Biodata: this is your life

Biodata is a little-known, yet potentially high-powered selection tool; we investigate what it can offer and the difficulties that stand in the way of its wider use.

The use of biodata to predict job performance can be traced back to the 19th century, when it was first used by insurance actuaries to forecast productivity and labour turnover data for life insurance sales personnel.

Biodata - biographical data - is based on the premise that past behaviour represents the best predictor of future performance. In the employment context, it aims to measure relevant past experiences (in previous work or life) and behaviours in order to predict future job performance, training or length of service.

An example of this might be when looking to identify candidates with strong leadership potential. Candidates will be questioned on their experiences and behaviours in leadership roles, not just at work, but across other domains of life, including sport, community organisations, school and university.

According to Adrian Furnham, professor of psychology at University College London, business adviser, broadcaster and writer, part of the attractiveness of biodata as a tool is that "it is very reactive and people are not aware that when filling out an application form, they are filling out a selection instrument."

While it consistently scores highly as a means of selecting the right person for the job, biodata is used by just a minority of employers - despite the fact that it has been rated in research by occupational psychologists alongside assessment centres in terms of the quality of candidate selection¹.

Research studies, including our own surveys, find that so few organisations now use the biodata technique that it disappears off the radar. It is likely that fewer than 1% of employers use it. So, given that it is known to be potentially so effective, why are recruiters not using this technique in greater numbers?

Complex and time-consuming

The answer to this conundrum probably lies in the complexity of the process required to identify appropriate biodata, the volume of information involved, coupled with the fact that each job function may need a biodata analysis in its own right. According to Professor Furnham: "Biodata has a number of limitations: it is expensive, it takes time to validate, does not 'travel' very well, and it perpetuates the same sort of person, giving rise to cloning."

Another major problem with biodata, as far as Professor Furnham is concerned, is the fact that it is based on information about the past, with the result that it is backward-looking. He says that biodata predictors are developed through "replicating the past and so, if in 1995 certain characteristics predicted success and you live in a stable environment, then things are okay. The trouble is you're selecting for things that worked in the past, and things change."

But what is it?

A questionnaire that uses biodata scales can appear similar to a personality questionnaire, but with some important differences. While a personality questionnaire focuses on attitudes, preferences and self-perceptions, biodata usually asks for historical information about past experiences and behaviours. This ensures that candidate responses can be more easily verified and, thus, more reliable.

These questions usually appear in a multiple-choice format, which serves two purposes: it is easier for applicants to respond; and it is possible to apply objective scoring procedures in the assessment of individuals' responses.

Typically, the multiple-choice questions look at work attitudes, achievements in work settings, education, life experiences, vocational interests, team orientation and other job-relevant characteristics.

Effective biodata questionnaires or inventories must be developed as a result of extensive research, usually with employees in the same or similar roles. The sample group is then split into two: those identified as good performers, and those who have consistently underperformed or have been poor performers. In order to identify biodata predictors, or criteria that indicate success in the role, these groups will answer an identical list of questions, based on biographical data or job-related functions. The responses of both groups are compared to identify predictors, or criteria, relevant to job success.

Sample sizes

"You get better generalisability with bigger samples, especially if you get a lot of wobble in your data. It depends on many factors, but I would be very hesitant with anything less than 100 [individuals] or more," says Professor Furnham. A study to develop biodata will only produce information that can be reliably used for roles identical, or closely similar, to the jobs that were originally studied.

In addition, biodata is highly susceptible to change. "The trouble with biodata is that any change requires the process to be conducted all over again," says Professor Furnham.

He feels that there is also potential for problems to occur from the very start, when the sample of individuals is being identified. "You go into an organisation and ask them for their good, poor and average people. The trouble is with regard to the criteria used to define who are the 'good', 'poor' and 'average' people. That is the same of all tests, but some organisations have good, robust data and many of them don't."

However, while this is the case for most job roles, it seems that "generalisability" can be improved when biodata instruments are used for supervisors and managers. Given the high level of transferable skills within these two positions, this is perhaps not surprising. Research conducted on instruments for both supervisors² and managers³ indicates that biodata developed for both these roles can have a longer "shelf life".

Predictors of success

Once the samples have been examined and the relevant data obtained, the potential answers to the multiple-choice questions based on the data can be weighted. Points can be assigned to answers and a cut-off score agreed. This scoring process means that organisations can set a score below which individuals will not progress any further in the selection process.

Professor Furnham explains that the essence of using biodata to identify predictors relies on the assumption that "we live in a stable environment where the predictors are always the same".

"Now, to some extent that is true," he adds. "We know that intelligence is a very good predictor of success at work in all jobs, at all times. But there are other factors as well that might come into play at certain times, and they might not be picked up from biodata that was done in the past."

Moreover, times change, and what were once powerful backward-looking predictors of success may no longer be as relevant. For example, Professor Furnham cites obtaining a university degree, undertaking foreign travel or owning a home computer as factors that used to be effective biodata criteria in some cases.

But, with the increasing number of university students, foreign travellers and owners of home PCs, these factors have lost their power and have little or no relevance today.

Suitability

There are particular jobs and employment situations that are more suited to using biodata - not least, because of the extensive research required to develop robust data. Such jobs and situations include:

The legal context

With the tightening up of discrimination law, organisations need to ensure that their use of biodata is fully compliant. The Acas Recruitment and induction guidelines⁴ provide a good starting point. They point out that: "Use of biodata, like other tests, needs careful control to avoid any possibility of discrimination or invasion of privacy."

Acas encourages organisations to consider why the test is being used, if it is necessary for the requirements of the job, and what infrastructure is in place to store results and provide feedback to candidates. Employers are also required to store record or data sheets in accordance with data protection legislation.

Research⁵ has found that legal challenges to an employer's recruitment process are less likely if the candidate feels that its selection methods include four characteristics:

Professor Furnham considers that employers need not be discouraged from using biodata as a result of increased legislation, quoting its greater level of use in the USA "where awareness of legal constraints has always been higher".

He believes that the empirical nature of biodata can produce robust findings that are not discriminatory in nature. "If one can demonstrate that your biodata predictors are not in fact related to sex or race, or other areas, then there should be no problem whatsoever." But, he warns, gathering information on whether applicants "come from a big family as opposed to a small family, if that is correlated with Catholicism or class - which it used to be - then you're picking up something that you can get in trouble with."

Biodata is no less reliable or robust than any other selection method, in Professor Furnham's view. Like other selection processes, "there are things people shouldn't be looking at: sex, ethnicity, etc. But you can't legislate against people looking at employees' files, and drawing their own conclusions. What we know is that it is illegal to discriminate on the basis of certain variables."

He points out that if an organisation notices that there is a predominance of, for example, successful men in a certain role, then its selection process as a whole needs to be examined for objectivity. The employer also needs to ensure that equal numbers of men and women have the opportunity to apply for the positions in the first instance.

He goes on: "If a powerful predictor such as sex is determined, it is not very much use to you because you can't discriminate, at least not very subtly, on this basis. The question is: if females are better at it or the other way around, how do you explain this? What is the origin of this? Is it the result of some sort of prejudice in the first place or is there some other factor going on? There are lots of explanations why this may be the case. And you might learn something about your selection process from it."

The future

The increase in recent years in the use of online recruitment and selection may represent the future for employers' use of biodata. For example, some graduate recruitment sites right now ask prospective candidates to answer a qualifying question relating to the type and quality of degree they are studying for. If the degree subject is not deemed desirable or appropriate, then candidates are screened out at the beginning of the process, thus saving both the organisation and the individual valuable time and effort.

Professor Furnham, however, feels that selection methods have far more ambitious futures: "It will be genetic testing soon; it won't be a questionnaire. Candidates will just get a mouth swab or [recruiters will] give you some coffee and take your DNA from your cup." Just imagine.

1. "The validity and utility of selection methods in personnel psychology: practical and theoretical implications of 85 years of research findings", F Schmidt and J Hunter, Psychological Bulletin, 124, pp.262-274, 1998.

2. "Biographical data in employment selection: can validities be made generalisable?", H Rothstein et al, Journal of Applied Psychology, 1990.

3. "Generalisable biographical data validity can be achieved without multi-organisational development and keying", K Carlson et al, Personnel Psychology, 1999.

4. Recruitment and induction, Acas, www.acas.gov.uk.

5. "Personnel selection", Ivan Robertson and Mike Smith, Journal of Occupational and Organizational Psychology, 74, 2001, pp.441-472.

This article was written by Noelle Murphy, researcher and writer on IRS Employment Review, noelle.murphy@irseclipse.co.uk.