Statistical approach to intelligent managerial decision making.

Simicevic, Vanja

Abstract: This paper emphasizes the importance of statistical approach to intelligent managerial decision-making. The research was based on an original empirical survey, conducted on the basis of a random sample of large Croatian firms, aimed at estimating the extent at which Croatian managers use statistical methods. The research results are presented in this paper, and suggestions are given for the promotion of the statistical education in order to increase the level of statistical thinking in Croatian firms.

Key words: statistical education, statistical thinking, intelligent methods, managerial decision making

1. INTRODUCTION

Today's manager cannot carry out great deal of modern management tools (Rigby, 2005) without knowledge of statistical methods and statistical thinking. Statistical thinking and applied statistical methods help managers to cope with modern business conditions and obtain best results for the firm and their personal carrier (Whitaker et al., 2001). Statistical methods can provide many possible solutions for solving the modern age paradox of more information but less understanding. Statistical thinking should have positive influence on managerial efficiency, but in the business practice it is rare to find a manager who thinks, in such a manner, which is a result of the traditional approach to managers' education. The aim of this study is to estimate the extent of use of statistical methods in Croatian firms and to suggest the promotion of the statistical education in order to increase the level of statistical thinking. To this end, an original empirical survey was conducted on the basis of a random sample of large Croatian firms. The results of this research, as well as suggestions how to include statistical thinking in the statistical education, are a scientific contribution of this paper to intelligent managerial decision making.

2. THE METHODOLOGY OF RESEARCH

Random sampling method was used to select 300 firms from the Croatian firms database Institute for Business Intelligence. Among them, 42 firms refused to participate in the survey, mostly due to the policy of business confidentiality. A total of 106 valid answers to the questionnaires of the mail survey were obtained, which corresponds to 35,33%, i.e., acceptable for this type of research (Kish, 1995).

3. RESEARCH RESULTS

It was found out that typical sample firms deal with financial transactions, industrial processing, or commercial trade, and are registered as LLC. The sample included both genders equally, with age range from 30 to 40 years. Majority of managers have a university degree; while minor part holds a MBA degree, or are high school graduates. Regarding the duration of the working experience, less then 10 years is by far the most typical. They mostly occupy the tactical management level, which is explained by their relatively young age (most are below 35 years). The interviewees themselves provided information on their employment position and department, as well as their own assessment about the defined procedures in managerial decision making, and about dominant decision making style in their firms.

The survey results indicate that statistical methods are mostly used in the departments that are traditionally oriented towards usage of quantitative methods in finances, accounting, planning, trade, auditing, production and marketing. Other departments are mentioned rarely, and two of them are not a typical part of the firms' functional organization. These are the insurance and public relations departments, which are normally parts of the firms' functional organization for finances and marketing services (Gogala & Simicevic, 2005).

The employment positions like directors of financial, marketing, planning and production departments use statistics frequently. The controllers, accounts manager, auditors, supply managers, actuary and import clerks are less indicated. Clearly, all of these employment positions correspond to the firms' area of operations. For example, the actuaries are mostly mentioned in insurance and leasing firms, while the auditors are most frequently mentioned in the auditing firms and banks.

The majority of interviewees are familiar with basic statistical notions and techniques, such as frequency distribution, mean, standard deviation and graphs. A minority is familiar with the median and time series indexes. Very few know about the more complex methodologies such as linear trend, regression analysis, hypothesis testing and confidence interval. It is worth emphasizing managers' rather high level of awareness about the existence of statistical methods, but they were not requested to assess the level of their own practical knowledge about their use. Only five managers declared that they are not familiar with any method, but they have only high school degree.

The level of use of statistical methods was analyzed, as well as who is preparing the analyses--managers themselves, or some other employees. Managers are mostly preparing analyses themselves by using basic statistical methods, such as graphs and mean values. Among the methods prepared by the other employees, priority is given to time series indexes, regression analysis, mean, linear trend and graphs. The least methods used are confidence interval for the mean, practiced only by four interviewees.

Finally, interviewees were asked to assess the importance of statistical methods in their business practice. Approximately one third feels that statistics are useful and important for their business. However, one quarter of interviewees are indifferent, and the same proportion believes that statistics are not important at all. Equally small number of interviewees (less then 10%) has both, extremely positive and negative opinion about the importance of statistical methods in their work.

4. STATISTICAL EDUCATION FOR MANAGERS

According to the research results, most managers are highly aware about statistical methods, but they often use only the basic methods, i.e., graphs, measures of central tendency, measures of dispersion and time series indexes. On the other hand, managers consider that their own better knowledge of statistical methods could be useful for their business, while a minor number thinks that they are really important for their business activities. The reasons why managers do not use statistical methods as much as necessary mentioned in the literature (Makrymichalos et al., 2005) are the following:

* In the past, firms were focused on research and development, as well as production, which are not considered important nowadays (Snee, 1991). The modern firms are consumer oriented. Changes are necessary due to highly competitive markets as the result of globalisation. Although statistical methods are frequently used in the fields, which are important for customer relations, such as data mining (Berry & Linoff 2000) the operational and tactical task assignments have still not been focused on the statistical approach.

* Statisticians are traditionally focused on technical aspects of statistical methods. With the exception of Deming's study, a small attention has been dedicated to improving business applications of statistical methods for the purpose of optimising business results.

* Managers feel that statistics offer greatest benefits in quantitative approach to business, and it has been used as alarm in case of a deviation from the expected business results.

* At present, managers are not prepared to change their way of thinking, which often does not include quantitative approach to the problems.

* Traditional approach to statistical education places accent on the mathematical details and computing procedures, and less on practical application (Deming, 1986).

* As a result of traditional approach to statistical education, managers are usually "scared of statistics". Statistics has been established exclusively as a mathematical discipline, with too little practical value and for use only by the mathematically gifted analysers.

In order to introduce the students to the statistical thinking it is necessary to achieve interaction between students and teachers, as well as interaction among the students themselves. Just as in any other discipline, the students should learn more efficiently statistics' subject matters if they are actively involved in teaching process. It is relatively simple to establish interaction among small number of MBA students. Managers should be encouraged to present their firms' experiences and discuss them together. However, even large groups of students can benefit from interaction with teachers. In such groups, it may be difficult to organize teamwork, but the teaching process should not be carried out without including all students in the discussions, though this calls for breaking down their fear of such teaching method.

An efficient way to learn is through the process of discovering. For example, measures of central tendency are one of the most important topics: students should be enabled to find the way of determining median value intuitionally, as well as to estimate the reason why it is so different from mean and mode, for example. Approach, which requires that measures of central tendency must first be calculated, followed by the presentation of the formula, normally does not result in students' understanding. In the area of the regression analysis students could be induced to draw regression line by means of scatter plot, as well as to demonstrate why least-squares method is much more precise than the subjective judgment.

On the basis of such an approach to statistical education for managers, they should be enabled to take on new roles that are expected from them in the era of statistical thinking. (Makrymichalos et al., 2005), as follows:

* Managers should modify their attitude regarding statistical methods. They should understand that statistics is not just a series of statistical methods, but also the way of improving business results.

* Management decision-making should be normally based on the carefully selected data instead of on the intuition. The data collection should be carried out in such a way as to enable search for the causes of variability.

* Managers should distinguish general and special causes of variability in the business process of their firms. Such knowledge could enable them to choose the most adequate options for the business reengineering procedures related to general causes, or for solving the ad hoc problems related to special causes of variability.

5. CONCLUSIONS

The nature of this study is practical: it is dedicated to companies' top management leaderships, which deal with management education, educators in the fields of statistics for the managers in Croatia. It is an imperative that statistical education gains its practical mission, by promoting statistical thinking instead of purely mathematical aspects of computing statistics. Research in the applied statistical methods in the Croatian economy has been rare, i.e., only few have been published (Pejic Bach et al., 1999). In this context, this study provides a preliminary platform for future research on statistical education for managers in Croatia.

6. REFERENCES

Berry, M. J. A. & Linoff, G. S. (2000). Mastering Data Mining, Wiley, ISBN: 978-0-471-33123-0, Chichester

Deming, W.E. (1986). Out of the crisis, MIT Press, ISBN: 0262541157, Cambridge

Gogala, Z. & Simicevic, V. (2005). Usage of Statistical Methods in Croatian Firms. Zbornik radova Ekonomskog fakulteta u Zagrebu, Vol.3, No.1, pp. 321-339, ISSN: 1333-8900.

Kish, L. (1995). Survey Sampling, John Wiley & Sons, ISBN: 978-0-471-10949-5, New York

Makrymichalos, M.; Antony, J.; Antony, F. & Kumar, M. (2005). Statistical thinking and its role for industrial engineers and managers in the 21st century. Managerial Auditing Journal, Vol. 20, No. 4, pp. 354-363, ISSN: 0268-6902

Pejic Bach, M.; Dumieic, K. & Gogala, Z. (1999). Using Forecasting Methods in Croatian Large Firms. Ekonomski pregled, Vol. 50, No. 7-8, pp. 780-803, ISSN: 0424-7558

Rigby, D.K. (2005). Management Tools 2005: An Executive's Guide, Available from: http://www.bain.com, Accessed: 2005-01-01

Snee, R. (1990). Statistical thinking and its contribution to total quality. The American Statistician, Vol. 44, No. 2, pp. 116-121, ISSN: 00031305

Whitaker, D.; John, J. A. & Johnson, D. G. (2001). Statistical Thinking for Managers, CRC Press, ISBN: 1584882484, Boca Raton