SPREADSHEETS

Spreadsheets in System Modeling

Spreadsheets in System Modeling

Spreadsheets are among the most common types of computer software used by people working in epidemiology and public health. When desktop computers were introduced in the late 1970s, the first 'killer app' ('killer application,' i.e., the software that everyone wants to have) was the spreadsheet. Visicalc, the first major spreadsheet application, was so useful that it justified the purchase of a computer (Langer, 2004).

For people working in fields such as statistical analysis, scientific research, economics, and finance, the ability to easily manipulate large amounts of numerical data presented an immense advantage over manual methods. The consequent time and cost savings easily paid back the investment in a desktop (or 'personal') computer. In fact, spreadsheet software quickly became so popular that it helped establish the notion of the personal computer—a computer used primarily by a single individual and small enough to sit on a desktop, in contradistinction to the mainframe computers that were far more common at the time (Keller, 2000).

When the IBM PC was introduced in the early 1980s, Lotus 1-2-3 became its killer app. Lotus became the accepted standard for over a decade. Its success was coincidental with the runaway success of the PC. By the 1990s, with the introduction of the visual interface of Microsoft Windows and Apple's Macintosh, Microsoft Excel overtook Lotus 1-2-3 as the market leader and remains the standard to this day. Although there are other choices in spreadsheet software, Excel has a market share of more than 90%. Because Excel is bundled with the dominant wordprocessing program Microsoft Word in the ubiquitous Microsoft Office package and because many other programs can use Excel data files, it has become the accepted standard (Frye, 2004).

Spreadsheet software (the name is derived from the spreadsheet used by accountants to record financial information) is a computer program that presents a rectangular matrix of rows and columns to display data (see Figure 1). Each cell can contain numerical or textual data. Columns are defined by letters and rows by numbers. Cells are referenced as the intersection of those two criteria, A1 or D37, for example. In this figure, each row contains the information for one case, in this instance for one patient. Each column represents a variable (gender, date of birth, etc.) for that patient. In database terminology, each row is a record and each column is a field in that record (Langer, ...