The following numbers all come from one ESL class in California of fewer than 30 students. The students range in age from teens to 50s. Some are full-time students, some are part-time and some work.
0.63 – Correlation between homework scores and final grades
Homework is a tiny percentage of the final grade, but predicts it to a fair degree. I usually gave homework a simple at-a-glance score between 1 and 5 points per assignment. The highest score for homework for this semester was 108, and the lowest score was 12. Homework is also correlated with attendance at 0.87. Attendance, on the other hand, is correlated with final grades at 0.72. This tells me I can cut the amount of homework (saving myself some grading time in the process), take attendance rigorously, and expect roughly the same distribution of grades. Next semester’s students will be happy to hear being physically present is such a strong predictor of English skill.
0.82 – Take-home essays/final grades
Take-home essays are also, surprisingly, a tiny part of the final grade. That is what makes this correlation so surprising. Take-home essays predict final grades even more than they do in-class essays (0.76), although of course take-home essays are themselves not a part of in-class essay grades, while they are of final grades. Blame my mild innumeracy for not knowing how much that should affect these numbers. In any case, it seems prudent to replace some of the other homework (see above) with essay-related work like planning, more rough drafts, and reflections.
-0.03 and -0.85 – Numbers of tardies and absences/final grades
This particular class met in the early morning, which accounts for the high number of tardies (8.0 on average for the class out of about 30 class meetings, with a standard deviation of 5.8), but they didn’t seem to do much. Students walking in late didn’t hurt their grades so much as annoy me personally. This probably has something to do with my practice of starting each class with some kind of task rather than a quiz, which is arguably a bit of a conceit of mine. Absences, on the other hand, were even more predictive of final grades than essays, and much more predictive than homework scores. Again, being a warm body in the classroom seems to be a reasonable heuristic for a lot of heady work.
-0.55 – Years in the US/final grades
This number comes from a survey we did at the beginning of our classes. I suppose this will surprise a lot of people who work in EFL – I certainly expected a more or less linear relationship between years in the country and degree of acculturation (similar to integrativeness, the motherlode of language learning motivation) before I got here (although I had reason to know better). In fact, particularly with people who were partly educated here, the sense that one doesn’t belong in ESL is a significant barrier to buy-in for class activities and willingness to communicative with classmates. ESL teachers often have a mix of eager international students with standard-issue grammar, communication and acculturation problems and jaded veterans of US society who have established an identity around their patterns of language use and definitely don’t see themselves as ESL students. My main takeaway from this is to acknowledge the different needs and motivations of my main two student constituencies near the start of the semester to defuse any feeling among the veterans that they don’t belong there or don’t need to work hard.
0.45 – Hours of sleep before test 1/final grade
I usually put one “gag” item at the top of my tests – “Name: _____ Student ID: _____ Breakfast this morning: _____ ” to give another example. Sometimes, these yield insights into my students’ lives (a lot have nothing but coffee for breakfast), and occasionally lead to educationally useful data. The only time I had a quantifiable “gag” item this year was on the first in-class writing test, on which students were asked to write how many hours they slept the night before the test. This number turned out not only to predict the scores on that test (albeit weakly, with 0.22) but their grades for the semester. Now, this isn’t a slam dunk, but it is more predictive than most individual homework assignments (whose correlations with the final grades ranged from -0.14 for the first homework of the semester to around 0.60 for essay-related stuff). Apparently the recipe for a high-scoring student is 8 hours of sleep a night before coming to class, not necessarily on time and not necessarily with any homework. As I said, at least this simplifies my grading.