segunda-feira, outubro 27, 2008

Vale a pena ler

James Stimson, sobre os resultados das sondagens nos Estados Unidos:

Saturday 10/25: Stability and Variability
Variability: This is a race of considerable variability in various organization's estimates of what should be the same quantity. And at the same time I have never seen such stability in my estimates of the daily lead. A typical day sees about ten organizations report an Obama lead varying between 1 and 14 points. Thirteen points difference is a lot, more than double what would be expected from sampling fluctuation alone. This arises chiefly, it appears, from two sources, (1) initial assumptions about the partisan makeup of the electorate, and (2) varying likely voter assumptions.

Ler o resto aqui.
Enviar um comentário