A SLOEDP EXPERIENCE: Demystifying the Numbers

2 Votes with an average with 1
In our second blog on elections data analysis we have looked at the numbers with regard to (1) their completeness, (2) how many voters did not go to the polls per district, and (3) predicting the elections results without going into the assessment of the how those who have voted for 16 parties already will vote. We looked at the summary of all elections results now that the NEC has placed them on their website and the first thing we noted was an error in the numbers we shared the other day. But no damage is done. Here’s what happened: in our haste or attraction to the 249,000-odd votes shared a day or two ago, but especially that we did not have all data from NEC, we missed the 23-odd thousand that were distributed to the other 14 parties. So in short, the number of 618,021 yesterday should be changed to 641,780. A difference of 23,759 voters. Pay attention to that number. We will speak to it later when we look at voter registration efforts in another piece. Next, we took the totals of actual votes as per the elections results published by NEC and contrasted it with the total voter register numbers, per district. From there we were able to tell the numbers of voters per district who did not go out to vote for whatever reason. This is valuable to both parties still, even as the author believes we are in a broken race based on all the confusion with regard to how NEC has handled these elections. So there are two broad categories of data we will be looking at overall. On one hand, we will look at those who registered to vote but did not vote at all. On another hand, the second series of analyses will deal with ‘what if’ situations in the movement of alliances and/or the effect of non-alliances and as other dynamics take shape. All fourteen parties aside from the APC and SLPP, make up a total of 356,892 votes out of 2,537,122 actual voters; or fourteen percent. The NGC and C4C carrying a bulk of that with 261,734 votes nationwide. We have created three pockets of data in these analyses: the SLPP strongholds (Bo, Kenema, Kailahun, Bonthe, Moyamba and Pujehun) the APC strongholds (Port Loko, Tonkolili, Karene, Falaba, Kambia, Bombali and Koinadugu), and the neutrals (Western Area Rural and Urban; and Kono). On some occasions, we will group them all together to show a national outlook for the two remaining parties still in the race (see numbers below). Meaningful data and real work. So this is the approach we will be using over the next few weeks to bring analyses and projections based on NEC’s published results. The idea is to have fun with the data, but in a useful and informative manner that others can try as we go through a particular practical exercise like elections.  Now let’s proceed to look at the reality behind some of the numbers. PART 2 With all that above said, in these national elections analyses series, we pay attention the number of votes attained percent-wise in both SLPP and APC strongholds. In addition, the picture will not be true with the inclusion of deemed neutral spaces that can sway one way or the other, or stay around the middle. And these are Kono and the Western Area (both Urban and Rural). The numbers show thus: If all the voters that are registered but did not vote as the list shows, voted, in the runoff, the two parties, would share a total of 564,816 out of the aforementioned 641,780. But wait a minute!  That number is overblown by those who have died moved out of the country or area, those who are apathetic (and this number may increase now that the NGC is out, e.g.); and most importantly, those who did not pick up their voter ID cards.  For the sake of argument, let’s assume there are 100,000 such people. Okay, make it 200,000 to be generous!  And considering that as of February 20th, 2018, 165,000 cards had not been picked up.  Yes, that's high if you're thinking so! So based on that assumption the total number of adjusted remaining voters are 441,780 (641,780 - 200,000) nationwide and therefore both SLPP and APC have that pool to work with. Now, knowing that not all 100 percent of the remaining non-voting but registered voters will participate, we can adjust the percentages to half (50%), three quarters (75 %), or even a quarter (25%) to project what could happen in different 'what if' scenarios.   We can also adjust the potential of 441,780 to be allotted to both parties in the runoff based on any given percent of choice or common sense.  Other scientific variations including the real percentages reflected in how people have voted already in round one (like we are doing in some cases). But with the caveat that the numbers may be distorted due to alleged rigging efforts that are still under investigation. All analyses are based on numbers NEC have shared with us as certified elections results.  Here is something to ponder: the SLPP had five districts in which they gained 80% or more, inclusive of one where they had 90%: Bonthe!  The APC, on the other hand, has gained 80% or more in only two districts; comparatively.  What does that mean, the numbers beg for interpretation?  I want to reiterate that this analysis speaks only to the numbers that are outside of those that voted. The only thing the current voting is related to is that we have used the percentages based on the actual results to make inferences on ‘what if’ conditions.  Hope you can bear with us for being elementary in our illustrations. At SLOEDP, we want others to start doing the same and become open elections data miners. With a sense of purpose!  More to come...

Add new comment

The content of this field is kept private and will not be shown publicly.

Plain text

  • No HTML tags allowed.
  • Lines and paragraphs break automatically.
  • Web page addresses and email addresses turn into links automatically.
One file only.
2 MB limit.
Allowed types: pdf doc docx xlsx csv jpg png jpeg.