So it turns out that others are also concerned about the differences between the Nanos poll and the Decima poll. The Ottawa Citizen has run a story about it today.
2 differences apparently account for the gap. One, Decima prompts people with a choice of who they are supporting whereas Nanos asks an open ended question – meaning the responding has to choose. The second is more interesting:
Nanos pays more to get cellphone exchanges included in its calling list. Cell users tend to be young and more transient than those with land-lines, Mr. Nanos says.
So, Nanos sample is actually more representative. Decima is going to be skewed towards older voters. What is more interesting is that this problem isn’t a new one…