Picking the Right Assessment Tools

A table to help determine which tools are best for you.

September 18, 2003
With so many different types of assessment tools available, deciding whichones to use can be confusing. Is it better to use background checks, talentmeasures, or structured interviews? Are drug screens better than knowledge testsor job simulations? In general, the answer is "it depends."

    There is no single "best" type of assessment tool. Assuming they areappropriately designed and administered, all of the major types of assessmenttools can effectively predict future candidate behavior. Appropriate use of anywell-designed assessment tool will substantially improve a company’s bottomline. The challenge is to determine which assessment tools will have thegreatest impact, given your company’s particular needs. Here are some generalguidelines to help you figure out which assessment tool is best for you.

Understanding what you’re trying to predict
    The first step in choosing an assessment tool is to determine what yourspecific needs are. This means understanding and clearly outlining what kind ofperformance you want to predict.

    For instance, if theft is the most important issue affecting organizationalperformance, then you should look at assessment tools specifically designed toidentify candidates who are likely to steal. These might include backgroundchecks or integrity tests. If customer service has the greatest impact onrevenue, then consider job simulations designed to assess candidates’ abilityto handle difficult customer interactions or talent measures that assesscandidates’ interpersonal style.

    This table summarizes the effectiveness of various assessment tools forpredicting different kinds of job performance. It can be used as a general guidein trying to determine what types of assessment tools to consider.

Qualification Screens Structured Interviews Job Simulations Knowledge Tests Talent Measures Background Checks Drug Screens Physical Tests Integrity Tests Cultural Fitness
Attendance/ Reliability         x       x  


      x       x


Teamwork   x x   x          
Customer   Service    x x   x        
Learning and Adaptability     x   x          
Deviant  Behavior (theft, drug use, violence)         x x x   x  
Physical Performance               x    
Problem    Solving    x x   x          
Leadership    x x   x          
Motivation/ Work Ethic   x x   x       x  
Specialist Knowledge  x      x            

The importance of validation
    Whatever type of assessment tool you are considering, it is important that itbe validated. Validation is a process for testing and documenting theeffectiveness of an assessment tool in predicting job performance. Any vendoryou are considering should be able to provide detailed "validation" datademonstrating the effectiveness of its assessment instruments in predictingdifferent types of job performance.

    Validity data tends to be presented in two forms: percentages and validitycoefficients. Percentages show increases in the number of successful hires thatare associated with use of a specific assessment tool. For example, Unicruconducted research showing that employees hired with their assessment tools hadturnover rates that were 25 percent lower than those of employees hired withoutuse of these tools. Percentages are easy to interpret, but they demonstrate onlythe effectiveness of tools that predict highly objective kinds of performancesuch as turnover or sales volume.

    Validity coefficients show the effectiveness of assessment tools forpredicting less objective forms of performance such as customer service,problem-solving, or teamwork. Validity coefficients are an index of anassessment’s ability to predict performance. They range from 0, whichindicates no relationship between an assessment tool and job performance, to1.0, which indicates 100 percent accuracy.

    The table below provides some general guidelines for interpreting the meaningof different validity coefficients. The validity coefficients of well-designedassessment tools usually range from .15 to .60, with most falling between .20and .40. Beware of anyone claiming validity coefficients above .70, as thislevel is unrealistic for practical purposes.

Validity Coefficient Range Effectiveness of Assessment Tool
Below .10 Not useful, might as well flip a coin
.11 to .20 Possibly useful, depending on the circumstances.
.21 to .30  Useful
Above .30 Very useful

The advantages of using several tools
    It is important to remember that job success almost always depends on severaldifferent kinds of performance. For example, an employee who is very good atcustomer service will be of little value if he/she frequently misses work. Youcan increase the effectiveness of a selection system by combining assessmenttools that predict different types of job performance. For instance, combining apersonality test with an ability test can increase the overall predictiveability of your selection system by 20 percent or more. Adding one or twoadditional assessment tools can often significantly increase the ROI of yourstaffing processes.

    It is important to choose assessment tools that predict the various kinds ofperformance that are most critical to the job. As a general rule, it is good tofocus on having assessment tools that predict both what candidates "can do"and what they "will do." This usually means having a mix of assessment toolsthat measure "hard skills" such as knowledge and problem-solving ability aswell as "soft skills" such as reliability and customer service.

    Of course, as long as it predicts a key area of performance, there is nothingwrong with using just one kind of assessment tool. However, hiring managers andrecruiters should be careful not to overly emphasize the kinds of performancepredicted by the tool at the expense of other factors that influence success.

    For example, if a test of knowledge or ability is used as the only assessmenttool, hiring managers may decide to hire only candidates who score in the toprange. While this will increase the likelihood of hiring good problem-solvers,it could decrease the number of hires who are likely to show high levels of workethic and reliability.

Workforce Online, December 2002 -- Register Now!