Chapter 9 – Making the Selection Decision
need to have an opinion and rationale for it
After studying this chapter, you should be able to:
• Discuss different ways of combining predictors
• Create a process for establishing cutoff scores
• Discuss the steps to take once the selection decision has been made
1. Combining Predictors
• Subjective approaches use human judgment, inference, or intuition when
combining predictor scores to make a final selection decision.
• Objective approaches use statistical or mechanical means and equations to
combine predictor scores to arrive at a final selection decision.
• Don’t ignore gut feelings but look at numbers
• Allows the power trippers to be put in place with numbers
• Practical reason – hires will be better, legally better position to defend, lets
people with all the power not calling all the shots
Major Approaches to Combine Predictor Scores
To select – one part of selection process, collecting that subjectively
other part, everyone has same score, relates back to job analyses, test job related etc –
come out with a score (objective)
Four Major Approaches to Combine Predictors
• Pure judgmental approach: Predictors are selected and combined exclusively on
the judgments of one or more decision makers.
• Not a great approach • Trait ratings approach: Judgmental data, such as interview scores, are entered
into a mathematical formula to calculate the total score for each applicant.
• Profile interpretation approach: Objective data about applicants are collected
but interpretation and combining is done subjectively.
• Probably fall in this one
• What does a great performer look like
• Statistical approach: Objective data are collected and combined using
mathematical formulae or other objective approaches.
• Very high validity
• Won’t get people to buy in, hard to convince higher authority
• Will probably get the best people though
Choosing an Approach
(the more you can move people towards objective data – the better, it just has to be
perceived, not even actual objective data)
• Research indicates that statistical and mechanical composite approaches are either
equal or superior to judgmental approaches to combining data.
• Reasons include:
– Limits on cognitive capacity of humans (decision makers don’t have all
the cognitive ability to process the information – people get overloaded)
– Human tendency to ignore sample size and quality before jumping to
– Human susceptibility to illusory correlation** (great communication skills
automatically mean great worker – bias)
– Personal biases
How to assign weights to various predictors?
• Multiple regression: Use regression formula to combine predictors and arrive at
a total score for each applicant.
• Graph* predictor (test score 0100 (x axis)) and performance (y axis 1
• If positively correlated going to be line from bottom left to upper right
• If you know equation for the line you can predict performance at different
• Things to be evaluated on in example, Test Score/Interview/Job
Simulation – assign weightings to each
• Higher the score, the higher the future projected job performance
• Very effective way to focus the team and objectify
• Multiple cutoff approach: Establish cutoff score for each predictor and reject
applicants who score below the cutoff on any single predictor.
• One predictor is so fundamental to performance, if you can’t do that
function that is absolutely critical you are no good for the job (minimal
level), doesn’t matter how they scored on the other ones • Examples on exam – if you have a job that have a few things that are
critical that they have to do a minimal level what method would they
choose? Multiple Cut Off
• If you have a few qualities that can be compensated by great score in
another – you do multiple regression
• Everyone does every form of selection process
• Multiple hurdle approach: Applicants must pass each predictor in sequence.
Unless the mi