Country Background Report – Denmark
Download 1.6 Mb. Pdf ko'rish
|
10932 OECD Country Background Report Denmark
Effectiveness
In line with international findings (Hanushek 1996, Hanushek 2003), a number of stud- ies find that the level of school expenditures per student generally has no or virtually no effect on student performance 36 (Houlberg, Østergaard Larsen & Rangvid 2013, Calmar Andersen & Mortensen 2010, Laier Christensen 2009). It should be noted that these studies, along with most other studies analysing the relationship between school expenditures and student performance, hinge on conditional independence in relation to causal inference and generally suffer from potential self-selection bias, especially when the unit of study is the school level. For instance, it is not random in which mu- nicipality people reside and which school parents choose for their children. 35 See Also appendix 1. 36 Student background controlled for a variety of socio-economic characteristics of the individual students. 87 Keeping this in mind, (Calmar Andersen & Mortensen 2010) 37 take the analysis a step further and find that increasing expenditures over the years in which the student at- tends school may be beneficial for student performance. More robustly and significant- ly it is found, however, that stable budgets in the years of schooling have a positive effect on student performance at the final exams of the Folkeskole (Calmar Andersen & Mortensen 2010). The design of the study provides no empirical explanations for processes and mechanisms behind the importance of stable budgets, but the study points to a set of theory-based arguments on economic and personnel stability as a pre- condition for stability in administrative systems, such as structural, mission, technolo- gy and procedural stability. Moreover, personnel stability may also require stable budgetary developments that allow personnel planning and adjustments made through natural wastage (Calmar Andersen & Mortensen 2010). Higher expenses do not imply better outcomes for the students when focusing on the effect of resources spent in the schooling sector on student performance measured as SES-adjusted, school-leaving examination marks calculated over a five-year period from 2002 to 2005 (Laier Christensen 2009). As schooling is a ten-year event, and the expense variables included in this benchmarking study only cover 1-5 years of the school period, which may be the explanation of the insignificance found (Laier Chris- tensen 2009). When including main explanatory variables as gross expenses for schooling per student and expenses for teacher wages per student at the municipality level, along with controls for share of low-educated mothers, share of low-income families, share of unemployed or studying mothers, share of non-western mothers, and real estate prices (square metres), a positive but insignificant association between ex- penditures and student performance is found (Laier Christensen 2009). The somewhat lacking timing congruence between the outcome and the explanatory variables, as well as conditional independence, is a problem in the study. When studying the effect of municipal school expenses on student performance meas- ured as school-leaving examination marks, it is found that spending over all years of schooling is a more important determinant of student attainment than spending in the last year of school attendance (Houlberg, Østergaard Larsen & Rangvid 2013). Keep- ing potential self-selection bias in mind, and the awareness that self-selection bias will tend to bias the estimated association between expenditures and student performance toward zero, (Houlberg, Østergaard Larsen & Rangvid 2013) not only focus on the average effects but also test for potential heterogeneous effects. A substantially limited but statistically significant positive effect of increased spending is found for municipalities with (need-adjusted) spending levels below the average, i.e. spending up to the average seems to be fruitful, while increased spending above the average does not seem to be beneficial for student performance. When it comes to stu- 37 Including a robustness test that – given the data available – seeks to control for self-selection bias at municipal level. 88 dent background, the spending level seems to be statistically beneficial for weak stu- dents but not for students with a more advantageous socio-economic background (Houlberg, Østergaard Larsen & Rangvid 2013). Despite the satisfying amounts of data included in the analysis, the study hinges on conditional independence. This is not a satisfactory design for establishing causality and does not provide a sound basis for calculations of potentials. Effects on 9 th form school-leaving examination marks of “structural conditions” relat- ing to teacher characteristics, class and school size are studied in order to assess whether high performance can be achieved by cost-neutral change in policies (Teglgaard Jakobsen et al. 2012). Effects of “structural conditions” are measured as teachers’ sickness absence, teachers’ teaching share, school size, class size, students per teacher, teaching hours per week, teaching hours per student and the student com- position at the school (Teglgaard Jakobsen et al. 2012). Only teachers’ sickness ab- sence, teachers’ teaching share, class size and the student composition are significant. The outcome variable of the study is (probably) school-averaged predicted exam marks. The predictions come from a model with exam marks as dependent variable and traditional socio-esocio-economicconomic variables as explanatory variables. The study hinges on conditional independence and is methodologically weak with regard to documentation of the outcome variable. (Teglgaard Jakobsen et al. 2012). Download 1.6 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling