Search This Blog

Monday, November 26, 2007

The Creative Arithmetic of The Steadmann Polls

I badly need some further education in mathematics to understand something that really puzzles me.

I wasn’t really all the bad in maths and I scored a decent credit in my O-levels, however my limited knowledge is obviously not enough to enable me fully grasp the latest Steadmann poll.

Here’s my simple question.

How is it possible for a presidential candidate to lead convincingly in 6 provinces but end up tying with his opponent who has a strong showing in only two provinces?

Don’t get me wrong. I am genuinely looking for answers here because I believe there is a logical explanation to this. One of the possibilities of course could be the fact that the single community, from which the latter presidential candidate is receiving most of his support from, outnumbers several other top communities in Kenya put together.

Having said that, let me add that the Steadmann folks are way too sensitive to effectively handle the very public and very political tool that they are dealing with here. Right from the word go senior executives have been leaving comments in this blog whose main aim is to ridicule rather than “educate” us so that we understand this polls thing better.

A recent comment by A Steadmann executive has even taken a cheap shot at this blogger. The comment appears below THIS RECENT POST which was written by Phil (and not by me as our Steadmann executive seems to believe).

At one point the obviously uninformed Steadmann sympathizer says of this blog and blogger;

So if you ask me all this is just propaganda to ensure Kumekucha gets more hits, have you people ever asked yourselves how much money this guy gets from this blog? did you maybe think that he cleverly makes up this stories to ensure you all get fired up and keep posting comments not mentioning visiting this blog every other day.

This cheap attempt to incite readers against this blog is uncalled for and it is clear that the Steadmann folks understand very little about blogging and this blog in particular. (Admittedly as little as we understand about their sensational and erratic polls).

For the record, a decision was taken a few weeks ago to allow top commentators to make uncensored posts in this blog. The idea was to ensure a balance and for this reason we now have top authors in this blog from across the political divide and beyond. For instance, there are those who are die hard ODM supporters (and have no qualms letting everybody know about it) and we also have fanatical PNU and ODM-K supporters. And finally we have the horseless or neutral guys. Incidentally the post that so irked the Steadman executive so much so that he felt obligated to take his cheap shot, was penned by Phil aka “rabidly ODM”.

My advice to Steadmann is not to waste energy targeting the messenger but rather to address themselves to the message. It is no secret that in most Kenyan’s eyes, Steadmann has lost all credibility, indeed many readers (from right across the political divide) have said so here. That is not Kumekucha’s fault.

Their energies badly need to be directed towards repairing their public image and in “educating” Kenyans more so that we can understand their polls better and thus say the more flattering things about them that they so badly crave.

P.S. I dare these so-called anonymous Steadmann symptahizers leaving comments here to use their real names and job titles and quit hiding behind the anonymous tag.


Anonymous said...

Its all very simply! if you don't like what Steadman is churning then ignore it!Steadman finally relocates to his familiar teritory as far as you are concerned.
What will happen if RAO tops the next poll ?will you embrace the results as you have been doing in the last 4 weeks?

Anonymous said...

Steadman is taking the Kenyan public for a ride. It is unfortunate how some clique of people underate our intelligence. Why rob Kalonzo (poor guy) to uplift Kibaki. The Dec poll will make or break steadman.

Anonymous said...

It requires no genius to figure out what Steadman is doing. Since Central province has more voters than say Western, he/she i.e. Steadman for instance picks 20 people from Central and 5 from Western. Obviously, if 15 from Central say they will vote for Kibaki, and 4 will vote for RO, the results will appear as if MK is catching up.

However, that way of doing things masks the truth that Central votes are for Kibaki at all times. However, these votes are not enough to give him victory since the other 6 provinces will vote against him.

If Steadman is telling us the truth, he should tell us what issues have made voters change their minds, and where that has happened. Otherwise picking many voters from MK strongholds is misleading.

Anonymous said...

The Steadman Group was on the defensive after politicians and readers took issue with the figures it released in its latest opinion poll.

Curious readers called The Sunday Standard saying the figures from Friday’s poll did not add up.

A casual addition of President Kibaki’s support per province as released on Friday then divided by eight (the number of provinces) showed that Kibaki had scored 39.5 per cent, Raila had 47.6 while Kalonzo had 8.5 per cent.

But Steadman, while releasing the polls on Friday, said President Kibaki and Mr Raila Odinga got 43.3 per cent and 43.6 per cent, an indicator that Kibaki, who had trailed Raila by over nine per cent two months ago, had caught up with the ODM presidential candidate.

There were also questions about figures that did not add up in the provincial vote. An addition of the figures from Nyanza Province showed that some 10 points were missing, making the figures not add up.

Alex said...

Taking the challenge from Chris, I post under my name but will not put my Job title or any other information since I do not yours or for any other person on this blog.

You probably have misunderstood issues here. in your great intelligence, you failed to notice that the issue was not that anyone is a die hard supporter of Steadman e.t.c. The issues are simple here. Steadman has been around with us for quite sometime now. For a larger part of this year, they have published poll results where Raila was on the lead and Kibaki trailed. Then the results were hailed as products of research per excellence. Now according to the polls Kibaki is closing on the gap (he is still not leading), and they have lost credibility, they are not objective, they are taking Kenyans for a ride, I dont know who was slapped by who in state house. Common Chris, this is crap! and by the way am sure you are aware of the other organisations doing weekly polling on the presidential candidates. would you care to tell us who is credible and has some moral soil under their feet??

My take is the election will be won on the ballot box on 27th not by steadman, infotrak, strategic research e.t.c. so if what they reflect is a manipulated poll or a personal opinion, so what? so do majority of the people here on this blog.(and others).

Lets do political commentary in an a way and that makes us more informed to make a decision on 27th (or ammend the ones we already have made.)

chris said...

Thank you Alex for taking up my challenge.

Obviously you only read posts in this blog where Steadmann is mentioned. No problem, I know how busy you must be. So let me give you a litle info clearly documented throughout this blog.

Kindly read the slogan at the top of this page. I AM NOT a supprter of RAO and if you must know my preferred presidential candidate it is one John Githongo. Sadly he is not standing because of threats on his life. But that's Kenyan politics for you, Alex.

So to be honest I don't care whether Father Xmas heads the polls the next time or Pius Muiru. What I would really love is for you to explain (I am sure there is an explanation as to how a candidate can lead in 6 provinces and tie with his opponent who is only strong in 2.)

This blog is a good forum for you to educate the public so that we all understand polling better.

By the way my main strength is that I am always willing to learn.

To prove my good will in this matter, if you can post an explanation here, I will re-post it as a main post in the main blog so that as many people as possible see it.

Please don't be fooled or put off by the biting and hurtful remarks you see here from readers, this is politics. Just be confident in one thing; our main aim here is sober debate. Of course it gives us more hits when those debates are controversial, but you guys also got very famous based on your polls which was a controversial new idea in Africa and in Kenya.

Controversy sells and it also helps position products in the market muc more quickly and at a much lower cost.

Once again thank you for taking up the challenge and thank you for commenting.

Now can we work together in repairing the Steadmann image as we get the truth out there (Kenyans will not accept anything less than the truth).

Have a nice day Alex, and I'm looking forward to your reply.


Anonymous said...

I was a former respecter of the Steadman opinion polls not anymore. I think it is a mark of desperation to resort to using such common and low down tactics to make it look like MW is winning. IF he was actually winning outright in the polls I would have no problem with accepting the results and so would all other kenyans but to try and blatantly manipulate the results and then try and pass it off as genuine.. that is despicable.
What I know for sure as a kenyan.. anybody who can stoop that low in a bid to look popular should never be voted for anything. If you cannot be honest with yourself how can you be honest to us. No wonder we have been subjected to the fallacy that democracy has become in Kenya all because some rich old guy is scared of facing his sunset years enjoying his wealth ( I will not even touch the ill acquired gains)

Anonymous said...

Leading in NE and coast and other densely populated provinces does not help, it is places like RV, C and Nrb tha count. Kibaki lead and support is in places that have more voters. Every % in RV means more votes than % in NE. It is simple really.

Anonymous said...

STEADMAN failed a simple arithmetic test. they have been correctly adding the provincial votes and dividing by eight. this time they didnt do that. infact, the previous results they did not do the provincial ratings.

the bitter truth which steadman has to deal with is that no one supporting ODM, PNU or ODMK will change sides now. so they just try to play with figures. seemingly, they are not good at it.

Alex said...

Chris, You are wrong again, I read this blog many times, steadman mentioned or not. I have only posted comments as annonymous before but will henceforth post using my name lest you brand me a coward (a word I detest.)

Some maths lesson for you - 90% of 10 is 9 and 10% of 100 is 10 - Obviously then, if a guy is said to have 90% of a province with ten voters and another has 10% of a province with 100%, the percentages notwithstanding, the one with 10% of 100 will win because winning is not about percentages but absolute numbers (a pity that the ammendment to have the president garner over 50% of all votes cast was rejected) then atleast that percentage would be important.
Second lesson, when calculating total percentage of the winning candidate in the polls, you dont take the percentages in all provinces and divide by 8. Big NO. you take the absolute numbers again and perform the maths afresh.

As for whether or not you support RAO or not...its out in the open...and there is nothing to hide from here - you have a right just like anybody else to take sides in this important. All me and others ask for is consistency and integrity in the debates!

Anonymous said...

Alex of Steadman,

Thank you for your contribution and explanation. I think it has helped a lot of readers!

Am not a Kenya but I have lived in Kenya for quite sometime and I fight the politics of Kenya very interesting. I also find Mr. Odinga a quite intriguing political operative...

I however, think that Steadman is not adhering to correct scientific polling practice. For example, a fortnight ago you released results and explained that you had changed your sampling method and failed transparency test when you also failed to release provincial/regional data. You further explained that you sampling had previously been done based on population distribution and now you do it based on registered voter distribution. But then you failed to confirm that those polled were actually registered voters... Could you explain how you came to the conclusion that Mr. Kibaki had closed in on Mr. Odinga yet the previous polls were based on different methodology. Is it a case of comparing apples and oranges? Am sure as a scientist you know what i mean.

I also remember that in your first polls you claimed that most of those polled had little likelihood of changing their minds before the elections. How then the drastic changes?

Last question, why could you not provide poll results using the two different sampling methods (registered voters and population size) so that the diference can be quantified? And according to Steadman which polls are correct, is it the first ones which used population size or the latter ones which used registered voter base... Why? And if so does that mean one set of results are inacurrate?

Please answer these questions objectively despite your local politics in this blog...

Thank you.

Paul Robert

Anonymous said...

To Alex,

You are only consistent in your inconsistency. It is simple, answer the issues objectively and stop taking us round. We are not ignorant as you may want to think. We have our own minds please.

Mundumwami said...

Lets all be honest in such times even objective press houses sometimes get biased. Lets ask what can make Steadman biased?:
For one its owned by Joe Wanjui who is a major financier of Kazi iendelee probably using profits from steadman. Now if the profits are pro Kibaki Why not ITS opinion for the time bieng so that its kazi iendelee. You see we may really not know what these guys thought when they coined this slogan

Some opinion i came across frm a biased odm supporter.

Turn- out

Province Registered voters Turnout% Turnout No.s

Rift Valley 3,358,285 70% 2,350,800

Eastern 2,516,998 70% 1,761,799

Central 2,186,315 99% 2,164,252

Nyanza 2,041,686 70% 1,429,180

Western 1,564,854 55% 1,048,452

Nairobi 1,275,021 55% 701,262

Coast 1,045,629 55% 575,096

North Eastern 315,664 65% 205,182


Voter turnout increases by 10% across the board from 2002 figures.

Turnout in Central increases to 99% due to efforts by the government to get even those who are sick in hospital beds to vote.

Nyanza turnout matches Rift Valley and Eastern i.e increase by 15% from 2002 as much more is at stake in this election.

Vote share

Percentage Number of votes

Province RO MK KM RO MK KM

Rift Valley 65% 30% 5% 1,528,020 705,240 117,540

Eastern 10% 40% 50% 176,190 704,759 880,949

Central 0% 100% 0% 0 2,164,452 0

Nyanza 90% 8% 2% 1,286,262 114,334 28,584

Western 85% 10% 5% 891,184 104,845 52,423

Nairobi 55% 35% 10% 385,694 245,442 70,126

Coast 60% 30% 10% 345,058 172,529 57,510

North Eastern 75% 25% 0 153,886 51,295 0 TOTALS 4,766,294 4,262,897 1,207,131


Results,Raila 47%, Kibaki 42%, Kalonzo 12%.

Raila beats Kibaki by more than 500,000 votes.


Kibaki improves on Banana vote in Rift Valley by 6% due to Moi’s efforts (Banana got 24% in referendum.

Raila gets only upper eastern (Muslim votes in Eastern). Kibaki gets all Meru/Embu votes. Kalonzo all Kamba votes.

Negative propaganda in Central ensures that Raila is completely locked out of Central. (All polls at the moment show Raila at 5% in Central).

Kibaki gets 50% of Kisii votes. Most opinion polls have Raila getting 75% of Kisii Nyanza support. (NB Turnout in Kisii Nyanza will be much lower than Luo Nyanza).

Kibaki manages to get 25% and above in 6 out of 8 provinces. Kibaki gets 25% in Coast and North Eastern despite being anti-Majimbo and being perceived to be anti-islam.

Kibaki manages to get 33% of the Bukusu vote despite the Dini ya Musambwa prophecy in Western that is all the rage at the moment.

Bill Clinton said...

EXAMPLE 30% of 1000 people and 60% of 2000 people is not simply 45% overall. It is (300+1200)/3000 = 50%. IF ODM CANNOT DO SIMPLE MATH HOW CAN THEY LEAD A NATION ??

Anonymous said...

province registered votes raila % kibaki % kalonzo % others raila % kibaki % kalonzo % others
RIFT VALLEY 3,358,285.00 55% 39% 4% 2% 1,847,057 1,309,731 134,331 67,166
EASTERN 2,516,998.00 6% 49% 44% 1% 151,020 1,233,329 1,107,479 25,170
CENTRAL 2,186,315.00 5% 92% 2% 1% 109,316 2,011,410 43,726 21,863
NYANZA 2,041,686.00 86% 2% 2% 10% 1,755,850 40,834 40,834 204,169
WESTERN 1,564,854.00 73% 21% 1% 5% 1,142,343 328,619 15,649 78,243
NAIROBI 1,275,021.00 40% 46% 2% 12% 510,008 586,510 25,500 153,003
COAST 1,045,629.00 51% 35% 10% 4% 533,271 365,970 104,563 41,825
N.EASTERN 315,664.00 65% 32% 3% 0% 205,182 101,012 9,470 -
14,304,452.00 6,254,047 5,977,415 1,481,552 591,438

bloggeratti said...

I'm neither a mathematical genius nor statistics prodigy, but perhaps we may be ignoring one simple fact: That Emilio enjoys pockets of support that are slowly emerging in hitherto ODM "strongholds."

May I remind you that your dealing with a more informed and unforgiving voter who won't be easily keeled over by flowery promises?

My humble opinion is that Hon. Raila and Co. went overboard on the promises - and you know the adage "when the deal is too good, think twice!"
One other possible reason which I overheard is that ODM lost some serious mileage in Rift Valley after the fiasco that was their primaries.

My advice? Don't look at the figures ONLY - there might be a human element that you are ignoring.

Anonymous said...


opinion polling is a science and captures the opinions of the people polled at that particular time regarding the questions asked.

I dont know what you mean when you tell us not to look at the figures only since opinion polls are done especially to quantify those human elements you talk about. As in if people are dissatisfied with Mr. Kibaki or Mr. Odinga the opinion poll will show this in terms of falling poll numbers.

Opinion polls if done correctly do not force the respondents to make a choice and it correctly polls the choices made by the respondents... Opinion polls will also show the amount of undecided and the number error ratio.

It is a well researched and proven science really...

KenyanGal said...

I really dont understand wht's with all the complaints against Steadman polls. ODM fanatics should know Steadman made Raila popular;- It's the Steadman poll that INFORMED us that Raila is leading. So dont complain to Steadman, wait for the real exam on 27th Dec. U have ur voters cards, dont you?!!!


Related Posts with Thumbnails