Data Collection
This study uses a unique dataset, which is a subset of the data that has been used by Bhansali and Zhu [12]. The dataset includes estimated IT expenditure of 329 large companies for 2005. The data was collected by phone interviews using a questionnaire designed by the research team. The questionnaire was distributed to the participants before interviews. The data was collected from approximately 600 firms; since some of these organizations are privately owned or provided unreliable/incomplete data, only 347 questionnaires were usable. As a result of mergers, acquisitions, bankruptcies and missing data, financial data for approximately 230 organizations were available for 2005.
The average annual sales in the sample period are $17.58 billion and the total sales are approximately $4.04 trillion. On average, organization spent $318 million annually on IT, more than a third of which was IT labor expenditure. Approximately 25% of the observations are from the metal related manufacturing sector, which is the largest, but not overwhelmingly dominant, area. The wood related manufacturing, and the finance and insurance sectors each account for more than 10% of the observations. The …show more content…
These variables are estimated by summing the related variables into one main variable. Three-step hierarchical regression analysis is used to test the framework. Hierarchical regression analysis is useful when prior literature or theory identifies sequence of variables to be added to the regression equation. Each step includes a set of variables which will affect the outcome variable in one direction. Therefore, using hierarchical regression will help us to identify the effects of control variables, independent variables and moderating variables separately. The list-wise case exclusion method is used to ignore missing values. Hypotheses are tested using the following four