HATS Research and Validation ...

Harrison Assessment Talent Solution
Get Deep Insight Into Your Staff
And Have 20 Years of Research to Back You

Dr. Dan Harrison talks on the Validity and Research of HATS

Listen to the audio of a Phone Conference with Dan Harrison discussing validation.

Sign up below to get updates and information in regards to the validity of Harrison Assessment Talent Solution (HATS)

Job Behaviour Validation

Personality validations started almost 50 years ago with tools like Myers Briggs.  When assessing personality you have no other mark - in other words you are looking for personality itself and the means of validating that is by looking at other measures of personality and then comparing them.

Whereas if you are looking at validating job behavior or looking at an instrument that is used for job selection or employee development it’s a very different process because primarily what you are looking for there is that each job that you use it for has it’s own set of validity.  In other words validity does not so much go with the instrument: it goes with the job.

Validation in regards to the Legal System

In the US, the government and the legal system there does not recognize any validity for an instrument itself.  However, what they look at is the validity of the instrument in relationship to that particular job. 

So in a sense it’s a misconception to think that the instrument carries validity itself.  Actually, the validity is in relationship to specific jobs. Having said that there is validity or there is sort of a potential for validity for instruments.  But that potential, for example, in the US, is based on construct validity.  And that means that if the instrument carries traits or carries measures that relate to specific job functions then that has face validity. Thus, that in itself has face validity in the US legal system and there is actually no need to go further than that to back your-self up. 

But as far as statistical validity goes, the US legal system does not recognize any statistical comparisons except as related to the specific job. This makes sense because in fact each time you calibrate the instrument to use it for a different job you are in essence using a different assessment.

So that is the primary thing to get a hold of in terms of the legal system in relation to the validity of the assessment. So when you are talking about job behavior assessment you are actually talking about specific jobs.  You are not talking about the global validity of an instrument, but some instruments are much better at having this construct validation. So for example, HATS uses traits that directly relate to job concepts, and that directly relate to job issues.  Therefore, it carries its own validity in terms of the legal system.

Whereas if you use personality instruments that have much more vague concepts to them and that are not necessarily job related then you have a serious problem, because literally it is impossible to statistically validate every job whatever instrument that you are using and therefore you actually have a legal problem. 

With Harrison Assessment Talent Solution (HATS), you don’t have that problem and that’s another reason the instrument is constructed the way it’s constructed.  The traits directly relate to job aspects and job functions and you can easily draw that face validity from it.

Listen to a 5 mins extract from the above audio in regards to legally defending an assessment instrument:-

The Research behind Validation and the Varibles Involved

In regards to specific research and specific quantities, if you are researching personalities, you can get thousands of people because there is no job related to it so you will just profile everybody you can get your hands on and then you start running validity to it.  But if you are actually talking about job validity that’s a different thing altogether since you will have to have a sample of people that are doing the exact same job. 

And if you do your homework well you have to actually narrow that sample down considerably.  So for example it just would not work to compare sales people who have been doing a particular job for 10 years with those who have been doing the same job for 1 year because you have a whole learning factor in that. And, let’s also say these sales people are from different territories with different degrees of  saturation. So you have to have control for a whole series of variables. Therefore, if people say “your sample is small”, well yes, we would much rather have a small sample that’s a good sample that is controlled for a whole series of variables than a large sample that does not.

As you may know, it is extremely difficult to get good performance criteria for large numbers of people, but we still do that. We have samples of 1 to 300 people which is a reasonable sample.  However, if you are talking about thousands and thousands of people, then that's really difficult to get. I would basically say that anybody who’s saying they have thousands and thousands is not controlling for those variables.

If you want to receive updates and other information in regards to assessment validation, please do sign up to our newsletter above.