Monday, February 13, 2012

Validity and Reliability

1. Why is it important to establish the “goodness” of measures and how is this done?

The goodness of measures indicates whether the scale we use are reliable and valid. In other words, they indicate to what extent we are accurately and consistently measuring the concept (reliability), and whether we are indeed measuring whatever we are supposed to measure (validity). If we are not measuring the intended concept accurately, then, our study will not yield the correct results.

Reliability is established by assessing the stability of the measure through test-retest reliability and parallel form reliability, and internal consistency of the measure through Cronbach’s alpha.

Validity is established through face validity, content validity, criterion-related validity (concurrent and predictive), and  construct  validity (convergent and discriminant).

2. “Whenever possible, it is advisable to use instruments that have already been developed and repeatedly used in published studies, rather than develop our own instruments for our studies". Do you agree? Discuss the reasons for your answer.

It is always advisable to use well-validated instruments that have been repeatedly used in published studies, rather than going through the cumbersome process of developing one's own instruments and laboriously establishing their validity and reliability. However, simply because an instrument is used in some published studies does not necessarily make it a good instrument. It is always wise to examine the psychometric properties established for the instrument before using it. In case this is not easily available, a pilot study can be conducted to see if the measures have at least acceptable reliability.

3. "A valid instrument is always reliable, but a reliable instrument may not always be valid". Comment on this statement.

If an instrument is valid, especially in terms of content validity, we would expect to have a reliable instrument. However, an instrument can have both consistency and stability, but may not be measuring what it is intended to measure -- i.e.  have low validity. However, if an instrument does measure the concept that it is supposed to measure, it will be reliable. Thus, a valid instrument will be reliable, but a reliable instrument need not necessarily be valid.

No comments:

Post a Comment