Methods of Analysis some illustrations

The sales manager of the Alloy Steel Company wanted to identify the characteristics associated with successful salesman. He believed that such information would help him carry out his recruiting and hiring responsibilities more effectively.

In order to study this matter, the sales manager compiled the following data for each salesman: whether or not the salesman made his quota this past year; the number of years the salesman had been selling alloy steel; and the salesman’s formal technical education after high school, stated in numbers of years. The sales manager asked the marketing research manager to analyze these data in order to identify the characteristics associated with successful salesman.

The marketing manager noted that these data consisted of two different types of variables. Whether or not a salesman made his quota, was a variable made up of only two categories – the salesman either did or did not make the quota. Each salesman’s years of selling experience and years of technical and years of technical education were continuous variables that could on values of 0.0, 0.5, 1.0, 2.5, 4.0, 9.0 or any other value. The marketing research manager knew he couldn’t use cross tabulation to analyze the data unless of the variables were in categorical form. He also knew he couldn’t use correlation or regression analysis because not all of the data were continuous variables. He would have to look for some other method to analyze the data.

When analyzing these data, it was marketing research manager’s intention to use the categorical variable (made quota / did not make quota) as the dependent variable. The two continuous variables would be used as the independent variables in the analysis. This meant that the marketing research manager would have to find a method that could be applied to data consisting of a categorical dependent variable and two continuous independent variables.

The marketing manager of the Consumer appliance Company wanted to learn how appliance purchasers obtained product information when making purchase decisions. Among other things, the manager wanted to know the number of different information sources people used when evaluating which brand to purchase. If some people used only one information source while others used several or many information sources, it would be helpful to know how the first group of people differed from the second group. The marketing manager felt she could more effectively design advertising and promotional campaigns directed at potential appliance buyers if she knew more about the factors influencing their purchase decisions.

The company carried out a large sample study in which they located 653 households that had recently purchased an appliance. The respondents were asked to identify all of the sources of information they used when making their appliances purchase decision. The researchers also obtained data from the respondents on some 25 other variables, such as the cost of the appliance purchased, family income age of the head of the household and other data.

One of the analyses the researchers wished to perform involved the number of information sources people used when purchasing appliances. That would be the dependent variable in the study. The other 25 variables would be used as independent variables in the same analysis in order to identify how respondents who used many information sources differed from respondents who used fewer information sources when selecting appliances.

The researchers knew that they could use cross tabulation to analyze the survey data. To do so, it would be necessary to treat the dependent variable as a categorical variable (0-1 information sources used 2—3 information sources used, 4—5 information sources used, 6 or more information sources used) rather than as a continuous variable (0,1, 2, 3, 4, 5 ….. ). If any of the independent variables were in continuous form, they would also have to be transformed into categorical form.

However, the researchers decided not to use cross tabulation because there were 25 variables that could be used to describe the characteristics of respondents. They knew that cross tabulation becomes quite cumbersome with so many variables. Because many of the independent variables were in categorical form, the researchers knew that neither correlation nor regression analysis would be appropriate to use. So the researchers had to find some other method to analyze the data obtained from the 653 households.