Issues of measurement
How are you going to measure whatever it is you need to measure? To some extent we have discussed this earlier: when you know what it is you want to assess you need to find a valid and reliable tool to do the job. If you are measuring blood pressure or wound infections or various aspects of the immune system, this can be easy, but if you are considering opinions, feelings, interactions or communication, you may need to hunt for or develop instruments. Typical instruments used in educational research are interview or focus group schedules, psychometric tests, questionnaires, observation schedules and standardised rating sheets. Never design your own instrument when a perfectly good one already exists – the development of measures is a time-consuming and costly job if it is to be done properly with all the issues of reliability, validity and piloting taken care of. Doing a search on the topic usually reveals a number of good instruments that can be used with or without modification.
Instruments are reliable (replicable) to the extent to which using them on the same individual at different times or in different circumstances still produces a relatively similar result. Instruments are valid when they measure what they set out to measure. With face validity the measure should simply make sense to individuals as a reasonable way to assess what it is you want to measure, for example an end of programme survey questionnaire looking at satisfaction. Criterion validity assesses the extent to which the measure correlates with other existing measures – for example, that a new measure of educational quality has a high correlation with another measure for the same purpose. Construct validity measures whether an instrument relates in a theoretically predictable way to another established variable, for example that quality of teaching relates to teachers’ years of experience, or that quality of the learning experience relates to the provision of learner support.