Have you ever been told to calculate the mean and standard deviation of dichotomous data, and then perform an ANOVA on it? If so, you’re not alone. I’ve seen this mistake happen to many students, and even some professors. But here’s the thing: it’s just wrong. Dichotomous data, by definition, can only take two values – 0 or 1, yes or no, etc. So, what’s the point of calculating a mean or standard deviation? It doesn’t make sense. And as for ANOVA, it’s a statistical technique that’s meant for continuous data, not dichotomous data.
So, what should you do instead? Well, there are a few options. One common approach is to use a chi-squared test or a Fisher’s exact test to compare the proportions of the two groups. Another approach is to use a logistic regression model to model the probability of one of the outcomes. But the key is to recognize that dichotomous data requires a different set of statistical tools than continuous data.
It’s understandable that your professor might have made this mistake, but it’s important to speak up and correct them. After all, as students, it’s our job to learn and understand the material, and as professors, it’s their job to teach it correctly.
What do you think? Have you ever encountered a situation where someone was trying to analyze dichotomous data incorrectly? How did you handle it?