Results in TIMSS linked with earlier studies in maths and science

Published

27 September 2022

at

For the first time has all data of grade eight students or equivalent of earlier international large-scale assessments in mathematics and science been linked and scaled to the series of TIMSS. This means that we now have comparable data for grade-eight mathematics and science since the 1960s.

International large-scale assessments have been increasingly important for researchers, policymakers, and politicians for the evaluation of educational systems as well as in the school debate. TIMSS (Trends in International Mathematics and Science Study) measures grade four and eight students’ achievement on the country level in mathematics and science and has been conducted every fourth year since 1995. In the last TIMSS participated over 60 countries.

Several earlier studies before TIMSS

Portrait of Erika Majoros. — Erika Majoros.

In her doctoral thesis, Erika Majoros has linked earlier large-scale assessments in both mathematics and science to TIMSS. The first international large-scale assessments on mathematics and science conducted by the International Association for the Evaluation of Education Achievement (IEA) were administered in 1964 and 1970-71. They were followed by a new set of assessments in 1980-82 and 1983-84 before IEA established TIMSS which has been ongoing since 1995. The IEA has not linked these earlier studies to TIMSS due to difficulties with comparability and technical possibilities.

"IEA did not make any link partly because of changes in their test administration, i.e. how the assessments are constructed and carried out, and partly because of technical challenges that made it difficult to handle all data. But methodological and technical improvements have allowed me to create a meaningful linking", says Erika Majoros.

A new method makes linking possible

In her thesis, Erika Majoros has applied a design that makes it possible to establish a link between the earlier studies and TIMSS.

"I found items that were repeated from the earlier studies in the TIMSS surveys. The repeated test items were not as many as between TIMSS assessments but I applied a design that combined different linking methods and thus allowed for creating a meaningful link", says Erika Majoros.

The earlier studies and TIMSS are different in several ways. In TIMSS every student only takes a sample of items from the whole item pool, while in the first mathematics study there were fewer items and all students made them all. Another difference is that the emphasis on different content domains, such as algebra or geometry within mathematics, has changed over the years. The definition of samples of students has also changed. In the early studies, it was based on age (13- and 14-year-old students) or grade, in TIMSS on grade (typically grade eight).

Great use for analyzing long-term effects

Thanks to Erika Majoros’ design we now have comparable data of grade-eight mathematics and science since the 1960s.

"It is of great use for researchers and others. The trend scale is now available for further analysis to investigate the long-term effects of for example policy, curriculum, and changes in the student population, says Erika Majoros. It should be noted that the sampling differences need to be considered when using the scales."