Search This Blog

Jun 19, 2009

The Spreadsheet fallacy

Perhaps people don't remember it, but VisiCalc, which was Excel's forefather, was one of the crucial forces to unleash the personal computer revolution. Excel is great; it allows running infinite number of scenarios quickly, make patterns and ratios visible, and generally, it is a good way of making a solid argument.

However, the whole idea of numbers came about when people started to deal with large quantities of essentially the same or very similar things, like bushels of grain, heads of sheep, or buckets of beer. Almost none of the hunting-gathering societies generally count beyond 3 or 4, because if you had 10 arrows, they were all different, and you remembered them all individually. If you saw a herd of animals, that's what you called it, and there was no need to count them. If you don't deal with identical things, you cannot count them. Or rather, they have to be identical in at least one practically important aspect. For example, you cannot count loafs of bread, if they are very different in size. I mean you can try, but won't be happy with outcomes, because you may end up with more loafs but less bread. So, the sameness of things you're counting is the most fundamental assumption of mathematics.

Here is where the Spreadsheet fallacy comes in. Excel spreadsheets are so compelling that people are tempted to count apples and oranges as pieces of fruit, but then make conclusions about their average skin-thickness. For example, we have some classes that are real classes, and other classes that are independent studies, and still others that are checkpoint courses (which are not classes at all, but tricks to make our registrar database keep the information we need). We have classes for each people are getting paid, and classes that are done as service. Some classes are methods, while others are theory; still others are tutoring classes. Some classes are co-taught by different content specialists, while others are co-taught because it is easier to have one large class than 4 smaller ones. These all require different enrollment caps, different forms of compensation, different rooms, etc. When you put all of these things into one spreadsheet, you must assume that 1 credit=1 credit, and 1 instructor=1instructor. But the basic assumption of sameness, of the consistent unit of measurement just does not work. If the only source of your information about reality is the numbers in the spreadsheet, you may see phantoms rather than the reality. By trying to be objective, you may actually become less objectives. Numbers are only good when they measure something real.

We receive a lot of spreadsheet reports from the University. I am always impressed with their authors' Excel skills, but can rarely see the raw data that goes into the calculations. In more than one occasion, I discovered that the input included numbers that cannot even be added together, because they compare incomparable things. Thank god, most of these reports have been inconsequential so far. However, when some actual decisions and policies can be made based on the Spreadsheet fallacy, remember the rule:


No comments:

Post a Comment