Janice Chalmers asks the experts...

Are you one of the many recruiters undecided about the value of screening online? Here, two psychometric testing pros give their views on the value, and validity, of online tests.

Professor Adrian Furnham, Professor of Psychology, University College London

What are the advantages and drawbacks of online candidate screening?
The online nature makes for quick and efficient scoring. But there are various drawbacks that people need to be aware of, such as: whether you can realistically do ability tests online how to check the candidate as opposed to others are doing the tests, and whether people with little computer experience (eg older people) are disadvantaged.

What, in your view, is the best type of testing and why?
I prefer to watch people doing these tests. Also, online is better with preference rather than power tests (ie personality rather than ability). Whether on or offline, tests should be used to confirm hunches or hypotheses, or perhaps to raise issues like a person’s ability level or motivation. Tests are always best used in conjunction with other selection methods, not just interview data, but the sort of data that is obtained at an assessment centre.

How does online compare with offline testing, for example, is online as valid?
The data that exists suggests few differences. But there are a lot of caveats and, at this stage, not that much research. The greatest problem with online testing of someone in another location is who is doing the test and how.

What is your view of the trend for recruiters to use their own tailored, ‘home-made’ tests?
The major issue is always the same - validity. I have nothing against homemade tests, as long as they can be shown to have content, construct, concurrent, predictive, discriminant etc validity. It is very dangerous and unwise to make people decisions on test scores that have no known validity. In fact, it is not only unwise but ethically and possibly legally wrong in that they may easily be discriminatory.

Finally, what should recruiters look for in an online test?
Evidence that tests are valid first; that people can’t fake it; that online experience does not effect test results, and that candidates fell fairly dealt with.

Ian Newcombe, Director of Product Development, PSL

What are the advantages and drawbacks of online candidate screening?
Recruiters find it hard to maintain focus and objectivity when wading through large numbers of applicants. Online testing is by nature a consistent and objective process. So long as the tests are properly selected and calibrated, each applicant is treated fairly and human error is eliminated from the process.

However, there are risks as well. For example, there is a great temptation to attempt to use it to create a shortlist. But testing is typically most effective at eliminating weaker candidates - it cannot distinguish between the better candidates with sufficient precision to create a shortlist. It is tempting for employers just to pick the top scorers, but this is rarely appropriate or justifiable.

What, in your view, is the best type of testing and why?
The area where the biggest impact is being felt is in sifting. Where an employer receives very large numbers of applicants, factual information, such as the possession of a work permit or specific qualifications, can be used to eliminate a proportion of applicants. But different forms of tests should be used at the different points of the selection process - for example, after this initial sifting questionnaire, those candidates that progress to interview stage may be invited to attend an assessment centre, and more in-depth, supervised testing will take place then.

How does online compare with offline testing, for example, is online as valid?
The evidence strongly suggests that the same test delivered online and offline is equivalent. The main area where there is cause for concern is candidate cheating. Techniques such as item banking (where questions are drawn at random from a large pool) reduce the potential for cheating. However, since most online testing is unsupervised, there is always the risk that the applicant has asked for help when completing the test. Recruiters need to have a method of verifying that the test scores are an appropriate assessment of the candidate, and this is often done with supervised re-testing for those that pass to the face-to-face stage of selection.

What is your view of the trend for recruiters to use their own tailored, 'home-made' tests?
Developing tests that are effective and legally defensible is a very specialised area. Unfortunately, [building your own] could prove to be a false economy. Another consideration is how these tests will be delivered. Only a handful of test vendors have created the sophisticated software necessary to deliver and score online tests reliably.

Finally, what should recruiters look for in an online test?
Ask other recruiters what has worked for them, and where they have encountered problems.

Guru guidance - Ten online testing tips:

• Ensure the test does not discriminate unfairly against any candidate group
• Do not use online screening to create a shortlist
• Use test results as only one of several sources for selection decisions
• Use different tests at different stages of the selection process, eg online test at early application stage, ability / reasoning tests at interview stage
• Use fact-based tests or questionnaires to filter out candidates early on in the application process
• To minimise the 'cheat' factor, have candidates re-tested when they come for interview
• Check if your test vendor offers 'item banking' (automatic generation of random test questions) to avoid candidate over-familiarity with tests
• Make sure you're covered legally for any recruitment decisions you make based on your online tests

Whether online or off, use full psychometric tests to follow up hunches.