« previous next »
Pages: 1 [2]

Author Topic: IPSOS Witchcraft Released  (Read 4739 times)

Offline Kim Jong-Un's Pajama Pants

  • Moderator
  • Enigma
  • *
  • Posts: 8783
  • Reputation: 106254
  • An oryctolagus cuniculus is feeding on my couch
Re: IPSOS Witchcraft Released
« Reply #20 on: July 23, 2017, 11:34:17 PM »
There is no evidence that Kenyans lie in polls.  If I support, kamwana, I am going to say so to a pollster.  Same thing for Raila.  But more so, kamwana.  Why would I lie and make it easier for Raila to reject the outcome?
"I freed a thousand slaves.  I could have freed a thousand more if only they knew they were slaves."

Harriet Tubman

Offline RV Pundit

  • Moderator
  • Enigma
  • *
  • Posts: 38334
  • Reputation: 1074446
Re: IPSOS Witchcraft Released
« Reply #21 on: July 24, 2017, 06:59:01 AM »
The problem I see with our polls is their inability to factor tribe - the overriding factor - esp when sampling in cosmopolitans areas like Nairobi & Rift Valley. The tribal breakdown data is available from census. They need to use it.

If you look at Synovate Ipsos & Iposos- here - they do a decent job - except in Nairobi & Rift Valley - which are cosmopolitan. Synovate has Raila at 64% of Nairobi and uhuru at 27% which is improbable. Infotrak get RV completely wrong...nearly 35% for Raila - with 55% for Uhuru.

http://www.ipsos.co.ke/report_downloads/downloads.php?dir=polls&file=Ipsos_KE_SPEC%20Poll_Press%20Release%20Presentation_%2023rd%20July%202017.pdf
https://pbs.twimg.com/media/DFbeedeW0AIOiqH.jpg:large
Pages: 1 [2]
« previous next »