Saturday, July 30, 2011

Logistic regression

I have decided to analyze the data for the dissertation by using logistic regression. Usually, when I mention that, people's eyes roll up in their heads, their eyelids flutter, and they appear to be hearing a high-pitched shriek, akin to that only heard by dogs.

Basically, this is defined as the pattern of dots on a graph. For example, if height increases, chances are weight will increase in a child (this is not always true for people as they age, but I digress). If the researcher charts the height and weight of each child but putting a dot on a graph, and examines how closely the dots follow a straight uphill line, this is regression analysis. Logistic regression means that one of the factors on the graph is a yes/no, which is similar to the height/weight.

If people express an interest in the process, and I try to explain it, they tend to edge away from me.

I have to admit, I don't get the full theoretic framework of the analysis process. I thought I'd look on Wikipedia to see if this information would be dumbed down enough for me to grasp the deeper aspects. This is what Wikipedia said:

Logistic regression analyzes binomially distributed data of the form where the numbers of Bernoulli trials ni are known and the probabilities of success pi are unknown. An example of this distribution is the fraction of seeds (pi) that germinate after ni are planted.

The model proposes for each trial i there is a set of explanatory variables that might inform the final probability. These explanatory variables can be thought of as being in a k-dimensional vector Xi and the model then takes the form

...and it gets worse from there. Seriously? How did the Bernoulli effect get in there? Isn't that the explanation of flight, where air goes over a curved wing at a different rate than it goes under, causing lift? Is this a way of telling me to stop obsessing and go on faith?

If I can't trust Wikipedia to make this simple, who can I trust?

I did bring home a basic statistics text from work, and it is helping. However, at what point do I just throw in the proverbial towel and view this process like I do air travel: I have a good basic understanding of the Bernoulli effect, but I don't need to know much in order to take my journey.

No comments: