Some concepts or variables are known or have been determined to be a good measure of other variable or attributes. These variables (concepts) otherwise known as criteria, can be used as basis for judging the validity of a newly designed study instrument.
There are two types of criterion – related validity. These are:
a) Predictive Validity – which measures how the newly designed measuring instrument adequately predicts certain attributes, it is
262
designed to measure. The emphasis here is more on the predictive ability of the instrument than on what the instrument measures.
b) Concurrent Validity - which measures other current behaviours or variables. It attempts to answer the question; does the instrument adequately and correctly measure the existence of some current conditions or criteria?
To determine criterion – related validity, the researcher must be able to determine the appropriate criteria that are being measured or against which the instruments outcome is to be compared. For example, company X has over the years come to associate the performance success of their managers with certain criteria, such as the type of university the employee attended and his ability to communicate fluently in English language. These two criteria have been measured over the years for every newly employed manager.
3.5.2 Reliability of Measuring Instruments
Reliability is the accuracy or precision of a measuring instrument. It is the consistency between different researchers under the same assumption and conditions.
A researcher who designs a measuring instrument should be concerned about how consistent the results he obtains with the instrument are. He will want to ensure that the instrument gives similar, close or the same result, if the study to which the instrument is applied is replicated. In other words, will the instrument give the same or similar results when it is used by different researchers under the same assumptions and conditions?
Test of Reliability
There are four ways a researcher can possibly test for reliability, these are:
i. Test – re – Test Reliability
In test – re – test reliability, the same measuring instrument is used to take two separate measurements on the same population at different times. The higher the correlation between the two measurements, the higher the reliability of the measuring instruments.
The problem with test – re – test reliability is that the first measurements may sensitise the respondents and make them appear to be consistent.
This may lead to some spurious reliability.
ii. Multiple (Alternate) Forms
Multiple forms reliability test is similar to the test of construct validity.
It attempts to test for reliability through the use of the same measuring instrument administered on different dimensions of the same variable.
These are a measure of the equivalence of the forms, and a high association among the forms a high reliability of the instrument.
Low association between the forms simply shows that the forms are not equivalent and may indicate how reliability.
iii. Split – Half Technique
In split – half technique, the assumption is that the measurement items can be randomly assigned to two equal parts. That is, the measurement items can be randomly split in half and each half is now treated as an alternative form of the same measurement. The reliability co-efficient formula used to calculate the reliability reads.
rtt 1 ve vt
Where rtt = the reliability coefficient
Ve = total expected variance between the measures
Vt = total variance between measures of the forms.
iv. Cranach’s Alpha Test
This test proceeds by associating each measurement items with every other measurement item and obtaining the average inter-correlation for all the paired associations. The computation is usually very long and tedious.
SELF-ASSESSMENT EXERCISE
Describe four scales for measuring data properties?
4.0 CONCLUSION
In this unit, we have discussed measurement and scaling in empirical research. We measure things casually every day. Research measurement is more precise and controlled. In measurement, one settles for measuring properties of the objects rather than the object themselves.
An event is measured in terms of its duration. What happened during it, who was involved, where it occurred, and so forth, are all properties of
264
the event. To be more precise, what are measured are indicants of the properties. Thus, for duration, one measures the number of hours and minutes recorded. For what happened, one uses some system to classify the types of activity that occurred. Measurement typically uses some sort of scale to classify or quantify the data collected.
5.0 SUMMARY
Measurement in research consists of assigning numbers to empirical events in compliance with asset of rules. Researchers measure indicants of the properties of objects using scales. Scaling describes the procedures by which we assign numbers to measurements of opinions, attitudes and other concepts. There are four scales types. In increasing order of power, these are nominal, ordinal, interval and ratio. Nominal scales classify without indicating order, distance or unique origin.
Ordinal data show magnitude relationships of more than and less than but have no-distance or unique origin. Interval scales have both order and distance but no unique origin. Ratio scales possess all of these features.
A sound measurement must meet the tests of validity, reliability and practicality. Validity reveals the degree to which an instrument measures what it is supposed to measure to assist the researcher in solving the research problem. Three forms of validity are used to evaluate measurement scales. Content validity exists to the degree that a measure provides an adequate reflection of the topic under study. Criterion- related validity relates to our ability to predict some outcome or estimate the existence of some current condition. Construct validity is the most complex and abstract. A measure has construct validity to the degree that it conforms to predicted correlations of other theoretical propositions. A measure is reliable if it provides consistent results.
Reliability is a partial contributor to validity, but a measurement tool may be reliable without being valid. Three forms of reliability are stability, equivalence and internal consistency. A measure has practical value for the research if it is economical, convenient and interpretable.
6.0 TUTOR-MARKED ASSIGNMENT
1. What do you understand by the terms “measurement” and
“scaling”?
2. What are the characteristics of a good measurement scale?
7.0 REFERENCES/FURTHER READING
Blumberg, B., Cooper, D.R. & Schindler, P.S. (2005). Business Research Methods. New York: McGraw-Hill Companies Inc.
Eneanya, A.N. (2012). Research Methods in Political Science and Public Administration. Lagos: University of Lagos Press Ltd.
Philips, B.S. (1971). Social Research Strategy and Tactics (2nd ed.).
New York: Macmillan.
266