1 and they usually find a correlation (does this surprise anyone? I mean, they're both standardized science exams.)

In sum, I think people have reason to be very skeptical of anyone who claims that preclinical grades do not correlate with step 1 performance.

---

Members don't see this ad.

 

[Dave89]

I didn't use any qbanks during M1. In fact, while I supplemented with FA for certain subjects (e.g., pathology and immunology), I studied toward the in-house exams (with judicious use of BRS books when appropriate).

But there is no way to "ensure" a 250+.

 

[seminoma]

http://www.ncbi.nlm.nih.gov/pubmed/26363626
http://www.ncbi.nlm.nih.gov/pubmed/20135567

The first study indicates that different courses have different weights as far as how well they predict and that some courses are not statistically significantly correlated with step 1 score. Still, major courses like pathology, epidemiology and pharm are significantly correlated (p<0.001). Anatomy, biochem, psychiatry, and some others were not significantly correlated. In any case, many courses correlate with step 1 performance. The second article indicates that cumulative preclinical grades predict step 1 failure.

In any case, even if studies were not finding a correlation, I'd think it's more likely that the study is wrong because it makes no sense. The preclinical courses are the subjects that step 1 tests. Findings indicating no correlation have a ton of prima facie evidence to overcome. Those studies are more likely to have something flawed in their methodology or be chance findings than they are to actually reflect that two things that are so obviously related turn out to have absolutely no bearing on one another.

The problem is that these kind of studies can't be carried out across institutions. Grading schemes are different and different schools don't even have all the same courses. The only standardized measures we have are the preclinical nbmes which not all schools use. Besides, studies have been done looking at the correlation between individual nbme exams and step 1 and they usually find a correlation (does this surprise anyone? I mean, they're both standardized science exams.)

In sum, I think people have reason to be very skeptical of anyone who claims that preclinical grades do not correlate with step 1 performance.

I only have access to the abstract at the first link, but it doesn't show anything other than low Path grade correlates to low step 1, which is only a single course grade and only correlating to a low score. As an analogy, the fact that smoking worsens your health below baseline doesn't mean not-smoking improves your health above baseline. In fact, as we all know, quitting smoking reduces the decline in PFTs and former smokers can eventually return to baseline-rates of decline, but they certainly will not have the lungs of an age-matched never-smoker.

Likewise for the second link, it only shows low grades with risk of failing.

Neither of these is really correlating scores or score ranges (beyond failing or <205) with grades. They both simply suggest that low preclinical grades results in low scores. No data out there to suggest average preclinical grades correlates with average Step 1 or high preclinical grades correlates with high Step 1.. and those last 2 are the ones people really care about. Obviously if you are near-failing preclinicals you are going to have knowledge gaps come step 1 time. Less obvious, and more interesting, is whether or not +1SD in school means you're on your way to a 250+. Anecdotally I know people with a higher class rank and higher overall exam scores than me who scored >20 points lower on Step 1. I also know two people who are a quartile below me in class rank who scored within 5 points on Step 1. Tiny n, but still.

If there really was a strong correlation between preclinical grades and step scores I don't see a reason for schools to push P/F. Instead they should have grades, show their students the data, and push their students to score as high as possible in classes because that will correlate to scoring high on step 1.