CHAPTER 1
Old and new selection methods
We've always done it this way

WHY SELECTION MATTERS

Clark Hull is better known, to psychologists at least, as an animal learning theorist, but very early in his career he wrote a book on aptitude testing (Hull, 1928), and described ratios of output of best to worst performers in a variety of occupations. Hull was the first psychologist to ask how much workers differ in productivity, and he discovered the principle that should be written in letters of fire on every HR manager's office wall: the best is twice as good as the worst.

Human resource managers sometimes find they have difficulty convincing colleagues that HR departments also make a major contribution to the organization's success. Because HR departments are neither making things, nor selling things, some colleagues think they do not add any value to the organization. This represents a very narrow approach to how organizations work, which overlooks the fact that an organization's most important asset is its staff. Psychologists have devised techniques for showing how finding and keeping the right staff adds value to the organization. Rational Estimate technique (described in detail in Chapter 14) estimates how much workers doing the same job vary in the value of their contribution. One 'rule of thumb' this research generated states that The value of a good employee minus the value of a poor employee is roughly equal to the salary paid for the job. If the salary for the job in question is £50,000, then a good employee, in the top 15%, is worth £50,000 more each year than one in the bottom 15%. Differences in value of the order of £50,000 per employee mount up across an organization. Hunter and Hunter (1984) generated a couple of examples, for the public sector in the USA.

• A small employer, the Philadelphia police force (5,000 employees), could save $18 million a year by using psychological tests to select the best.
• A large employer, the US Federal Government (4 million employees), could save $16 billion a year. Or, to reverse the perspective, the US Federal Government was losing $16 billion a year, at 1980s prices, by not using tests.

Some critics see a flaw in such calculations. Every company in the country cannot employ the best, for example, computer programmers; someone has to employ the rest. Good selection cannot increase national productivity, only the productivity of employers that use good selection methods to grab more than their fair share of talent. At present, employers are - largely - free to do precisely that. The rest of this book explains how.

RECRUITMENT

Traditional methods

Figure 1.1 summarizes the successive stages of recruiting and selecting an academic for a British university. The advertisement attracts applicants, who complete and return an application form. Some applicants' references are taken up; the rest are excluded from further consideration. Applicants (As) with satisfactory references are shortlisted, and invited for interview, after which the post is filled. The employer tries to attract as many As as possible, then pass them through a series of filters, until the number of surviving As equals the number of vacancies.

Figure 1.1 Successive stages in selecting academic staff in a British university.

Recruitment sources

There are many ways employers can try to attract applicants: advertisements, agencies - public or private, word of mouth, 'walk-ins' (people who come in and ask if there are any vacancies), job fairs, and the Internet. Employers should analyse recruiting sources carefully to determine which find good employees who stay with them. Employers also need to check whether their recruitment methods are finding a representative applicant pool, in terms of gender, ethnicity, and disability. Newman and Lyon (2009) investigate targeted recruiting, through the wording of advertisements for job. They suggest that saying the organization is 'results oriented' will tend to attract more As high in conscientiousness, and saying the organization is 'innovative' will attract more As high in mental ability. Later chapters will describe research showing As high in conscientiousness and mental ability tend to make better employees. Newman and Lyon suggest the right advertisement can attract such applicants, both overall and from minorities, so meeting the twin aims of many employers: good employees and a representative workforce.

Realistic job previews

Any organization can paint a rosy picture of what is really a boring and unpleasant job because they fear no one would apply otherwise. In the USA realistic job previews (RJPs) are widely used to tell applicants what being, for example, a call-centre worker is really like: fast-paced, closely supervised, routine to the point of being boring. Earnest, Allen and Landis's (2011) analysis confirms the results of several earlier reviews that there is a very modest link with reduced turnover, suggesting RJPs may be worth using, given that RJPs cost employers very little whereas turnover costs them a lot. Earnest et al. suggest RJPs work by making As see the employer as more honest.

Informal recruitment

Applicants are sometimes recruited by word of mouth, usually through existing employees. Besides being cheaper, the grapevine finds employees who stay longer (low turnover) possibly because they have a clearer idea what the job really involves. Zottoli and Wanous (2000) report informal recruits on average do slightly better work; the difference is small (d = 0.08; page 31) but is achieved very cheaply. However, fair employment agencies, for example the (British) Equality and Human Rights Commission, generally frown on informal recruitment; they argue recruiting an all-white workforce's friends is unfair because it tends to perpetuate an all-white workforce. Weller et al. (2009) report data from the German Socio-Economic Panel Study, nearly 3,000 people, representative of the whole German working population, tracked over five years. Weller et al. confirm that informal recruitment results in lower turnover: more employees recruited through agencies or advertisement leave in the first two years, and leave sooner, departures peaking at 9 months compared with 17 months for employees recruited through informal contacts.

New technology and recruitment

Advertising, making applications, sifting applications and even assessment can now be carried out electronically, which can make the whole process far quicker. People even talk of making 'same-day offers'. More and more jobs are advertised on the Internet, through the employer's own website or through numerous recruitment sites. People seeking jobs can post their details on websites for potential employers to evaluate, which gives the job seeker an opportunity that did not exist before. Most employers now use electronic application systems, eliminating the conventional paper application form. Internet recruitment can greatly increase the number of As, which is good for the employer if it broadens the field of high-calibre As, but it does also create work sorting through a mountain of applications.

APPLICATION SIFTING

The role of the application form (AF), or its new technology equivalent, is to act as first filter, choosing a relatively small number of applications to process further, called sifting. Sifting can take up a lot of time in HR departments so any way of speeding it up will be very valuable, so long as it is fair and accurate. Research suggests sifting is not always done very effectively. Machwirth, Schuler and Moser (1996) used policy capturing analysis, which works back from the decisions HR make about a set of applications to infer how HR decides. Machwirth et al. showed what HR do often differs from what they say. Managers said they sifted on the basis of proven ability and previously achieved position, but in practice rejected applicants because the application looked untidy or badly written. McKinney et al. (2003) analysed how US campus recruiters used grade point average (GPA; college course marks) to select for interview. Some chose students with high marks, which is the logical use of the information, given that GPA does predict work performance to some extent, and that it is linked to mental ability, which also predicts work performance. A second large group ignored GPA altogether. A third group selected for lower GPA, screening out any As with high grades, which does not seem a good way to sift, given the link between work performance and mental ability. The choice of strategy seemed essentially idiosyncratic, and not linked to type of job or employer.

Accuracy and honesty

Numerous surveys report that alarming percentages of AFs, résumés and CVs contain information that is inaccurate, or even false. These surveys often seem to have a 'self-serving' element, being reported by organizations that offer to verify information supplied by As; not much independent research has been reported. Goldstein (1971) found many applicants for nursing vacancies exaggerated both previous experience and salary. More seriously, a quarter gave a reason for leaving that...