A statistical variable is one which is measured or enumerated (counted). When
interest focuses on physical measurements, such as the height or weight of an individual,
the indications as to what should be done and what measurement scales should be used
are obvious. In educational research, however, the measurement of statistical variables
such as attitude score, verbal reasoning scores and achievement scores is not so obvious.
When designing a study a researcher has to operationalize abstract ideas or concepts of
interest. For example, a consideration of the concept of self-image includes defining the
term and how it will be observed or measured so that numerical values for the
constructed statistical variable self-image can be recorded. A student’s self-image cannot
be measured directly as can, for example, his or her height. Instead it is measured
indirectly by the researchers’ constructed variable. The extent to which constructed
variables do actually measure the concept of interest, which in this instance is that of self-
image, is referred to as the construct validity of the measure. This idea was referred to in
Chapter 1.
It is important to recognize that measurement, in this example, relates to an attribute of
the student which is abstract. Measurement in education is a means whereby abstract
attributes are quantified and used to describe, for example, the amount of self image,
ability, achievement, or understanding possessed by an individual. The development of
educational and psychological measures with construct validity presents a challenge to
educational researchers.
Whenever an attribute is quantified, that is when a number is used to quantify the
amount or type of an attribute (variable of interest), then the statistical variable so formed
is classified according to the following convention:
As a discerning parent what variables might you consider important in
choosing a new secondary school for your daughter?The following variables might be included in a check list when visiting the school:
- Distance from home (kilometres)
- Whether sibling attends the school (yes/no)
- State of school buildings (general appearance of buildings is excellent, average, poor)
- Position of school in Education Authority league tables of school examination results
(e.g., 1, 33, 56) - Denomination of school (e.g., Roman Catholic, Jewish, Church of England)
- Average turn-over time for staff (time staff remain in post in months)
Statistical analysis for education and psychology researchers 18