Which of the following will reduce the precision of our estimate of the partial effect of salaries on average testscores? (Q A larger correlation between expenditure on salaries and materials. C) A larger sample size. 0 More variation in salaries across schools 0 Less variation in the unobserved error, u. We are conducting a stud},r relating school expenditures to average student test scores (student academic achievement] in primary schools. The data we have is a large sample of schools from the Department of Education. For each school, we observe: average Year 5 test score, averaged across all their students in Year 5 in 2016 (testscore) a full set of annual expenditures in dollars recorded into 4 broad mutually exclusive but exhaustive sub-categories: teacher salary expenses (salaries) teaching materials (materials) maintenance of buildings and school grounds (maintenance), and other (other) total annual expenditure in dollars (totexp). Our model of interest is: testsaor'e = g -l ,61 salaries + 5} materials -l- ,83 maintenance -l- ,84 other + a [C1] Why can't we also include totexp as an additional explanatory variable in model C1? O It would result in micronumerosity. O It would cause perfect multicollinearity. O It would cuase heteroskedasticity. O The sample size is too small