@innairobi - Actually the polarization is so bad that even an informed Kenyan like you or me would still have a problem coming up with a proper and representative sample.

Even a single area like Kibera, the distribution (by tribe for example) is not uniform. I am made to understand that different parts (of Kibera) are balkanized along tribal lines. There is an area (or areas) inhabited predominantly by luos, another by kambas, another by nubians another by kikuyus (?) etc. And worse still (from a sampling point of view) is the fact that the percentage of these groups is not equal. These is replicated in most informal settlements (and sometimes even in mainstream estates).

The only way to remedy this is to get samples from EVERY POLLING STATION and weight the number of respondents based on the number of registered voters. And this would still be prone to sampling errors due to the fact that the voter turn out will not be equal across the polling stations!

These Harvard pro's are Idle.

Somebody should request them to conduct a new "Poll"

Who is the president of kenya.

Raila would definately be leading and u guess it he will lead forever in "Exit Polls" either conducted by Harvard or Ambitho

Whatever people may say or think, the damage is already done.

Kenya is now in a stalemate position. Trust will be hard to find, and excellence will be stifled.

that ship has sailed....

Kwani hauna habari????