Martin Caust

Thesis abstract

Abstract

The documentation of learning is a weakness of all schools and systems, leading to complaints about the lack of information and a press for teacher accountability.  Current solutions to increase information about learning and improve accountability promote standardised (and national) testing of student cohorts and/or better use of often-archaic classroom assessment results.  System-wide testing, while not without value for some purposes, is very limited in its contribution to improving classroom practice. In particular testing is a process detached from the needs of classroom teachers and given the time for results to be returned, unhelpful in timely decision making.

Assessment of students by teacher judgement is a general feature of classroom teaching but its quality is often unknown.  This thesis addresses the history and application of teacher judgement assessment and then analyses teacher and test assessments of the same populations of students (from South Australia in 1997 and 1998).  The analyses establish the comparability of the assessment processes, and thus one basis for inferring the quality of teacher judgement.  The purpose is to test the feasibility of using teacher judgement assessments, calibrated to scales of learning, as the prime data to record, manage and report learning and monitor its change over time.

In curricula structured in levels, as apply in some Australian school systems, one possibility for recording assessments is in the form of the level judged to be most recently achieved.  Over an extended time frame a general trajectory of learning for each student can be documented. If the progress made as a student learns new skills, knowledge and understandings could be assessed and recorded by a teacher in finer detail than a level, a basis might exist for documenting learning with utility for teachers, students and all other parties interested in being kept informed. These two broad ideas, the teacher’s concept of learning in a specific strand of the curriculum and the mandated test as one method to describing that learning, are brought together to appraise the feasibility of creating methods of assessing and recording learning, built upon the constructs rather than any particular test or assessment process.

The data analysed are unique.  They are limited to two calendar years (1997 and 1998) for two learning areas and are useful in estimating the potential for teacher and test assessments to track the learning development of students over time in the same fashion.  Within the limitations of the data the potential of teachers to record the learning development of students directly, using broad scales to locate their current learning status is confirmed.  Very strong similarities are found in the general characteristics of the data once the teacher scale is transformed to the scale of the test.  Both assessment processes show increments in mean leaning for age cohorts grouped in 0.1 of year of age and smooth growth trajectories with age and Year level.  Both processes show marked gender differences for English language, trivial gender differences for mathematics.  Both processes show within Year level patterns by age and gender that are consistent with test data analyses found elsewhere. 

When case studies for individual schools are examined, it is clear that at some sites teachers assess with high correlation to the test scores, indicating the potential for easily recalibrating some teachers to increase the match of the assessments from the two processes. It appears potentially feasible to design classroom and school assessment systems on the basis of teacher judgement assessment data as the prime data source.  Test data can be integrated readily and usefully into the scheme.  The issues that need further consideration are outlined along with the general implications for support to teachers, training and re-training and some broader data management issues for classrooms, schools and systems. Subject to the resolution of a number of design issues, schools and school systems might then optimise the skills of teachers as both managers and documenters of learning.   This would allow for the professional skills of teachers to be acknowledged and capitalised upon.  Rather than the assessment skills of teachers being directly derided, or derided by implication as a consequence of externally imposed testing procedures, testing arrangements might be reconfigured to support and confirm the quality of teacher judgement assessments.

Download full thesis here.