Academic journal article Psychological Test and Assessment Modeling

Large-Scale Assessments: Potentials and Challenges in Longitudinal Designs

Academic journal article Psychological Test and Assessment Modeling

Large-Scale Assessments: Potentials and Challenges in Longitudinal Designs

Article excerpt

1.Introduction

Large-scale assessments are of utmost importance in educational research. Studies like the Programme for International Student Assessment (PISA), the Third International Mathematics and Science Study (TIMSS), or the international Progress in Reading Literacy Study (PIRLS) lead to fundamental knowledge gains concerning students' competencies at different ages, competence distributions in total as well as in subgroups, and covariations of students' competencies with variables at the family level (social or migration background) or the school level (teacher and school characteristics, school type, class composition). Among other things, PISA allows us to analyze the competence level of 15-year-olds in great detail with a special emphasis on their learning environments. By means of the PISA data of 2012 (OECD, 2014) it can be shown that within the OECD countries 23% of all students show very low mathematical competencies (Level 1 or below) whereas 3% reach very high mathematical scores (Level 6). The corresponding analyses conducted show correlations between mathematical competencies and gender, as well as between migration background and social status. However, these results do not allow us to derive any explanations of how and at what age different competence levels of adolescents develop and which learning environments (not only formal, but also nonformal and informal) foster or compensate the effects of given background aspects. Furthermore, the effects of a given competence level on mid-term or long-term educational and vocational trajectories and life-course development cannot be studied by means of the PISA data (because they are cross-sectional). Concretely, PISA does not allow us to answer any of the following questions: How will competencies develop in the future? How successful are adolescents with low, median, or high competence levels in achieving a school degree, finding an apprenticeship position, or entering higher education and the labor market? What conditions can promote the educational and vocational processes even in adolescents with Level-1 (or below) competencies or hamper these processes even in the most competent Level-6 students?

Because studies such as PISA can only take a snapshot of the competencies of students of a certain age group, questions about developments, about supportive or destructive environments, about processes (e.g., in aspiration setting) cannot be dealt with. Instead of referring to a single-time-point measurement, questions such as those above can only be worked on by using longitudinal data (also called panel data). This paper aims to highlight the potential of large-scale assessments following a longitudinal design. For this purpose, the National Educational Panel Study (NEPS; cf., Blossfeld, Roßbach, & von Maurice, 2011) with a special emphasis on the educational pathways of ninth graders is used as an illustrating example of a general elaboration on the specific characteristic of such a design. The NEPS is based on the principles of life-course research (Elder, Johnson, & Crosnoe, 2003; see also Elder & Giele, 2009) as well as the perspective of lifespan developmental psychology (Baltes, 1990; Baltes, Reese, & Lipsitt, 1980). Against this theoretical background, the clear focus is on educational trajectories as well as competence development and on the relevant processes behind them.

The NEPS was built up with six different starting cohorts (SC) ranging - at Measurement Point 1 - from early childhood through Kindergarten and school to higher education and adults aged 23 to 64. Figure 1 shows the design of the multicohort sequence design implemented in the study.

In order to follow educational processes and competence development in these six starting cohorts over time, appropriate sample sizes had to take into account the expected variations of multiple life trajectories. All cohorts were sampled at an individual or institution-based level - all are representative of Germany. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.