The businessmen and women who lost billions in the hedge fund crisis of 1998 should have read this book, or at least given it to their researchers. They were victims of what Trevor Webb would call the "nice to know trap", that is, they had read research-popular at the time-which confirmed what they wanted to hear: those fat juicy fruit were the thing to grow. This book is aimed at experienced business researchers who are charged with explaining the behaviour of such areas as markets or product

more »

... atterns, under different foreseeable conditions. It sets out a disciplined, scientific approach to research in order to avoid the assumptions and intuition on which many businessmen and women base their decisions. Webb believes that there are two main skills a researcher needs: the ability to provide reliable knowledge on which to base a decision, and a means of assessing the quality of opinion surrounding a subject. The practical means of using these skills is the model. This is a replica of a situation (such as a market for razor blades) in which the researcher sets out the range of events (for example, interest rate changes, legal developments or population shifts) which would change the behaviour of the research object. One of these variabies at a time is introduced to test the reaction ofthe object so that eventually the researcher could present the businessman or woman with a predictable pattern of behaviour. The book assumes a degree of previous experience in modelling since it does not give the reader an indication of what the model should Librry & Information Research Nervs (LIR\ Volume 23-Number 73-spring 1999 ISSN 0141-6561 contain (this is dependent on the researcher's particular resources and time limtts).It does, however, give fips on procedure-such as checking the behaviour of the object at different points in time (which is simply a random way of checking your predictions) and not relying too heavily on one source of information-so that the generalist manager who is likely to read the final repoft does not need any speciaiist knowledge. The information which is gathered to put in the modei should be selected against a number of criteria or quality control points. Webb warns against falling into the trap of using simplistic number crunching methods such as aggregates and residuals, and advises that the researcher is alert to bias when using research previously produced by others. He comes up with some useful mental tricks to make the process more efficient: by formulating a null hypothesis, that is, instead of researching whether profits would go up if there were a change of chairman one should look instead to see if not changing the chairman would increase profits. In this way the researcher guards against providing the answer the managing board is hoping for. The research cycle is completed by an audit of the procedure after the decision has been taken to see what predictions failed to materialise, and why the researcher did not accurately predict it. Researching for business is unusual in its field in attempting to supply a comprehensive framework for business research. Most studies are produced by business people themselves and are inherently subject-specific or padded out with descriptions ofthe person's heroic past. The book needs to be read several times over, and in conjunction with real research, because the language used is often abstract and sweeping with a lack of examples. This is not a book for use on an enquiry desk but contains a solid mental lramework for established researchers.