Chapter 2
Measurement Issues
Educational researchers are often interested in students’ knowledge, abilities, aptitudes or
personality characteristics. Many of these variables are not directly observable and this
creates particular measurement problems, which is an important consideration for
researchers reviewing other researchers’ studies or designing their own study.
Fortunately, many measurement procedures, such as intelligence tests, tests of reasoning,
or tests of verbal ability have been developed and evaluated and are accessible to
researchers. Whereas these tests and scales can be applied in many research contexts, the
proper use of these techniques requires an understanding of the fundamentals of
measurement and familiarity with statistical ideas which underpin test design and
interpretation of test data. Important concepts to be introduced or reinforced in this
chapter include classification and measurement, validity, reliability, criterion and norm-
referenced tests and standard error of measurement scales.
This chapter begins by considering measurement issues in education and the
classification and measurement of statistical variables, and goes on to consider properties
and assumptions of measurement scales, practical measurement decisions faced by
researchers, statistical ideas underpinning the use of psychological tests and scales and
finally concludes with advice on choosing a standardized test and interpreting test and
scale scores.
2.1 Measurement in Educational ResearchMany observations in education can be described in comparative assessments such as
‘this group is better at reading than that group’, ‘this child’s verbal reasoning ability is
well above average for his or her age’, or ‘this child’s work is not good enough’.
Statements such as these require judgment. One application of statistics in education is in
the use of numerical characteristics to replace such wordy descriptions and thereby to put
data into a context in which its meaning can be better evaluated and communicated. Put
simply, measurement means the rules that are used to assign numbers to statistical
variables. That is the process whereby particular properties or attributes of a variable are
given a numerical value. Educational researchers, whatever their interest, will invariably
be concerned with measurement; either how to measure variables in their own study or
how to interpret other researchers’ measurements.