Progress and setbacks in educational performance in Latin America: factors influencing learning achievement over the last decade

• In Event: Education Development, Policy and Education

Proposal

On average, Latin America has not achieved an improvement in learning achievement results in the last decade (UNESCO, 2021, 2015; OECD, 2019). Although some systems achieve improvements between 2010 and 2020, most have not made significant progress. This raises the question: why the region has not improved its educational performance. The literature identifies multiple variables and factors, which makes answering the question extremely complex. The improvement in educational quality depends on factors at the system level, at the school level, as well as at the individual and family level (Woessmann, 2016). Institutional and structural characteristics of educational systems such as administrative dependence, geographic location, socioeconomic differences, teaching policies, administrative and structural organization of the school have a significant association with the performance of their students (Parvez & Laxminarayana, 2022; Eigbiremolen et al., 2020; Dronkers & Robert, 2008; UNESCO, 2021; McKinsey & Company, 2007; Mourshed et al., 2010).

At the student level, the importance of the socioeconomic and cultural background of families on academic performance has been reported (Recalde Correa, 2018), as well as the negative association between school absenteeism and learning achievement (Aucejo & Romano, 2016; Simon et al., 2020; Vesic et al., 2021). On the other hand, attendance in early education is shown to be a favorable predictor of future learning achievement outcomes (Manfra et al., 2017; Schmerse, 2020; UNESCO, 2015, 2021).

At the school and classroom level, the importance of various factors with educational achievement is also recognized, such as teacher practice and development, classroom climate (Fernández et al., 2018), learning opportunities (Santibañez & Fagioli, 2016), pedagogical support and teaching organization (Miranda & Treviño, 2020) and, negatively, violence, victimization, and bullying (Laith & Vaillancourt, 2022; Mundy et al., 2017).
International large-scale assessments provide an opportunity to contrast differences between different education systems (Manfra et al., 2017; Schmerse, 2020), However, few studies consider most Latin American countries at more than one measurement point in time. Thus, this research analyzes the relationship of structural, individual, and school-level factors with learning achievement in Latin American countries during the last decade, based on the two UNESCO measurements (TERCE in 2013 and ERCE 2019).

Method:

Using a two-level hierarchical regression model, this study seeks to explain differences in learning achievement through the relationship of institutional and structural factors, individual-family, and school context and management variables. For this purpose, the results of the UNESCO Regional Comparative and Explanatory Regional Study (ERCE) from the periods 2013 and 2019 are used. This study measures the learning achievement of third and sixth grade students and its associated factors. The results of the study are reported on a scale with a mean of 700 points and standard deviation of 100. In turn, the indexes provided by these studies were used, which were normalized for both periods (2013 and 2019).
Six different models were tested, successively incorporating variables on family and individual student characteristics (family socioeconomic level, gender, preschool attendance, non-attendance, grade repetition, family expectations, parental involvement); structural variables (classroom socioeconomic level, administrative dependence); and school and classroom context (victimization, school violence, learning context and processes, work climate, teacher collaboration, school leadership).

Preliminary results (only for the reading test):

The intra-class correlation (ICC) of the null model indicates that 40.5% of the variance in sixth grade reading achievement can be attributed to differences between schools.

On the individual and family level variables, it is observed that grade repetition is negatively related to a little less than half a standard deviation in sixth grade reading (44 points). The socioeconomic level of the families also presents a high relationship with learning achievement in reading of up to one third of a standard deviation (37 points) when it increases by 1 standard deviation. Early education attendance is associated with up to 14 points in reading (or 14% of the standard deviation). To a lesser extent, truancy has a smaller but significant and negative relationship of up to 8 points on the test. Parents' and families' expectations that their children will attain higher education, which reaches up to 40 points (more than one third of the standard deviation).

At the level of structural variables, it is observed that a standard deviation in the average socioeconomic index of the students in the classroom leads to a positive effect of up to almost half a standard deviation (45 points) in the sixth grade reading test. On the other hand, when the school is private, it explains up to 16 points.

Finally, the school and classroom context incorporates students' perceptions of violence within the school, classroom disruption, challenging classroom composition, and teaching difficulties. All of these are negatively associated with reading achievement, but modestly so. The same is true when including the leadership and work climate variables.

Discussion:

Most of the variance of results among students in Latin America is explained within schools, only 40% is explained between schools. When performing the different specifications, the variables at the family level, personal attributes and students' educational trajectory explain a considerable part of the differences between schools, which is consistent with the literature. However, they explain very little of the variance within schools, leaving most of the variance unexplained. This is reinforced by the fact that the average NSE of the school becomes the most important factor in explaining the differences in performance between schools.

On the other hand, both the school context variables (such as violence, classroom disruption, difficulty in teaching and challenging classroom composition) and the leadership and work climate variables, although significant, their association is very modest, so that the models used in this research do not offer a convincing explanation for the differences in learning achievement. The challenge arises of measuring this and other phenomena in order to explain a greater proportion of the variance within schools.

Author

• Francisco Gatica, UNESCO