The Selection Crystal Ball

Picking up the thread from my last post on recruitment, let us assume that you have the perfect job description, the inclusive advert in various advertising mediums, and your diverse selection of panellists. You are now left with the task of figuring out who amongst those who have expressed an interest in the role would be appointed to the job. In modern-day work, we predict the success of an individual in a particular position based on some form of assessment. This is usually in the form of a test, an interview or other forms of assessment. ​There are various types of tests available for use. Some are noted below:  

• Computer adaptive test- this can be used to estimate an applicant's actual ability as it uses a technology that provides candidates with questions of increasing difficulty. This is regularly used in the armed forces.  
• Speed versus power test- a speed test is one in which you give a candidate a series of tasks and candidates are expected to complete as many as possible within a short timeframe. You can use this in administrative roles or in data entry roles. The power test has no time limit, but candidates are expected to complete complex tasks.  
• Performance test- This relates to activities with manual dexterity. Popularly used in warehouse operations machinery and the medical field (surgeons).  
• Psychomotor tests- measures speed and accuracy of motor and sensory coordination. For example, used for fighter pilots.  

In addition to actual tests, we could look at a candidates' cognitive ability through general cognitive ability tests or specific cognitive ability tests. Both test intelligence. The general cognitive test is often used to select roles requiring complex decision-making. Specific cognitive ability test follows a similar principle but tests intelligence in a particular job. For example, clerical ability is tested by focusing on the speed and accuracy of verbal and numerical data.  

Psychometric testing has also become a popular form of selection. Quite a few psychometric assessments are available, and I won't delve into this. A note of caution, the result of a psychometric should not be used as the sole basis of an appointment. Research indicates that candidates can answer these assessments in ways that make them desirable candidates. Psychometrics should be used as at least another form of assessment.  
We could also use work samples or assessment centres to assess the candidates. Work samples are when you give actual life situations to the candidate to solve. Assessment centres can be an in-basket exercise or a leaderless group discussion.  

In choosing our predictors of success, we have to be very careful about what we rely on. I recently read some research done on selection predictions, and it tells a cautionary tale (Levy., 2020). The work cited here shows the validity coefficient. The validity coefficient shows the strength of between a predictor (assessment test) and an outcome (performance on the job.) 

Figures to note. 

• General cognitive ability is a valid predictor of performance by 48% 
• Specific cognitive ability is a valid predictor of performance by 40- 50% 
• A personality test is a valid predictor of performance by 21% - 37% 
• The work sample is a valid predictor of performance by 39% 
• An assessment centre is a valid predictor of performance by 28% in a 2007 study and 40% in a 2008 study 
• Structured interviews are a valid predictor of performance by 71% 
• Unstructured interviews are a valid predictor of performance by 20% 

​All studies will have their limitations. However, despite the limitations in the various studies, the research tells a story. We cannot rely only on one assessment method to assess candidates. We need to use at least a mix of two or three, but we must be clear on what skills, knowledge, or behaviour is being tested with any available mediums. In my next post, I will discuss how you could use these predictors to select the 'one' 

---