'Random adults' and ANWR

Allahpundit notes a 10-point disparity between a Gallup survey and a Rasmussen survey on the question of drilling ANWR. Observing that the two surveys were taking a month apart, Allah says this "probably means opinion on ANWR has shifted somewhat too."

No, it means Gallup polled "random adults." Look:

Results for the Gallup Panel study are based on telephone interviews with 1,013 national adults, aged 18 and older . . .

Now, notice that Rasmussen polled "1,000 likely votes." This undoubtedly accounts for all the difference in the two surveys.

Compared to "random adults," actual voters tend to be older, more educated, and with higher income. "Likely voters" pay more attention to current events and public affairs than do "random adults."

One of the most dishonest things that polling firms and news organizations do is to ask political questions in surveys of "random adults," then report those results as if they were politically significant, when they are not. People who don't vote, don't count when it comes to politics. The disinvolved and uninformed tend to be more liberal than actual voters; pollsters and media know this, and thus use "random adult" surveys as propaganda to make liberal policies seem more popular politically than they actually are.