Measuring Reliability

The Myers-Briggs Type Indicator (MBTI) is one of the most popular personality tests in the world today.  However, data reviewed from more than twenty MBTI research studies has indicated that the test has inadequate validity.  In other words, MBTI does not adequately measure what it is meant to measure.  Moreover, MBTI has reported reliability coefficients for its four scales on general population samples in the ranges from 0.61 to 0.87.  Experts on psychometric testing reveal that reliability coefficients for short personality tests should actually be in the range of 0.7 to 0.8 (Psychometric Success).
Does this mean that MBTI is unreliable? – No.  While this important personality test sometimes reveals a lower reliability coefficient than the other personality tests, even when its reliability coefficient is close to 0.61, the test is said to give us a dependable measure of the constructs it measures at least 61% of the time.  To put it another way, at least 61% of the people who use the MBTI once would see that their scores on the different constructs of the test would not change on a second testing session and all subsequent testing sessions.
Reliability is also known as “repeatability” or “consistency.”  A test with perfect reliability has a coefficient of 1.0.  A test with no reliability whatsoever has a coefficient of 0.  A test with perfect reliability is one which gives us absolutely consistent results with the same sample.  The higher the reliability, the better is a test, scientifically, given that we can trust the results of this test with confidence, knowing that they will not change drastically through retesting (Trochim, 2006).

With MBTI, we can be confident that the test would show us a true picture of an individual’s personality at least 61% of the time.  Although this value is better than 50%, MBTI would have given us greater confidence in terms of its measurement if its reliability coefficient always fell in the range of 0.8 to 0.9.  This would have led us to believe that MBTI is most definitely a trustworthy measure of an individual’s personality type.
References
1.Psychometric Success. “Myers-Briggs Widely Used But Still Controversial.” Available at http://www.psychometric-success.com/index.htm. (31 January 2007).
 
2. Trochim, William. (2006). “Theory of Reliability.” Research Methods Knowledge Base. Available at http://www.socialresearchmethods.net/kb/reliablt.php. (31 January 2007).