The maximum course load for undergraduates for a semester is 17 credit hours and for graduate students is 16 hours. The maximum load will increase to 18 credit hours for undergraduates at the beginning of open enrollment for each term.
According to international teacher and writer Melissa Morgenstern , junior year is the most common time for college students to study abroad. Around this time, you might also be thinking about career and life plans that would make it better for you to finish your degree in less than four years. For example, you might meet with a future employer who really wants to hire you as soon as possible, or who tells you the job market in your field is better at this moment than it will be a year in the future.
You might have a romantic relationship that would work better if you could finish early and move to a new hometown with or marry your partner.
But, generally speaking, taking a maximum or an overloaded number of courses may be harder or even impossible in junior year. Be sure you are taking at least the minimum number required, and more than that if necessary to stay on course for graduation.
Remember, while everything you are doing junior year is important, performance in your classes is the key to all of it. Senior year is often thought of as the victory lap of college. The time to blow off steam, take a few easy classes as possible and relax with your work practically over. Just remember that you or someone you love are paying good money for every moment you spend at school, so you should work hard to get as much out of them as possible.
There is a balance to be struck senior year. The sooner you can get to the job market the better in most industries. Just remember that most upper-level courses are designed to get you right to the professional level of expertise, so they will be hard and require a lot of hours of study. If you do think you need to be there for your full senior year, consider taking the minimum number of credits needed to stay enrolled full time.
Enjoying a class on an interesting subject from an expert teacher and just doing so for the joy of learning is a great experience. He has 10 years of collegiate communications experience and has worked with hundreds of college students. Learn More. Jacob Imm Oct 01, Upper Division Courses - - Courses usually with prerequisites; primarily for third and fourth year undergraduates. Contact the Office of the Registrar. Send a message AD 83 E. Boone Avenue Spokane, WA Do Not Edit:.
Don't Edit This Field:. Thank you! Comment or Question. Excluding students who dropped out after one semester, students take a course load other than their modal course load i. In the context of the university, there is qualitative evidence suggesting that a fair amount of below-full-load taking behavior occurs because students try to take overenrolled classes, do not manage to register from the waiting list, and do not replace the class with another one Moore and Tan To the extent that this is the driver of within-student variation, concerns about selection bias in estimators using within variation are minimized.
A clear source of endogeneity in estimating the effect of course load on student performance is ability bias. Higher student ability should both lower the effort cost of taking more classes as well as increase the expected grade performance, leading to a positive non-causal observed relationship between classes taken and grades.
In this paper we use within-student variation to account for ability bias. Putting ability bias aside, we take into account the theoretical explanations covered in Section II.
The academic momentum theory implies a positive effect of course load on student performance that, if it exists, should be part of the effect identified.
Time allocation theory also implies a causal effect: if students substitute away from study time on other courses to some degree when they add a new course, course load should reduce performance. However, in addition to its implied causal effect, the time allocation theory suggests several possible biases in the results, which we outline below.
If a student faces a negative shock requiring them to work in the labor market or aid their family more, they may be likely to reduce both the number of classes taken and the effort spent on each class at the same time, driving a positive correlation between classes taken and grades. Similarly, if course load is chosen before effort, then a consumption shock occurring between those two choices may require the student to lower effort more sharply in each class if enrolled in more classes, driving a negative bias.
The bounded nature of course loads can also drive a positive bias. If a student is incentivized to spend less time on school in a given period, but is already taking few classes, they can only reduce effort rather than taking fewer classes. Alternately, if students are planning to take four classes but one is overbooked Moore and Tan , they must replace the class to remain adequately enrolled. But for a student taking five classes and failing to enroll in one, replacing the overbooked class is optional, and more motivated students may be more likely do so.
Finally, there is the endogenous feature of the difficulty of courses. Students may choose the number of courses to take on the basis of how difficult they expect their course mix to be Volkwein and Lorang ; Cornwell, Lee, and Mustard If students take more courses when their courses are easier, the impact of course load will be positively biased.
Identifying the effect of the course load on performance should then focus on within-student variation to account for unmeasured ability. Further, it must take into account time-varying external factors likely to drive the time allocation decision, including the difficulty of the courses being taken in a given term, which is also endogenous.
However, this reduced-form effect is of interest, since any impact of a policy that increases course load on how students choose which courses to take should be a part of the effect of interest. Some analyses will attempt to disentangle the mechanisms at play by controlling for different features of the course mix.
Basic estimates will be biased by this correlation. This section gives our empirical results. We first report standard fixed effects estimates investigating the effect of course load on student performance, and then follow with examination of potential observed and unobserved sources of bias in these basic findings.
Table 2 shows our fixed effects estimates explaining student performance. Standardization occurs within the section, rather than for the same class across all sections. All standard errors are clustered at the student level.
In Column 2 we add student-level characteristics. The within estimates regressing standardized grade on taking a full course load appear in Columns 3—5 of the table, where the Column 3 results are for all students and Columns 4 and 5 results are limited to students who successfully graduated, or did not graduate, respectively. As shown in Column 1, taking a full course load is associated with a 0. Column 3 shows a within-student estimate of the effect of classes attempted on grade.
The effect is positive, but is so small 0. Importantly, we can reject to a reasonable degree of precision that the within-student relationship between a full course load and GPA is negative.
Additionally, Column 3 gives some insight to the likely correlation between fixed unobserved student heterogeneity and classes attempted. The reduction in the estimate from Column 1 to Column 2 to Column 3 implies that it is higher-performing students who do tend to choose higher course loads, and that many of the relevant high-ability characteristics are not measured in our data.
Footnote 3. Columns 4 and 5 of Table 2 show that the effect of a full course load is not meaningfully different for graduates or non-graduates. The effect is not large or significant for either subgroup, but is slightly more positive for non-graduates, who are less likely to choose higher course loads in the first place.
Examining variation in the effect in other ways besides eventual graduation success, we allow for the possibility that the effect of course load on grades differs by key student demographic characteristics. In results available upon request, we find that none of the coefficients for these interactions are significant the lowest p -value for a joint test of a set of interactions is 0.
Among categorical interactions, there is only one case—the second-lowest parental income bin—where the point estimate for that group is negative although it is insignificant. Finally, we examine whether the effect varies along the GPA distribution. It is reasonable to expect that high-ability students might have a null or positive effect of increased course load, but that students already having difficulties might not be able to handle the additional work.
Surprisingly, we find that the effect is most positive at the low end of GPA, and actually turns negative and significant for the 90th percentile. The potential concern that strong students drive the nonnegative result, and that full course loads would still be too much for low-GPA students to handle, is not supported. Student ability is one determinant of course load, which can be addressed using fixed effects, but there may be other important factors. Table 4 demonstrates the predictors of course load, which may help to determine the extent of the known bias in Table 2 and how it might be reduced.
As shown in Column 1, students taking more classes tend to have higher high school GPAs. Women are more likely to take full course loads. White students are most likely to take a full course load, with black students coming close behind. The other columns in Table 4 show results giving the influence of time-varying student characteristics on the number of classes taken, and include student fixed effects.
As shown in Columns 2 and 3, students tend to take more courses when their courses are easier. However, this is based not on the raw grades given out in class Average Grade in Course , but rather the grades given out relative to what might be expected based on the demographic characteristics and cumulative GPA of students taking the class Population-Adjusted Average Grade.
Columns 1 and 2 display the results when we add average course grades. While the left-hand side variable is already adjusted for average grades in each course, it is still possible that multiple easy classes may allow a student to over-perform in all of them at once. However, neither addition eliminates the positive relationship. The effect also is not eliminated by controlling for recent or prior performance in Column 3, indicating that the positive relationship is not a result of students, for example, taking more classes as they discover which fields they are good at.
Finally, in Column 5, we include all of the previous controls using the non-population-adjusted Average Grade, although it does not make much difference and also include fixed effects for declared major to account for differing institutional standards for how many courses students should take, especially in STEM.
The effect remains positive. Columns 3 and 4 are especially important. As will be discussed in the next section, a potential source of omitted variable bias is time-varying pressures from work or family, which may reduce course load and, at the same time, harm performance in class. It is reasonable to expect that these time-varying pressures may be serially correlated and would also affect grades. If they are, then lagged GPA acts as a proxy for these time-varying pressures, and so Columns 3 and 4 act as a partial test for this source of omitted variable bias, and fail to change the result.
The relationship between courses taken and grades remains positive and statistically significant. While the effect is small, the qualitative result of importance is that the relationship is not negative, even accounting for time-constant skills with fixed effects, the tendency to take many easy courses at once, the tendency to take more courses when one is doing particularly well, or institutional differences between majors.
There is an important confounder missing from our analysis in Sect. At the setting studied, significant portions of students work either part or full time while attending classes.
Theoretically, we would expect that changes in course load driven by factors that increase work hours would reduce grades and courses taken at the same time, leading to a positive bias.
Under the assumption that these pressures are serially correlated, lagged GPA is a proxy for these time-varying pressures, and in Table 5 a control for lagged GPA does not change the result. Still, controlling for lagged GPA is unlikely to eliminate all bias. Since we do not observe work hours, we perform a simulated sensitivity test, inspired by Rosenbaum bounds. In other words, for this analysis to be heavily biased enough by an omitted predictor such that we should be reporting a negative effect of number-of-classes on grade performance, that omitted predictor would need to have a correlation strength on the drawn boundary or to the top-right of it.
In the base model without additional controls, such an omitted predictor would need to have a correlation of about 0. Simulation described in Sect. Slight positive slopes are due to indirect manipulation of correlation and a discrete search space. In absolute GPA terms this translates to about 0. We would consider these to be fairly strong within-student correlations, possibly implausibly high. In addition to the simulation of Rosenbaum-like bounds, we also examine coefficient stability using the method in Oster With all controls included, unobserved controls would need to have variation orthogonal to the included controls that explains at least There is some possibility for tradeoff between explaining residual variation in the treatment and the control.
For a large negative effect to emerge with proper controls, time-varying outside pressures would need to be extremely strong determinants of course load, which contradicts our earlier discussion that in this context, much of the course load decision has to do with course bottlenecks outside the control of the student. In Table 6 we show some supplementary results related to the effects of taking more classes each term. In Columns 1 and 2 we predict persistence to the next term. These results address the possibility that additional classes, even if they do not weaken performance, may lead to burnout so that students are less likely to return.
Consistent with much of the literature in Sect. Columns 3 and 4, which are performed on a one-observation-per-student basis, examine the relationship between taking more classes per term and the rate of graduation as well as the time to graduation. Without any within-student variation, the causal identification for these estimates is very weak, and requires that the list of controls is sufficient. So we consider these non-causal. There is a very strong relationship between taking more classes and graduating; students who take one more class per term graduate more than 30 percentage points more often.
There is also a negative relationship between classes per term and time-to-degree. Students taking an additional class each term take on average 1. Finally, there are one-unit classes and a small number of nonunit classes in the data.
We ran all results again using a more direct count of classes excluding one-credit classes rather than the number of units divided by 3. In results available upon request, we find that the findings are very similar, and while point estimates differ slightly and some of the positive-and-significant-but-very-small results that we emphasized as being upper bounds became positive-and-insignificant none of the substantive conclusions about the effect of class taking changed.
Two results did change: 1 the effect of class-taking for the top decile of students changes from negative and significant to negative and insignificant, and 2 results were different for Table 4 , in regards to predicting which students take full course loads.
In particular, racial and gender effects become less prominent in Column 1, and the influence of average class grades is now negative Columns 2 and 3. The negatives of long time-to-degree are clear: earnings penalties for some Witteveen and Attewell , poorer overall performance, perhaps due to skill atrophy Brugiavini et al.
One way to improve time to degree would be for institutions to support efforts to increase credits per semester. However, there is a concern that additional credits may harm student performance. We find no evidence that increased course load harms performance, which supports the use of time-to-degree policy that uses course load as a lever. What impact can universities actually have on course load? One avenue is to give students more options in course scheduling, such as evening and on-line courses Witteveen and Attewell Advising can also play a prominent role in increasing credits per semester.
However, advising on course load may need to be intensive; in a randomized experiment, an informational intervention designed to increase course load had no impact by itself Huntington-Klein and Gill, Course load policy does not need to be the only tool used for reducing time-to-degree.
Increasing the number of summer classes is one way to increase the speed at which students accumulate credits. Policy on institutional barriers and resources, and student performance, is of course also important. Pike and Robbins find that the predominant factors affecting graduation rates are institutional characteristics that are either invariant or cannot be quickly or easily changed.
Yet, they do find that per pupil expenditures for instruction are associated with better 4-year and 6-year graduation rates. Though not stated explicitly by the authors, factors such as peer tutoring and university efforts at remediation typically fall into this category. Walvoord and Pleitz provide non-causal evidence that peer tutoring is associated with higher first year grades. Additional support for the notion that advising, tutoring, and mentoring matter to degree completion is found in Deming and Walters The authors find both positive effects of school spending on degree completion and that spending on categories of academic support services such as advising and tutoring are particularly hard hit when there are state budget shocks.
Student support services and student-engagement programs have the potential to improve time to degree, particularly in the current COVID environment. Based on survey results, Graham-Smith and Lafayette report that students with disabilities themselves highly value a caring staff that provides safety and security.
Gaddy outlines strategies campus disability service providers can take to improve student performance, including the use of test-taking and writing strategies.
Larson et al. Evidence on the importance of student-engagement programs on retention and degree receipt is mixed. For example, Caviglia-Harris and Maier find that living-learning communities are positively associated with retention.
In contrast, Culver and Bowman find no effect of first-year seminars on retention or four-year graduation. Johnson and Stage find little to no impact of so-called high-impact practices on 4-year or 6-year graduation rates at four-year public institutions. Course load policy should be thought of as sitting alongside this array of interventions that focus on student performance, and requires similar institutional attention.
Aforementioned advising efforts will be less effective if students are unable to take their preferred classes at the times needed to finish their course of study in a timely manner. This barrier to timely completion may be particularly relevant at a large public university such as ours. Course scarcity could cause students to take fewer classes, take the wrong classes, or change majors, all of which would increase time to degree Kurlaender et al.
Reporting on a survey of students and focus groups from three universities in the CSU system, Moore and Tan indicate that students themselves rate course availability as the number one obstacle to timely completion of their degrees. It is important to note, however, that causal evidence that course scarcity affects time to degree is mixed.
Kurlaender et al. Robles et al. In this paper we use administrative observational data in order to assess the causal effect of taking a full course load on student performance. We first focus on within-student variation in course load to avoid bias arising from student ability. We find no evidence of a negative effect of a full course load on student grades, and instead find a small positive effect.
Our baseline estimates are likely to be positively biased due to a relationship between course difficulty and course load, or the presence of time-varying outside work or family pressures. We find that controlling for course difficulty does not change results. We cannot measure outside pressures, but we provide four pieces of evidence supporting our conclusion. First, we note that, in the context studied, a significant determinant of course load is registration bottlenecks Moore and Tan , which would not bias results.
Second, we control for lagged GPA, which is a proxy for time-varying academic performance as well as time-varying outside pressures under the assumption that those pressures are serially correlated, and find no difference in results. Third, we perform a simulation and find that outside pressures would need to have a correlation of around 0.
Fourth, we show using Oster that obtaining a meaningfully large negative effect of course load on grades requires a fairly large 0. There are two important takeaways from the evidence presented here. The first is that we find, to a reasonable degree of certainty, that there is no meaningfully large negative causal impact of increasing student course loads. Policy directives to improve four-year graduation rates by increasing course load are unlikely to have meaningful negative effects on student performance and learning.
It is also unlikely that those average effects mask strong negative effects for weaker students, although it is possible that the As mentioned, our university policy encouraging students to take more credits was enacted several years after the first-time freshmen cohorts we study started their coursework.
Future work looking at the causal effects of full course load as induced by policy would be useful. An analysis comparing the effects of course load for transfer students, as opposed to first-time freshmen, may also be a productive avenue for future research, but transfer students were not available in our data.
0コメント