Blog | Admin | Archives

Seniority Bailout

As originally suggested by Nordsieck, I have compiled a couple of graphs depicting the relationship between voting on the bailout plan and the seniority (time since first elected) of the voter. Check it out below, and feel free to draw your own conclusions. My conclusion? DC corrupts.

For more damning evidence, check out the relationship between campaign donations and voting in the House of Representatives.

2 Responses to “Seniority Bailout”

  1. Scott Says:

    Please send me your data. The statistically sound way to determine whether seniority is statistically significantly correlated with voting yes or no would be to use a logit regression model with dependent variable binary yes/no coded either as (Yes = 1, No = 0 or the opposite) and independent variable years length of seniority. The problem with doing the analysis the way you have is that it suffers from what econometricians call hetoskedasticity. More can be read about logit model here:
    If you would like a good software program for doing the analysis, I would recommend R:
    If you would like to just do the analysis quick and dirty the following website has a useful web applet

    FYI if you do end up taking the regression analysis class we talked about, they teach these techniques, which is very cool because then you are empowered with the tools to generate your own conclusion and not subject to what some statistician with a particular bent feeds you.

    Also, another very good article that I came across that contradicts my response to you posting from yesterday, but seeing as I like to facilitate the dissemination of information from all view points, I think it is definitely worth a read.

    Also, some great responses from the U. Chicago Professors in Economics and Finance, a school whose unique thought process has significantly influenced me.

  2. Spencer Says:

    This is a suggestive pattern, but I agree with scott that the linear regression isn’t really what you want in this situation. What you’re really interested is how the probability of voting yes changes as a function of time in office.

    However, without doing any further analysis, what are the r^2 values for the regressions? Those correlations look pretty week to my eye.

Leave a Reply