polls and past reliability ref 1994

[bullimiami]

I have been wondering about how the polls did in 1994 when the repugs swept through the government. apparently they did NOT predict this monumental event.
With an historic turnout we can create out own non-predicted monumental event.




from a 1996 david tipton article

""However, ladies and gentlemen, I am not the president. I do not let polls determine how I act, and I'm especially not allowing them to determine for me whether and how I'll vote. I have my principles and beliefs, and I refuse to cede my vote to some polls that tell me what I'm supposed to do a month from now. It isn't like polls have never missed results before. Remember 1994? If you listened to the polls, the Democrats would take a beating, but they were still going to hold on to the House and Senate.

"Come on, nobody gets 40 seats in a midterm election, and when you Republicans factor in the seats you're going to lose, you just can't do it. Try again in 1996." There was only one major problem with that -- voters. The Republicans gained eight Senate seats, over 50 House Seats, and not a single Republican incumbent running for anything anywhere in America lost. That's unheard of. Polls aren't oracles, and I refuse to treat them as such.

There's one thing their polls can't predict -- the human factor. This far out, there's still a fair amount of the electorate that hasn't started paying attention yet. They won't until about two or three weeks before the election. And once that curtain shuts, once people get away from their families, friends and pollsters, once they stare at that list of candidates, you can throw a poll out the window because you can't predict what many of these undecideds will do.""

[jobycom]

1. Not accurate, that's comparing national polls to local polls

For a national election like the presidency, polling is more constant, more accurate, and done by many different organizations, so the results can be compared to each other, and a big picture can be seen.

In 1994 they were working off local polls in small towns and states across the country, often taken weeks before the election, and in many regions, there were no polls at all.

The pundits who were saying the Repubs would make strong gains but probably wouldn't take the House were going off conventional wisdom, not polls. The polls I remember showed the Republicans making great gains. There were upsets in some places were polls weren't even bothered with, like Thomas Foley's district, where he was considered automatic, and wasn't even watched closely.

On the other hand, in the high profile Senate races, where statewide polls had a better chance of measuring sentiment than local polls, the polls were very accurate.

I hate it when liberals and Democrats slip into this "Polls are BS" mode. I've heard the Republicans do that before every election they've lost, and it just makes them look clueless. Now we are the ones doing it. The polls are showing a statistical tie, a close race. The fact that they give Bush an edge doesn't overcome the fact that his edge is within the margin of error, that the incumbent scores lower in elections than in polls, and that there are millions of new voters that the polls aren't picking up, and that these new voters are breaking for Kerry.

What it all means are that the polls are correct right now. It's a close race. It could go either way. Most polls predict that, not dispute it. Kerry could win in a landslide and most polls will still be correct within the margin of error.

That doesn't mean that all the pundits and "journalists" in America intepret them right. Journalism, these days, is BS.