How the Study Was Done ?

How the Study Was Done ?

Hi Friends, 

You can see how the study was done.The process resulted in minor attrition and a usable sample of 497 firms. Besides, all firms did not respond to all the items in the InformationWeek survey.

We obtained data for our study from two sources:


InformationWeek, a weekly magazine for business and information technology managers, and Compustat, a database of financial and market information on public companies. In 1994, InformationWeek surveyed Computstat’s top 500 firms for data on IT budgets, replacement costs of computers, client/server expenditure, and a breakdown of IT budget in terms of hardware, software, telecommunications, and IS staff expenditure. Corresponding financial data for these firms were obtained from the Compustat database.

For instance, only 231 firms reported data on IT budgets, and only 192 firms reported data on replacement costs. The validity and reliability of secondary data is often debated in the literature. Data obtained from secondary sources may be problematic due to inconsistencies in definitions and, consequently, the data reported. This problem is further accentuated when multiple secondary sources are considered. The data reported by InformationWeek has been found comparable to data from International Data Group (IDG), a leading IT media, research, and exposition company.

The same study also found that the total annual values from IDG data were comparable to those reported by the Bureau of Economic Analysis (BEA), a division of the U.S. Dept. of Commerce providing various economic and statistical data. The InformationWeek data is thus consistent with data from other secondary sources, such as IDG and BEA. Table 1 presents the study variables, their sources, definitions, and descriptive statistics. Replacement cost of computers was taken as a measure of IT capital [7]. Breakup of IT budget was reported by InformationWeek as expenditure on hardware, software, telecom, and IS staff. Client/server expenditure was reported separately as a range; the midpoint of this range was used as an estimate of an organization’s expenditure on client/server computing. Client/server expenditure represents hardware, software, telecom, and IS staff expenditures devoted to client/server systems. Firms in the sample were categorized as one of two sectors: manufacturing and service.

Primary Standard Industrial Classification (SIC) codes (published by the U.S. Dept. of Commerce) were not used to categorize firms into different sectors of the economy as this would have resulted in a too-small number of firms within each group to permit meaningful analysis. We considered six performance variables: business output as measured by value added and by sales; financial business performance as measured by two accounting-based ratios (ROA and ROE); and intermediate performance as measured by labor productivity and administrative productivity. Administrative productivity is the ratio of value added to the total administrative costs of the firm (similar to Strassman’s [11] concept of management productivity); labor productivity is the ratio of value added to the total number of employees. Descriptive statistics show that average IT budgets are greater than IT capital, suggesting a high rate of obsolescence for IT.

On average, 25% of organizational IT budgets are devoted to client/server computing. We applied the production function approach (a methodology relating outputs to multiple inputs) to measures of firm output—an approach employed in earlier research in IT investments [7, 9]. For analysis, we used the natural logarithm form of the Cobb-Douglas production function: Log (Performance) = ß0 + ß1Sector + ß2Log(Capital) + ß3Log (Labor) + ß4Log(IT input) + e IT factor inputs considered individually were IT capital, IT budget, client/server investment, and IT infrastructure investment.


Each component of IT infrastructure was entered in a separate regression equation, as dollar values of these investments are highly correlated. The IT input variables were entered after removing variances in firm output associated with differences in capital, labor, and sector. The effect of IT investments on firm business and intermediate performance was analyzed in the tradition of earlier research on IT and business value [2, 7, 10, 12]. The association of IT inputs with business and intermediate performance was estimated using hierarchical regression. Both sector and size were used as control variables. IT inputs were entered after the variance due to size and sector was eliminated. As size and IT inputs were skewed positively, their natural logarithms were used. The estimated regression equations were

Post a Comment