The management requires thoughtful decision making. Regardless of the decision area, the cost of the wrong course might be high. The improper investment and faulty advertising projects result in profit loss. The wrong course of employment results in decrease of customers quantity. The wrong pricing strategy decreases the sales. The manager’s experience and sixth sence are useful, and can be used in everyday situations. However, in serious cases, when there is responsibility before the board of directors, the intuition and other subjective means are inappropriate. Therefore, managers need reliable methods to use in decision making. The statistics makes possible conversion the numerical data into information (Keller & Warrack, 2000) that can be used for decision-making and forecasting.
The statistical data processing starts with the descriptive statistics, which provides the graphical and numerical methods for data summarizing. Data summarizing aims to make the numerical information presentable and understandable. The descriptive statistics includes measures of central tendency and variability, and testing data for normality. The measures of central tendency give information about the values that represents sets of data: mean, median or mode. If the data are numeric (sales, profit, number of customers), mean is the best representation. When the nominal characteristics are measured (described by qualitative characteristics), mode and median are used. The variability (standard deviation and variance) is the measure of dispersion around the mean value. The low value of standard deviation means that the data are clustered around the mean value, while higher value indicates that the data are scattered far from the mean values (Keller & Warrack, 2000). The frequency distribution is a valuable graphical representation of data. It gives the reader understanding of the data, and allows assessing the distribution. The statistics assesses data as “normal” and “not normal”; depending on this, the certain techniques and tests are applied (Srivastava & Rego, 2008).
While the descriptive statistics characterizes a certain object (a store, or an investment project), inferential statistics makes assumptions about the whole population (all stores in the city, the situation in the industry). The methods of inferential statistics enable managers making decisions on the basis of sample data. For this, the sampling distribution and confidence interval are used. The sampling distribution is the representation of all possible variables in the data set. The confidence interval is used to show the interval of the population parameter, calculated basing on the descriptive statistics data. The 95% confidence interval for the number of customers at evening hours is 45-55, and this means that the population is characterized with a value ranging from 45 to 55, and the researcher is 95% sure that the real value of the number is within this range (John, Whitaker & Johnson, 2001).
The analysis of the numerical data often generates new questions. Is the number of customers at late hours on Monday significantly lower than on Tuesday? Are the sales of the assistants significantly different? Is performance of the night and day shift equal? In statistical language, the alike questions are called hypotheses. The procedure of answering the question is called hypothesis testing. There are two types of hypotheses: null and the alternative. The null hypothesis states that the value of the sample (calculated from the data) is equal to some value. The alternative hypothesis is opposite to the null, or the value is not equal (greater or smaller) than the comparison value. Then, the certain statistical test is performed. Basing on its value, one of the hypothesis is accepted, and the other is rejected (Keller & Warrack, 2000). Therefore, the research question is answered.
The statistical test is chosen depending on the data involved. If a manager sets a certain goal value for the sales, and the monthly sales are compared to this value (the mean with the standard deviation is compared to the certain number), then one-sample t-test is applied. When it is necessary to compare the values of two months (two mean with their standard deviations), the independent t-test is applied. When the sales of several (more than two) branches need to be applied, ANOVA, or analysis of variance, is used. The test is performed by calculating the test statistics (Srivastava & Rego, 2008).
The test results are evaluated by comparison with the critical value. If the calculated value is less than the critical value, the null hypothesis is accepted; if it is greater, than the alternate hypothesis is accepted. The critical value is a table value, chosen at the significance level. The significance level refers to the probability of choosing the wrong hypothesis. Typically, for managerial practice, 5% or 10% significance level is applied; when higher probability of the error is accepted, 20% significance level can be used (John, Whitaker & Johnson, 2001).
The statistics is a valuable tool for managers. It enables managers to summarize data and organize them in a presentable way. The statistical tests are reliable scientific methods to assess and compare data, and to transform the numbers into information which can be used in the decision-making process.
John, J. A., Whitaker, D., & Johnson, D. G. (2001). Statistical thinking for managers. Boca Raton: Chapman & Hall/CRC.
Keller, G., & Warrack, B. (2000). Statistics for management and economics. Pacific Grove: Duxbury Press.
Srivastava, T. N., & Rego, S. (2008). Statistics for management.