|
Article Excerpt
Expertise can be acquired in different ways. Some professions, such as medicine or the law, require that practitioners acquire specific educational qualifications before they are permitted to practice. Other types of expertise are acquired through practice 'on the job', perhaps supplemented with skills-based training through commercial courses and workshops. Typically three practices are involved in the acquisition of expertise: education (bringing up individuals and groups so as to form habits, manners, intellectual and physical aptitudes); training (instruction and discipline in or for some particular art, profession, occupation or practice); and development (the unfolding and bringing out of full potential through relevant experiences) (Harrison, 1992). Learning is said to have resulted when there are relatively permanent changes in behaviour that occur as a result of these practices. The purpose of this paper is to address a previously unanswered question: does the way that professional skills are acquired influence the level of expertise developed? In other words, is it better to learn from a book or from experience?
Although previous studies have looked at the differences between the performance of experts and novices, or the differences between experts with differing levels of experience, these studies do not compare the expertise acquired through different mechanisms. This is because experts usually differ from novices in that they are at different stages in the process of acquiring their expertise via the same route. For example, practising clinicians differ from graduate students in that they have both formal education and experience. In areas such as medicine it would be impossible to contrast graduate students with clinicians who had had little or no formal education because legal restrictions prevent such a group from existing. This therefore restricts investigations to the consideration of the impact of experience superimposed onto a base of education.
The domain studied in this paper, trade credit management, is unusual in that education and experience are largely separated. Typically, credit managers have little formal education in their field. Yet the principles of credit risk assessment, one of the central parts of their skill set, are routinely taught as part of degree programmes in finance and accounting. Students taking these courses rarely pursue credit management as a profession, and usually move into more general careers in accountancy and corporate finance. Lecturers teaching these subjects are also unlikely to have been professional credit managers. This therefore provides us with a rare--possibly unique--opportunity to compare groups who have acquired expertise in credit-related tasks either through experience or through formal higher education.
The domain: An overview of trade credit management
Most trade between firms takes place on trade credit terms, with customers in the UK typically being allowed between 30 and 45 days credit on purchases. Trade credit management is the management of a company's trade credit relationships with its customers or, equivalently, the management of a firm's accounts receivable. These accounts receivable will be a major component of the firm's current assets (along with cash, marketable securities, pre-payments and stock), so the performance of the firm's trade credit managers can have a significant impact on its financial health. Yet these credit managers are not required to have specific professional or academic qualifications before entering the profession, and are rarely required to undertake specific professional training thereafter.
There are six main functional responsibilities associated with the granting of credit to customers (adapted from Mian & Smith, 1992): (1) assessment of the customer's credit risk; (2) making the credit granting decision with regard to credit terms and, where relevant, credit limits; (3) collecting the receivables (debts) as they fall due and taking action against defaulters; (4) monitoring customer behaviour and compiling management information; (5) bearing the risk of default or bad debt; and (6) financing the investment in receivables. The tasks used in this study fall within the first two functional areas; credit risk assessment and credit granting. The principles of these tasks, along with trade credit management, are taught in higher education in courses on corporate finance or financial management (typical textbooks are Brealey & Myers, 1996; Keown, Martin, Petty, & Scott, 2001).
Firms will generally undertake a risk assessment when dealing with credit applications. This will usually be based on accounting information such as that used in this study. This information can be obtained from a credit reference agency, but a credit manager could also request that the company supply copies of its accounts. The output of the risk assessment will be a prediction of creditworthiness--will customers pay their bills? will they pay on time? (1) Although these are predictions about payment behaviour, in the UK (particularly prior to the advent of payment behaviour scores supplied by credit reference agencies) this has usually been done using corporate failure prediction models. Two of these corporate failure models, Altman's Z-score (Airman, 1968) and Taffler's PAS score (Tafler, 1983), are particularly well-known in the credit management field and are used to diagnose a firm's financial health. Altman's model is also discussed in standard texts on corporate finance such as those mentioned above. In addition credit managers will typically use financial ratio analysis to analyse potential customers' accounts and to assess their financial stability, (2) and again this approach is taught in higher education courses on financial analysis and corporate finance.
Developing expertise
The study of expert performance has drawn on two contrasting perspectives: those of cognitive science and behavioural decision research (e.g. the reviews in Ericsson & Smith, 1991). The cognitive science approach focuses on the differences in how experts and novices approach problems. Chess experts, for example, differ from novices both in how they encode information on positions and in their evaluation of potential future moves (e.g. Chase & Simon, 1973a, 1973b; DeGroot, 1965, 1966; Holding & Reynolds, 1982; Simon and Gilmartin, 1973). Charness (1991) provides a review of this research. Expert computer programmers produce a richer encoding of available information than do novice programmers (e.g. Aldeson, 1981; Erlich & Soloway, 1984; Rist, 1989; Soloway & Erlich, 1984). Expert physicists tend to classify problems in terms of the underlying physical laws or principles (e.g. acceleration, inverse square law), while novices give greater weight to the surface features of problems, e.g. classifying problems according to whether they involve springs or pulleys (see, for example, chi, Felovich, & Glaser, 1981; Chi, Glaser, & Rees, 1983; Heller & Reif, 1984; Larkin, 1983, 1985; Larkin, McDermott, Simon, & Simon, 1980). The underlying message of the cognitive science approach is that experts structure and represent problems significantly differently than do novices (Lonka, Joram, & Bryson, 1996).
In contrast, the primary focus of behavioural decision research is on the performance of experts relative to novices or to statistical models. Many such studies involve predictive tasks where the relationships between inputs and outcome are probabilistic and characterized by potentially imperfect information, using domains such as psychological diagnosis, graduate admissions and economic forecasting (e.g. Armstrong 1978; Dawes, 1971; Dawes, Faust & Meehl, 1989; Einhorn, 1972; Goldberg, 1970). Generally, these studies have not found that expert judges do much better than novices. As Camerer and Johnson (1991) put it, experts 'know a lot but predict badly'. They conclude their comprehensive review of the literature by observing that 'the depressing conclusion from these studies is that expert judgments in most clinical and medical domains are no more accurate than those of lightly trained novices, [and] expert judgments have been worse than those of the simplest statistical model in virtually all domains studied' (p. 203). They cite a long tradition of research showing that models based on simple combinations of variables could outperform experts in a number of decision tasks (Dawes, 1979; Meehl, 1954; Sarbin, 1944).
Shanteau (1992) has investigated what distinguishes those domains where good and poor expert performance are to be found, and suggests three factors: availability of feedback (no feedback means poor experts), predictability of the outcome (poor predictability means poor experts), and whether the decisions involve people rather than objects or things (objects are easier to predict). In addition, two other factors also differentiate between domains of good and bad expert performance. First, feedback must not only be available, but it must come more-or-less immediately (e.g. Thaler, 1987). Second, experts do well when all information is openly available (the chess board is fully visible, the physics problem is fully specified), but do less well when information is 'beneath the surface' (people have hidden agendas and the stock market is chaotic). (3)
In this study we compare judgmental expertise developed through education versus experience. Our initial question relates to the quality of the judgments made by experts who acquired their knowledge and skills through the two routes. As indicated earlier, previous studies in medical domains have indicated that training and education can improve performance but that experience does not cause further improvement (see Garb 1989, for a review of such studies). There are several studies in other domains that support the suggestion that education or training can quickly improve performance, even bringing it to a level beyond that achieved by experts with many years' experience. In a particularly engaging study, Biederman and Schiffrar (1987) investigated the impact of instruction on the ability to sex day-old chicks. Novices who received instruction were compared with experts who sexed chicks for a living. (4) The novices first used 'their own intuition' to identify the sex of chicks from photographs, which included particularly difficult cases. Half of the novices then read an instruction sheet that explained how male and female chicks could be distinguished. These 'trained' novices improved their performance markedly, reaching the same level of accuracy as the experts. Yet the experts estimated it had taken them between 2 and 6 years of experience to reach their level of expertise. In a study where the context is more closely related to the current one, Hershey and Walsh (2000) investigated the impact of training on the performance of novice Financial planners. The training involved introduction to a conceptual model of the task and worked hypothetical examples. In later testing, trained novices made better decisions than expert financial planners.
In both the cases described above, it was possible to provide novice learners with clear principles underlying good task performance (what to look for in the case of the chicks; a conceptual model of the financial planning task). Doing this accelerates the learning process because individuals are guided to the relevant areas of the problem space. These results therefore suggest that training can be an efficient way to acquire expertise, and that in some circumstances it can even outperform experience.
Sometimes both training and experience help. Indeed, this is at the heart of Alan Lesgold's (2001) influential work, which involves guided 'learning by...
|
