“The effect of increasing class size in tertiary education isnot well understood,” schrijven zij op
The organisation of university education is increasingly in thespotlight, both in academic and policy circles. Recent research hasstressed the importance of higher education in providing positiveexternalities within firms (
Concurrently, most OECD countries have adopted policies that haveled to dramatic increases in university enrolment during the lastdecade. The average annual increase in university enrolment in OECDcountries during the period 1995-2005 was above 4%. In the UK, thisgrowth has occurred at both the undergraduate and graduate leveland across a wide range of universities. The UK is actually at thelow end of enrolment growth within the OECD. Between 1998 and 2005,the US experienced a 30% increase in student enrolment, forexample.
Such breathtaking increases in enrolment inevitably lead touniversity students facing larger class sizes. The effect ofincreasing class size in tertiary education is not wellunderstood.
The established literature on class size effects in primary andsecondary schools provides useful guidance (
Identifying class-size effects from within-studentvariation
To address this policy question, we estimate the impact ofclass size on the final exam marks of graduate students in aleading UK university between 1999 and 2004 (Bandiera etal. 2010). As we observe the same student being exposed tovery different class sizes, we estimate the effects of class sizeon students’ exam performance by comparing the samestudent’s performance to her own performance in courses with smalland large class sizes.2
It is important to stress that, on average, most of the variationin marks is due to fixed students’ characteristics and notuniversity inputs.3 On average theperformance of a given students only varies by around 7% of theaverage mark across her courses. We shed light on how much of thiswithin-student difference is attributable to differing class sizesthe student faces.
The effect of class size on students’ performance is – as expected- negative; students do worse in big classes. Namely, a givenstudent receives lower marks in courses withlarger classes, everything else equal.
To get a sense of the magnitude of this effect, our estimates implythat a one standard deviation increase in class size from the mean(that is going from the average class of 56 to a class size of 89)would decrease the mark by 9% of the observed variation in markswithin a given student. These estimates, however, mask twoimportant forms of heterogeneity: (i) the impact of class sizevaries across the range of class sizes; (ii) the effect of classsize varies across students.
On the first form of heterogeneity, the negative effect of classsize on student exam performance is large and negative only in thesmallest and the largest classes. There is no class size effectacross a wide range of intermediate class sizes. The magnitudesimply that moving the average student from a class of 10 to a classof 25 leads to a drop in exam performance of around 12.5% ofwithin-student standard deviation. Increasing the class size from25 to 45 determines a further 12.5% drop.
In contrast, there is no impact in a wide intermediate range, whilemoving from 80 to 150 determines a further drop of 25% in thewithin-student standard deviation. If moved from a very small class(of size 10) to a very large class (of size 150), the averagestudent can be expected to suffer a loss corresponding to about 50%of the overall variation in exam marks the average studentexperiences across all of her courses.
The second form of heterogeneity concerns students’ ability.Students at the top of the mark distribution are those mostaffected by class size. The effect is almost four times larger forstudents in the top 10% of the distribution of exam marks than forstudents at the bottom 10%, and about 50% larger than the averagestudent. This heterogeneity is most apparent in the largest classesand virtually non-existent for a range of intermediate class sizes.This implies the highest-ability students would benefit the most,in terms of academic achievement, from any reduction in classsizes, when class sizes are initially very large.
To shed light on the underlying mechanisms for the class-sizeeffect, our analysis uses information on teachers’ assignments toclasses and on students’ characteristics. We find no evidence thatdepartments purposefully assign faculty of differing quality todifferent class sizes, and we find no evidence that faculty membersalter their behaviour when exposed to different class sizes. Itappears that the preparation and delivery of lectures isindependent of the number of students taught.
On student characteristics, the class-size effect does not varywith proxies for students’ wealth. Hence if larger classes resultedin lower grades because students had more limited access to librarybooks or computer laboratories, the effect should have been smallerfor students who can purchase these inputs privately.
Moreover, the class-size effect does not vary with student’sfamiliarity with this particular university as an undergraduate orwith the UK system generally. This casts doubts on the relevance ofmechanisms that work through the information students have, such astheir awareness of other local resources (for example otherlibraries in the area), or their knowledge of the characteristicsof faculty, courses, or departments.
Discussion and policy implications
Against a backdrop of rapidly increasing enrolment ratesin tertiary education, our analysis has important policyimplications. Class size matters for student performance andparticularly for the most able students.
However, reducing class size is not always an effective strategyand is certainly not effective for all students in the same way.Reducing the size of very large modules (above 100) could be acost-effective way to improve students’ performance. For modules inthe range 30-100 reducing class size could be a rather ineffectivestrategy, while for classes below 30 it could be a valid but notnecessarily cost-effective strategy.
Attention should be devoted to other inputs in such cases, and morerefined and cost-effective solutions than pure number countingshould be identified. To this end, it is important to have a betterunderstanding of the mechanisms that link class size andperformance.
Although student-to-staff ratio is a commonly used indicator ofquality both in national and international comparisons, this mightbe a noisy measure of quality over this intermediate range of classsizes.4 Given the mechanisms our datarule outs, there appear to be at least two ways that larger classesreduce students’ performance. First, changes in student behavioursuch as their attentiveness or participation. Second, reducedresource availability, such as library books or faculty time duringoffice hours.
As the best students are the most affected, that could imply thatlarge classes induce a reduction in tutoring activity rather than asubstantial deterioration in classroom conditions. It is reasonableto expect that the best students are able to compensate classroomdeterioration at least as well as other students. However, the beststudents are also those that benefit the most (in terms of bothlearning and motivation) from contact with teachers. They,therefore, suffer the most in terms of reduced performance whensuch contacts or tailored feedback is less frequent.
1 This is particularly evident in the UK, where concernson the increasing student-to-staff ratios in higher educationinstitutions have recently been expressed in a report of theDepartment for Innovation, Universities, and Skills and by the mostimportant unions of university teachers.
2 Our estimates are therefore purged from confounding effects thatarise because students choose which modules to take. For instance,if more able students were to choose smaller classes, across-student comparison would capture both the effect of classsize and the effect of student ability. The within-studentcomparison only captures the former.
3 Characteristics like ability and motivation are certainlyaffected by previous experiences and schooling, but they can betaken as given for what concerns university policy.
4 This choice is somehow justified by the fact that, in order toevaluate teaching quality, the student to staff ratio probablyremains the only globally available and comparable indicator.
Auteurs: Oriana Bandiera, Valentino Larcinese en ImranRasul