Calling The Election
OK folks, don't eat me alive (or do backflips)...
I have been playing with data (I love data) from the Mars website, and it appears to me that with these three critical assumptions --
(1) that the provisional results reported are accurate,
(2) that the county results reported are representative of the county as a whole, and
(3) that the turn out in all counties is the same, then:
The worst Uhuru could end up with is 50.28% and the best Raila could end up with is 45.70%. Wow, UK has clearly won this election, but I would posit that it is too close to call it in the first round (any one or more of my assumptions may be off).
I used the confidence interval of proportions... Happy to share the spreadsheet, perhaps my statistics are not spot on, but seems clear to me.
Here's how I calculated it.
I have been playing with data (I love data) from the Mars website, and it appears to me that with these three critical assumptions --
(1) that the provisional results reported are accurate,
(2) that the county results reported are representative of the county as a whole, and
(3) that the turn out in all counties is the same, then:
The worst Uhuru could end up with is 50.28% and the best Raila could end up with is 45.70%. Wow, UK has clearly won this election, but I would posit that it is too close to call it in the first round (any one or more of my assumptions may be off).
I used the confidence interval of proportions... Happy to share the spreadsheet, perhaps my statistics are not spot on, but seems clear to me.
Here's how I calculated it.
- I downloaded the provisional results from the Mars website.
- I calculated the proportion of votes that UK, Musalia and RAO had in each county.
- I then calculated the confidence interval of error for proportions from here and calculated the worst Uhuru could do per county and the best that Raila could do in the same county.
- I multiplied those numbers by the numbers of registered voters and simply summed down.
0 Comments:
Post a Comment
<< Home