Reliability
Some researchers have interpreted the reliability of the test as being low. Studies have found that between 39% and 76% of those tested fall into different types upon retesting some weeks or years later.
One study reports that the MBTI dichotomies exhibit good split-half reliability; however, the dichotomy scores are distributed in a bell curve, and the overall type allocations are less reliable. Also, test-retest reliability is sensitive to the time between tests. Within each dichotomy scale, as measured on Form G, about 83% of categorizations remain the same when individuals are retested within nine months, and around 75% when individuals are retested after nine months. About 50% of people tested within nine months remain the same overall type, and 36% remain the same type after more than nine months. For Form M (the most current form of the MBTI instrument), the MBTI Manual reports that these scores are higher (p. 163, Table 8.6).
In one study, when people were asked to compare their preferred type to that assigned by the MBTI assessment, only half of people picked the same profile. Critics also argue that the MBTI lacks falsifiability, which can cause confirmation bias in the interpretation of results.
Read more about this topic: Myers-Briggs Type Indicator