More statistical analysis of MPs' expenses

Following my recent post on the lack of correlation between MPs' staff costs and the amount they spent on office costs, I thought of another interesting test I could do.

TheyWorkForYou publishes figures detailing what proportion of letters sent via their website are actually responded to within a reasonable time. We would expect that MPs with large staff costs should be able to get prompt replies more often than their understaffed colleagues, wouldn't we? (Actually, given we think they're paying their wives and families to do nothing, we wouldn't expect this at all, but let's play along with the hypothesis, shall we?)

Here's the graph:

Again, I've plotted a best fit line so we can understand what's going on a little better.

This time, there is a microscopic correlation, but even so, it's still not good news for our friends in Westminster. The slight downslope to the graph actually seems to show that an MP with low staff costs is more likely to answer his correspondence on time than his high claiming counterpart.

What possible explanation can there be for this anomoly?

I just can't imagine.... 

Comments

[Ian]

If you removed the outlier (£0 staff cost?!?), the R2 would be pretty close to 0.

[Bishop Hill]

The outlier is correct, apparently. Philip Hollobone does all his own typing.