INTERNATIONAL DIFFERENCES IN SELECTION METHOD USE

Before turning to consider these differences, it is worth noting that there is clearly considerable variability within a particular national context in the selection methods used. For example, Di Milia, Smith and Brown (1994), looking at management selection in Australia, report differences between manufacturing, retail and government sectors compared to the business sector. The latter made significantly greater use of cognitive testing and assessment centres than the two former sectors. Gowing and Slivinski (1994) similarly demonstrate that different methods of selection are used for different types of employee within North America. Moreover, there is even variability in the way a particular method is used within a national context. For example, Lievens and Goemaere (1999) interviewed representatives from a sample of Dutch companies that used assessment centres and found a large variability in practice. 'Interpretative flexibility' (Bijker, Hughes & Pinch, 1987) is not constrained to national 'sense-making'. Thus, variability exists even within a particular national context. However, evidence suggests that there is even greater variability in the use of selection methods across countries.

There have been a number of studies and reviews of the research that have focused on national comparisons in selection practices. Much of this research has concentrated on only a limited number of countries and most of the work has been done on comparisons across Europe. This has demonstrated that the use of interviews and application forms is very common throughout companies in Europe (Shackleton & Newell, 1994). These methods are also very common in North America (Rowe, Williams & Day, 1994) and Australia (Di Milia, Smith & Brown, 1994). However, given the interpretative flexibility of selection method use, referred to above, it is clear that even though the same method may be common across many countries, it may actually be used in very different ways. So, for example, even though the interview is almost ubiquitous in selection, its style and format can be quite different. Thus, in a comparison of management selection in British, German, Italian, Belgian and French companies, Shackleton and Newell (1994) found:

• German companies were unlikely to use more than one interview while French companies virtually always used more than one interview.

• Italian companies rarely used panel interviews while they were much more common in Germany.

• In Italy a representative from the Personnel/HR department is very rarely present during the interview while in German and Belgian companies a representative is very typically present.

• German and Italian companies used one-to-one interviews less often than Belgian companies.

Comparing the results from this European study with surveys conducted in North America also suggests differences, particularly in terms of the use of structured interviews. For example, Gowing and Slivinski (1994) found that 82% of the respondents in their survey of North American organisations structured their oral interviews. This is a higher percentage than is typically found in surveys of European companies. In terms of other less common methods of selection, there is also variability in use. So for example, in terms of Assessment Centres, these will rarely include psychometric tests in Germany while these will be common in a British Assessment Centre (Mabey & Thompson, 1993; Shackleton & Newell, 1994).

Moreover, there are also some established differences in the actual methods that are used. Thus, for example, in a comparison of five European countries, Shackleton and Newell [1994) conclude: 'The regression analysis shows that it is country rather than company size, or other company variables such as number of managers recruited, which is the dominant influence on the selection methods used' (p. 101). Similarly, in one of the most recent studies, Ryan, McFarland, Baron and Page (1999), comparing the use of 11 different selection methods across 22 countries, concluded: 'National differences accounted for a substantial portion of the variance in the use of fixed interview questions, as well as sizeable portions of variance in using multiple methods of verification, testing and number of interviews. This confirms previous research indicating variability across nations in selection practices' (p. 371).

Some established differences include:

• While a significant minority of French companies (including French companies operating outside France) use graphology to assess the personality of potential recruits, this method is very rarely used elsewhere, at least in Europe (Shackleton & Newell, 1991; Clark, 1993; McCulloch, 1993; Levy-Leboyer, 1994; Ryan et al., 1999). Bruchon-Schweitser (1996) also found that astrology and graphology are still used by some French companies while these methods are almost unheard of outside France.

• Situational tests and assessment centres are used more often in the UK, Germany and the Netherlands than in France and Belgium (Shackleton & Newell, 1994; Levy-Leboyer, 1994; Ryan et al., 1999), while assessment centres are not used at all in Spain (Schuler et al., 1996).

• There is a greater use of tests in France and Belgium than in the UK and Germany (Levy-Leboyer, 1994).

• There is a somewhat greater use of references by British companies, at least as compared to France, Germany and Belgium (Shackleton & Newell, 1994; Clark, 1993; McCulloch, 1993).

• In Greece selection methods are very primitive and simple compared to methods used in other European countries, at least those that are members of the European Community (Tsannetou, 1996).

• In the US drug testing and honesty (or integrity) testing are becoming popular but these selection methods are very rare elsewhere (Shackleton & Newell, 1994).

• In China, Wang (1997) reports that selection decisions rely heavily on personal and economic information and that there is little emphasis on assessing whether the candidate has the competencies to perform the job tasks.

• Italian companies make little use of any method except the interview (Shackleton & Newell, 1994).

This is not an exhaustive list of national trends in selection method use, but it does illustrate the variability that has been observed by empirical research.

Taken together this evidence suggests that more valid and reliable methods are commoner in some countries (e.g. the UK, Germany, North America) than others (e.g. China, Greece, Spain). However, it should be noted that there is a difference between the potential of a method to produce a valid result and its use in practice. For example, Di Milia and Gorodecki (1997) showed through looking at practice in a major Australian company, that although structured interviews may have the potential to produce high levels of validity, in practice this was often not achieved because the interviews were not undertaken as required. Similarly, in terms of the use of psychometric tests, research has demonstrated their inappropriate use in a number of countries:

• Mardberg (1996) reports on fairly extensive misuse of tests in Sweden.

• In New Zealand Smith and George (1992) report on the problem of indiscriminate use of personality tests.

• Rees (1996) notes the problem of inappropriate test use in the UK.

• Engelhart (1996) identified that tests are often used badly in France.

• O'Gorman (1996), looking at selection methods in Australia, identified a lack of training and unprofessional use of selection tests.

Moreover, even though there are variations across nations in terms of the use of more valid and reliable selection methods, the research findings also suggest that in virtually all countries there remains a heavy reliance on those methods of selection and assessment which are not the most valid. In other words, the methods used are not the most valid in terms of accurately measuring individual differences and so deciding on how suitable a person is for a particular job (Di Milia, Smith & Brown, 1994; Rowe, Williams & Day, 1994; Shackleton & Newell, 1994).

Therefore in virtually all countries, 'best practice' in the use of selection methods, at least as advocated by academic research from a criterion-related validity perspective, may be the exception rather than the rule. Nevertheless, taken together the evidence appears to suggest that there are national differences in the selection methods used. Moreover, given what was said in the introduction about the problem of using survey instruments to identify these differences, the evidence probably underestimates the variability that exists. Thus, while obviously there will be variability in selection practice at the level of the firm, it is also clear that there are national differences that need to be explained.