Selection involves picking out the most suitable employee (or employees) from a pool of candidates generated with the assistance of the recruitment process (Holzer, 1987: 1). The decisions taken by organizations during the process of selection establish a sine qua non for efficacious organizational functioning along with having far-reaching consequences of pertinence to individuals (Wilk and Capelli, 2003: 103). In particular, with respect to the choices concerning which method (or methods) should be employed when selecting candidates for a job, organizations are faced with a predicament to either concur with the 'more is better' perspective (Terpstra and Rozell, 1993; Capelli and Wilk, 1997), overcoming the probability of committing selection mistakes on the one hand, or as posited by Guion (1998) on the other, given the substantial cost involved in the selection process be exacting about which practices to adopt in picking out employees. Instead of using all the methods each time, organizations do indeed make decisions as to which method (or methods) shall furnish the most appropriate information about the candidates to be selected for the job. Moreover, given the absence of a 'best practice' to assist organizational decisions regarding the selection method(s) to be deployed, this paper has attempted to explore a set of bases as a guide to how organizations should decide which selection method (or methods) to use when selecting employees for a job. For this reason, a 'contingency approach' (Capelli and Wilk, 1997:6) has been deemed appropriate.
Deduced from a wide body of empirical evidence (e.g. Wilk and Capelli, 2003, Holzer, 1987; Terpstra and Rozell, 1993; Capelli and Wilk, 1997; Barber et. al, 1999; Jackson, Schuler and Rivero, 1989; Ryan, McFarland, Baron and Page, 1999) the identified contingencies serving as cues of which selection method (or methods) should be adopted by organizations, have been categorized into three broad heads for the purpose of the present paper. The clusters are: characteristics of work, organizational characteristics and extra-organizational characteristics. At this stage, it is essential to point out that the identified contingencies are by no means exhaustive and do not establish a normative framework that should be adhered to by organizations while deciding which selection method (or methods) to use when picking out employees. In addition, this paper argues that it may be fruitful for organizations to look into the theoretical issues of validity (content and predictive validity), fairness and feasibility of the method (or methods) when deciding which one(s) would be most appropriate for selection purposes.
In order to expatiate on how organizations might reach a decision regarding which selection method or methods to employ, the following structure has been adopted for this paper. Firstly, a systematic analysis of the three identified grounds (mentioned above) has been performed. Next, it has been endeavored to provide an insight into the validity, fairness and feasibility issues which might play a decisive role in determining organizations' choice as to which method (or methods) to use when selecting candidates for a job.
As noted by Wilk and Capelli (2003), characteristics of work have been found to be related to selection procedures by both job and utility analysis literatures. Through job analysis it is possible to delineate verifiable job behaviors and activities (Harvey, 1991) required by candidates to be successful at their job, having implications for the type of information that selection methods should reveal. Hence, it has been hypothesized that there should be a relationship between the characteristics of work and the firms' decisions regarding selection methods, provided the selection process is preceded by job analysis (Wilk and Capelli, 2003). Similar conclusions have been drawn by Boudreau (1991) regarding the usage of utility models. Given the importance of characteristics of work as a basis for deciding which selection practices to adopt, Wilk and Capelli (2003) in their study of employer choice of selection methods further distinguished work characteristics into three: skill requirements, training and wages. With respect to skill requirements, they found that interpersonal and cross-functional skills (essential for job rotation) were predictive of all types of selection methods i.e. academic achievement methods (such as teacher references, other high school information, and transcripts), methods reflecting work experience (e.g. use of applications, interviews, employer references, and resumes) and test performance methods (for e.g. work sample tests). Their results were further corroborated by similar findings of Capelli and Wilk (1997). Specifically, for low-skilled but physically demanding jobs Holzer (1987) reports the use of physical exams in generating apt information for the job in question. On the other hand, it has been argued from a utility perspective that for low skilled jobs providing piece rates, the skewed trade off between costs and benefits of conducting selection may prompt employers not to screen candidates attentively using a particular method (or methods) (Capelli and Wilk, 1997).
In addition, the increased movement towards teamwork (Ployhart, Schneider and Schmitt, 2006), has necessitated organizations to also delve into KSAs required for performing in a team (Stevens and Campion, 1994). Consequently, this has prompted organizations to address the challenge of 'fit' not only at person-job, but also at person-team level when deciding selection methods. It is believed that team skills can be adequately gauged through references/testimonials and assessments centers (Anderson et. al, 2004). As such, there can be observed a certain degree of overlap between selection methods used for tapping into interpersonal and team skills. Further, it has been found that firms engaging in formal training of employees tend to utilize academic achievement and test performance methods of selection and not methods revealing candidates previous experience (Wilk and Capelli, 2003; Capelli and Wilk, 1997), since these methods possibly divulge more about candidate's potential to learn. Finally, studies by Wilk and Capelli (2003) and Capelli and Wilk (1997), wherein use of wages as a criterion to determine which selection method (or methods) should be chosen, revealed work experience and test performance methods to be appropriate. Despite a robust empirical evidence to its favor, recent convergence on competency based organizations (Lawler, 1993) and the issue of 'bimodal prediction' (Herriot and Anderson 1997 as cited in Anderson et al., 2004: 489) are challenging the efficacy of work characteristics in determining which selection methods should be used.
Along with characteristics of work, organizational characteristics i.e. organizational size, organizational strategy, organizational level and industry to which the enterprise belongs have been assumed to serve as beneficial grounds of employer choice of selection methods (Terpstra and Rozell, 1993; Holzer, 1987; Barber et. al, 1999; Jackson, Schuler and Rivero, 1989). First of all, in a sample of 201 American companies, Terpstra and Rozell (1993) observed organizational size to be a determinant of selection methods chosen by employers, such that relatively larger firms made greater use of validation studies and cognitive ability tests. Research by Barber et. al (1999) indicates that larger firms (< 1000 employees) focus on objective, job-related qualifications, therefore, prefer to take up selection methods revealing academic achievement in contrast to work experience. On the other hand, their findings exhibit the tendency of smaller firms (> 500 employees) to resort to selection practices such as interviews that disclose more about candidates' interpersonal skills, reverberating smaller firms' emphasis on subjective, general qualifications. However, in opposition to the results of Terpstra and Rozell (1993) and Barber et. al (1999), it has been reported by Barclay (1999) that organizations of all sizes use structured interview in the selection of candidates, thus raising questions on whether organizational size should or should not be deemed germane in deciding which selection methods should be used.
In supposing organizational strategy to be a footing through which organizations can make decisions about selection methods, a 'downstream from strategy' rather than 'resource based view' is implicit. Schuler and Jackson (1987) suggest that subject to diverse organizational strategies (such as dynamic growth, exact profit or turnaround strategies) a variance in HRM practices, in particular, performance appraisal, compensation and training and development can be noticed. But, since staffing is one of the HRM practices and given that organizations adopt entire sets of HRM practices (Miles and Snow, 1984b) in relation to specific organizational strategies (Kochan and Chalykoff, 1987), it may be assumed that organizational strategy will lead to variance in the adoption of selection methods by firms, though there is a paucity of research to substantiate this claim.
Schuler and Jackson (1987) have also noted that there is a greater variance in HRM practices within organizations, especially with regards to organizational level. Specifically, selection for managerial posts involves greater use of structured interviews than other lower level posts (Barclay, 1999). Terpstra and Rozell (1993) established the utility of industry type as a ground for organizations' judgment as to which selection methods to employ. In particular, they found service industry firms to employ recruiting studies to a far greater extent than organizations in the manufacturing industry. At a generic level, even Jackson, Schuler and Rivero (1989) have spoken of variability in choice of personnel practices (and not selection methods explicitly) as a function of industry type. Thus, there is ample demonstration through empirical studies to validate organizational characteristics as a criterion by which organizations could decide which method (or methods) to use when selecting candidates. However, given the recent upsurge in the determination of person-organization fit, a salient lack of evidence for organizational culture as being a basis for employer choice of selection procedures needs to be addressed.
Finally, extra-organizational characteristics - national and cultural context, may be looked into by organizations when deciding which selection method (or methods) to use in picking out employees from a pool of candidates. With reference to national context, various studies have found that French tend to use graphology to a greater extent, whereas there is a greater use of references in the UK (Clark, 1993; McCulloch, 1993). Ryan et al. (2006) in a survey across 20 nations observed that organizations in cultures with high uncertainty avoidance are more likely to use job-trials and structured interviews. This might be an important factor to consider for multinational organizations, as candidates in such cultures would perform better, other things equal, if the 'uncertainty' in methods used to select them was 'avoided'. Hence, it shall be advantageous for firms to delve into extra-organizational characteristics when deciding which selection methods to use.
In addition to work, organizational and extra-organizational characteristics, firms should consider the theoretical issues revolving methods of selection themselves i.e. validity, fairness and feasibility. The worth of validity of selection methods as an issue to be considered has been brought out succinctly by Schuler and Jackson (1987: 126) that 'all (organizations may) find it most useful to select individuals using valid as opposed to invalid selection tests'. Thus, construct and predictive validities of selection method (or methods) could also be a footing for enterprises to decide which one (or ones) to use for selecting candidates. Regarding fairness of selection practices, Arvey and Renz (1992) have suggested that organizations should select those methods which are objective (involving minimal subjective interpretation on the part of the employers) and tap into job-related and merit-based facets, maintain confidentiality of data and are consistent. Lastly, the cost-effectiveness and time taken to administer i.e. the feasibility of method (or methods) of selection should be gauged before organizations take a decision as to which method (or methods) shall ultimately be used.
In conclusion, it should be highlighted that in reality, due to constraints of time and resources, organizations do not consider all of the aforementioned bases and ground their decisions regarding selection methods on their most pressing strategic requirements. Subsequently, substantial deviations in decisions regarding selection method(s) to be utilized can be witnessed across organizations. The identified contingencies along with the theoretical issues of validity (content and predictive validities), fairness and feasibility of the method (or methods), nonetheless, do help to provide some insight as to how organizations should decide which selection method (or methods) to use when selecting employees for a job.