Till sidans topp

Sidansvarig: Webbredaktion

Sidan uppdaterades: 2012-09-11 15:12

- Göteborgs universitet
- Forskning
- Tackling Unobserved Heter…

Till innehåll
Läs mer om hur kakor används på gu.se
# Tackling Unobserved Heterogeneity in Socioeconomic Status and Opportunity to Learn, and their Effects on Academic Achievement

## Sammanfattning

Konferensbidrag (offentliggjort, men ej förlagsutgivet)

Författare |
Victoria Rolfe Kajsa Yang Hansen Rolf Streitholt |
---|---|

Publicerad i | Paper presented at ECER 2017 Copenhagen, Denmark |

Publiceringsår | 2017 |

Publicerad vid |
Institutionen för pedagogik och specialpedagogik |

Språk | en |

Ämnesord | Unobserved Heterogeneity, Educational Equity, Socioeconomic Status, School Mix, Opportunity to Learn, Factor Mixture Model, TIMSS, PISA. |

Ämneskategorier | Internationell pedagogik, Pedagogik |

The TIMSS 2015 results revealed an improved mathematics score for 8th graders in Sweden. However, the achievement gaps with respect to socioeconomic and ethnic background also have increased dramatically (Mullis, Martin, & Loveless, 2016). It is thus interesting and important to examine the factors that might be lying behind the observed trends in educational equity and quality. A possible precursor of this period of fluctuation in achievement and the noted achievement gap is the influence of wide-ranging structural school reforms implemented since the late 1980s. These reforms, notably the free school choice policy, decentralization and deregulation, have transformed the school system creating a marketplace for providers and may have triggered changes in the characteristics in school compositions, in turn have an influence on student outcomes (Thrupp, 1995; Thrupp & Lupton, 2006). Another possible explanation is that changing opportunities to learn (OTL) have had differential effects on equity and efficiency. Many countries have revised their curricula, and other reforms, such as choice of schools have changed the social and ethnic composition of the schools. Changes in the learning and teaching environment may, furthermore, constrain or strengthen the OTL (e.g., Authors, 2016a; 2016b). In the study we focus on changes in OTL of mathematics content and SES between 1995 and 2015 in Sweden. We try to establish a causal link between the changing SES, OTL and the changing achievement gaps between unobserved subgroups of students. This theoretical framework opportunity to learn (OTL) distinguishes between the prescribed curriculum, the taught curriculum, the assessed curriculum and the achieved curriculum. The prescribed curriculum is usually specified at the system level. The taught curriculum is at the classroom level. The assessed curriculum refers to the tasks included in the achievement tests, and the achieved curriculum refers to what individuals have learnt, as reflected in the assessment (e.g. Schmidt, Zoido & Cogan, 2014). The theoretical model also distinguishes between antecedents and contexts at three levels: the educational system level, the school/classroom level, and the student level. In the proposed study, we focus on the antecedents at the school/classroom context is determined by allocation of resources and translation of the prescribed curriculum into teaching and the antecedents at the student level are individual characteristics such as characteristics of the home, which influence the achieved curriculum. This multilevel model thus specifies direct and indirect effects on the achieved curriculum of the other factors in the model. According to the model, there are direct effects of OTL and student characteristics; while the factors at the school/classroom and educational system levels influence the achieved curriculum indirectly. As mentioned before, the free choice of schools accompanied with other reforms in Sweden has changed the landscape of Swedish schools. As a consequence, the school mix in terms of student intake body’s characteristics and school resources differed largely across schools. We have good reason to assume that different groups of students and schools may have rather differentiated factor structure in the measurement and structural relations. The unobserved heterogeneity in the sample population need to be taken care of otherwise the estimates of effects will be biased. Against this background, the paper seeks to answer the following questions: 1. How do SES and OTL impact mathematics achievement in TIMSS 1995 to 2015? 2. Does the factor structure of SES an OTL differ over some unobserved subgroups of students and schools in each TIMSS studies from 1995 to 2015 in Sweden? 3. Does the relationship among SES OTL and math achievement differ across the subgroups of students and schools? Methods/methodology This paper uses data from TIMSS 1995 to 2015 pertaining to grade eight students in Sweden. Factor mixture modeling is used to examine unobserved heterogeneity in the factor structure of SES, school mix and OTL for each TIMSS study. It is assumed that the factor structures are not the same for the sample in each TIMSS study and the differentiated factor structure over the unobserved heterogeneous student groups are captured by the latent class variables included in the factor mixture model (Lubke & Muthén, 2005). The variables included in this model are drawn from the student and school questionnaires in TIMSS studies since 1995. Two-level factor mixture model (Henry & Muthén, 2010) is used to account for the hierarchical data structure and allow simultaneously detect the differentiated factor structure as well as effects on math achievement for unobserved subgroups of students. It is also interesting to see how the subgroups change the characteristics over different TIMSS studies over time. The analyses are conducted in Mplus (Muthén & Muthén, 1998-2015). Outcomes Preliminary analyses indicate that socio-economic and ethnic inequality in mathematics achievement differ significantly across schools for some countries. The school student mix too appears to have a varying effect on the between-school differences in the relationship between student’s socio-demographic characteristics and their math achievement. We expect that, in addition, differences in OTL between 2011 and 2015 have an effect on the sociodemographic inequality in mathematics achievement across different educational systems