The human assessment has become a well-known and widely accepted activity with several leading assessments claiming a majority of the U.S. business market, since the 1940’s and ’50’s. Many years ago, these leading assessments started doing repeat score reliability studies to separate the non-viable human assessments from the viable ones.
This test/retest benchmarking has served to help people in business know which assessments are legitimate enough to bring into the Human Resources world for use to develop leadership. Recently, however, retesting has been conducted at increasingly short intervals after the initial test.
Does Timing Matter?
The validation reports generated 10-20 years ago demonstrated a repeat score reliability in the low 70% range when retested within 90 days of the first assessment process. The industry accepted a 70% retest level as the line of viability.
Any assessment which had lower than 70% repeat score reliability was judged to be non-viable. Some tests that achieved a score just above the minimum level saw their repeat score reliability fall markedly below 70%, when they retested people a year later or five years later.
The second level of validation began being important when too many instruments surpassed the 70% 90-day reliability hurdle. The viable assessment companies began doing construct evaluations between instruments. Several questions became standard construct questions.
1. Does our assessment identify the same personality characteristics as other 70% reliable instruments?
2. Does our assessment identify the same characteristics in retest individuals as it did in the first test?
It can be argued that the only real measure of validity is repeat score reliability over time. This is why repeat score reliability was chosen as the measure of assessment viability in the first place.
The Core Values Index™ of Taylor Protocols has just completed its third, third-party repeat score validation process with CVI scores going back to 2002, including people who have taken the CVI up to five times or more, with at least 90 days between testing and sometimes as much as ten years between repeated CVI completions.
We have published the complete results that were derived by Seattle Research Partners in this study, without change or subtraction of data. The CVI has been found to be approximately 97.7% repeat score reliable year over year and decade over decade.
If it is possible to attain such a high repeat score reliability, why is the standard accepted as 70%? Companies wishing to develop their human resources potential would benefit from a higher standard, giving them more reliable test results.
Currently there is no clear way to choose one assessment instrument over another except for repeat score reliability. Reading through test/retest validation reports reveals that the increased reliability seen today comes mostly from retests completed in less than 90 days. Reliability scores from retests where more than 1-3 years have lapsed continue to fall off below the 70% validity hurdle. Is our industry ignoring longer duration studies and accepting claims of validity based only upon test/retest validation that occurs in less than 90 days?
What do you think? Is the 90-day 70%-reliability standard high enough to inform and advise the users of assessment instruments?
Article Source: Taylor Protocols