All the specifications estimated below include time fixed effects captured
by year-test specific dummies to allow for intercept heterogeneity in the six-
teen cross sections (four tests in each of the four years): λt = Pj4=1 P4k=1 δjkDjk,
where Djk is a time dummies for year j and test k, and δjk is the correspond-
ing parameter. We also control for individual characteristics such as year
of registration, gender, foreign language, and live away from home, and in-
clude sets of dummy variables for high school type, parental education and
occupation, and province of origin.
5 Results
This section presents the estimation results. We start by estimating equa-
tion (1) by OLS, and examine the impact on the estimated coefficient for
attendance of controlling for unobservable factors such as ability, effort and
motivation. We then consider the results obtained with IV for the same set of
specifications. Next, we present estimates obtained for panel data estimators
(random effects and fixed effects). Finally, we examine the respective effects
of lecture and class attendance on performance.
Table 2 reports OLS estimates of alternative specifications of the rela-
tionship between academic performance and attendance. All specifications
produce a coefficient estimate for attendance that is positive and statisti-
cally significant at the one per cent level. In the basic univariate specifica-
tion (column 1), the point estimate indicates that one additional percentage
point of lecture attendance corresponds to a 0.09 percent improvement in
performance. As reported in column 2, the addition of a set of controls for
individual characteristics does not affect the estimated coefficient for atten-
dance. In this specification, year of enrollment and live away from home are
negatively and significantly associated to performance.
Next, we consider how controlling for unobservable factors, such as abil-
ity, effort and motivation, affects the estimated coefficient for attendance.
Adding either the set of ability proxies (column 3) or the set of effort and
motivation indicators (column 4), the estimated coefficient for lecture at-
tendance falls to 0.076 and 0.087, respectively.18 Adding both sets of indi-
18 Note that, to the extent that GPA reflects the effect of attendance in other courses,
and that attendance is positively correlated across courses, the inclusion of this variable
could lead to underestimate the effect of attendance on performance in Introductory Mi-
croeconomics.
11
More intriguing information
1. Psychological Aspects of Market Crashes2. Restructuring of industrial economies in countries in transition: Experience of Ukraine
3. BEN CHOI & YANBING CHEN
4. NEW DEVELOPMENTS IN FARM PRICE AND INCOME POLICY PROGRAMS: PART I. SITUATION AND PROBLEM
5. The name is absent
6. Reputations, Market Structure, and the Choice of Quality Assurance Systems in the Food Industry
7. Political Rents, Promotion Incentives, and Support for a Non-Democratic Regime
8. Towards Teaching a Robot to Count Objects
9. Modelling Transport in an Interregional General Equilibrium Model with Externalities
10. Plasmid-Encoded Multidrug Resistance of Salmonella typhi and some Enteric Bacteria in and around Kolkata, India: A Preliminary Study