Where PISA is taking us: A Latent Growth Analysis of PISA Literacy from 2000 to 2012

Author(s):

João Maroco(submitting)Rosário Mendes(presenting)Vanda Lourenço

Conference:

ECER 2016

Network:

09. Assessment, Evaluation, Testing and Measurement

Format:

Poster

Session Information

09 SES 04.5 PS, General Poster Session

General Poster Session

Time:

2016-08-24

12:00-13:30

Room:

NM-Concourse Area

Chair:

Contribution

The Programme for International Student Assessment (PISA) is the most well-known large scale international student assessment project ongoing (OECD, 2014). It started in 2000 and has been repeated triennially with the purpose of evaluating 15-years olds’ knowledge and skills in three major knowledge domains: Mathematics, Science and Reading. Although PISA results were standardized in the first application to a 500-point average and 100 points standard deviation, and the overall average has been maintained close to the initial standard average, a large variation in results evolution of the around 70 participating countries and economies has been observed. While some countries have shown no significant evolution in their 15-year olds skills and knowledge according to PISA (e.g. Luxembourg, Finland or Spain); some others have greatly improved (e.g. Poland, Portugal or Brazil) while others have actually shown a drop in the PISA series results (e.g. New Zealand or the Netherlands). In this communication we report the results of a Structural Conditional Latent Growth Analysis on the PISA Mathematics results of the 57 countries (23 non-OECD countries; 34 OECD Countries) that have records in at least 3 of the 5 PISA cycles (2000, 203, 2006, 2009 and 2012). We hypothesized that student, school, family and country specific characteristics and their evolution in the PISA life-cycle may explain the evolution verified in these countries and therefore may be useful to educators and policy makers alike to pinpoint the variables that better explain the latent growth variation observed between countries participating in the PISA studies.

Method

A Linear Discriminant Analysis (LDA) was used to identify student, family, school and country predictor variables that could discriminate quartiles PISA results for Mathematics in the 57 countries pool that participated in at least 3 of the 5 PISA cycles. A conditional Latent Growth Model (LGM) was then used to study the countries’ growth in the PISA Mathematics literacy results from 2000 to 2012. A series of causal Latent Growth Model using the LDA identified variables as predictor variables was then fitted to the overall PISA mathematics results Latent Growing Models to identify which variables could explain the growth variation observed in PISA results amongst countries. SPSS Statistics and SPSS AMOS (v. 22; SPSS Statistics, An IBM Company) was used for LDA and LGM analysis.

Expected Outcomes

A LGM model fitted to the PISA mathematics results between 2000-2012 PISA cycles showed a good fit to the observed countries variation in PISA mathematics literacy (X2(7)=4.562; p=.713; RMSEA=.01; CI 90% ].000; .123[). The overall average growth of students’ mathematics literacy from 2000 to 2012 was statistically significant [Mean growth rate (b1)=5.353, p=.049], there was a significant variance of the mean growth rate (V(b1)=341.529; p=.006) suggesting that the growth rate is not homoscedastic between countries. Further, the average starting value in PISA 2000 (b0) was negatively correlated with the mean growth rate (r=-.832; p <.001) suggesting that the countries with higher starting PISA scores are growing at a slower rate in PISA than the countries with lower starting PISA scores. A conditional analysis of OECD vs. non-OECD countries showed that OECD countries had higher scores than non-OECD countries in PISA 2000 by a mean difference of 73.38 points (p<.001). However, the average growth rate of non-OECD vs. OECD countries improved by a mean difference of 15.85 points during the PISA life cycle (p=.003) showing that non-OECD countries are improving in PISA at a faster rate than OECD countries. The LDA identified as the main predictors of students differences in mathematics literacy the students’ ESCS and the Countries’ human development index in addition to the Per Capita Gross Domestic Product and countries’ expenditure in education (CEE). Focusing on CEE, the LGM showed that in the PISA life-cycle the participating countries increased their expenditure in education (as % of GDP) at an average rate of .454 (p=.002). On average, OECD countries had invested more 1.23% than the non-OECD countries at the beginning of the PISA life-cycle period (mean difference=1.23; p<.001) although there was no statistically significant difference in the CEE’s mean growth rate between OECD vs non-OECD countries (Mean difference=-.02; p=.864).

References

OECD (2014), PISA 2012 Results: What Students Know and Can Do – Student Performance in Mathematics, Reading and Science (Volume I, Revised edition, February 2014), PISA, OECD Publishing. http://dx.doi.org/10.1787/9789264201118-en

Author Information

João Maroco (submitting)

IAVE, I. P.

Lisboa

Rosário Mendes (presenting)

Instituto de Avaliação Educativa (IAVE, I. P.)

International studies

Lisboa

Vanda Lourenço

IAVE, I. P., Portugal

Search the ECER Programme

Search for keywords and phrases in "Text Search"
Restrict in which part of the abstracts to search in "Where to search"
Search for authors and in the respective field.
For planning your conference attendance you may want to use the conference app, which will be issued some weeks before the conference
If you are a session chair, best look up your chairing duties in the conference system (Conftool) or the app.