Nah, they're simply not as far off as they usually are... LA Times would rather argue its a "dirty campaign" bounce than a convention bounce.
The decoder key to LA Times polls are their partisan sampling... This one shows a decent comparison of Republicans to Demonrats.
Thanks for posting the sampling results. I refuse to register with them, so I was wondering if it was 14% Dim heavy like a few moths ago.
Down Goes Kerry! Down Goes Kerry!!Down Goes Kerry!
Here are the LA Times August 21st-24th, 2004 published poll internal numbers for political affiliation for the three-way race between Bush, Kerry and Nader. These LA Times poll data are in a table format for easy reading.
This national three-way race poll which had 1,352 Registered voters, now shows Bush in the lead by 3% with Bush 47%, Kerry 44%, Nader 3%. The MoE is ±3.0% for this LA Times poll question number six.
Source: LA Times Presidential poll - August 21-24, 2004 PDF file, released August 25th, 2004: Question #6, page 9 of 25.
Here below is the distribution of the 1,352 registered voters necessary to recreate the numbers show in the table directly above. Remember that the numbers below must be rounded to NO decimal places to match the above LA Times published data.
So the partisan affiliation weightings that were used by the LA Times are indicated in the table below. Remember, because of LA Times rounding of their poll numbers, the below breakdowns might have a ±1.2% variance. That being said, it appears that this latest LA Times three-way race presidential preference poll has 9.1% more Democrats than Republicans in its sample population. A 2%-3% advantage for Democrats over Republicans might be a more realistic set of demographic splits.
So what would the LA Times poll look like with a more reasonable national sample of say 35.1%(R), 37.1%(D), 27.8%(I)? Look below and you would see that Bush would now have a 7.8% lead, which would be clearly outside the 'Margin of Error' of the LA Times poll. Kerry has been sinking swiftly...
Hope this helps...
dvwjr