Developing and testing a theoretical framework for assessing extended response questions in GCSE Science

This study aimed to develop, test, and validate a theoretical framework that could be used to write levels-based mark schemes for extended response questions in GCSE science. The work focused on questions which require students to give a scientific explanation or provide an argument. The development was informed by the work of researchers who have evaluated argumentation in the science classroom and also took into account the grade descriptors that are used to determine the cut scores for key grade boundaries during the awards process for GCSE science. The framework was used to write mark schemes for five questions from the January 2012 GCSE Science examinations. The mark schemes were used to mark scripts (n = 19 to 26) from those examinations. The marks awarded were compared with those given by the examiners who originally marked the questions. To ensure the theoretically-based mark scheme could be used by others, three senior examiners also marked two of the questions. Senior assessors (n = 12) for GCSE Science were asked, through an open response questionnaire, to comment on the framework and on its potential usefulness for writing levels-based mark schemes. Comparison of marks awarded using the theoretically-based scheme with those using the original scheme showed that the two schemes produced similar rank orders (Kendall’s coefficient τ = 0.61 to 0.83, n = 19 to 26). When other examiners used the theoretically-based scheme they awarded similar marks to those given by the researcher. These two outcomes suggest that the theoretically-based framework could be used for the proposed purpose. The senior assessors were generally positive about the usefulness of the scheme as a starting position for writing mark schemes and some recognised its potential to provide consistency of standards between different papers in the same session and across time.