Daily Presidential Tracking Poll Sunday, March 21, 2010: Hussein at -16, Health Takeover at -19

The Rasmussen Reports daily Presidential Tracking Poll for Sunday shows that 26% of the nation's voters Strongly Approve of the way that Barack Obama is performing his role as President. Forty-two percent (42%) Strongly Disapprove giving Obama a Presidential Approval Index rating of -16 (see trends).

As the House prepares to vote on the health care plan proposed by the President and Congressional Democrats, just 41% of voters favor the plan while 54% are opposed. Those figures include 26% who Strongly Favor the plan and 45% who are Strongly Opposed. Most voters believe it will raise the cost of health care and reduce the quality of care. Still, nearly two-out-of-three voters believe it is at least somewhat likely to pass and become law.

>> Weekends do produce different results <<

Not really very much. When I look at the data going back several months, I can't detect any systematic, repeating weekend effect of more than a point or so -- if that. The probable reason for this felicitous outcome is that by weighting his samples in terms of D's and R's, Rasmussen eliminates most of the "weekend bias" that otherwise would favor the Dhimmis.

In other words, today's huge pro-0bama bounce probably reflects nothing more than a very "favorable" -- but random -- Saturday night poll for the Big Ø, just the kind of statistical noise that always accompanies small-sample polling. To get a six-point swing in this morning's reported Index value, last night's raw data probably saw something in the neighborhood of an 18-point swing -- a magnitude that can't possibly reflect a true, underlying change in public opinion.

Moreover, Rasmussen's "three-day methodology" means that last night's results will continue to influence the Index report for the next two mornings. So look for the value to stay somewhere near its present -16 for the next two days.

Next, my odds calculator says the value ought to click down a point or two on Wednesday morning, bringing it closer to a true, underlying value of ca. -17 or -18.

On the other hand, if we DON'T see a down-tick on Wednesday, it will probably signal that the Annoited One has rallied his base once again. But also as in the past, any such rally will be temporary.

Forecast for tomorrow:

60% probability that the Index is at -16 or higher.

40% probability that the Index declines to -17 or lower.

Other info:

I just did a linear regression on the Index for the 44 days since February 6 -- the point at which a new trend appears to have begun. This calculation yields a trend value for today at -18.3, whereas the 10-day moving average stands this morning at -17.5.

Which method will have greater predictive power? Yet to be determined. So in the meantime, "You pays your money and you takes your choice."

An exponential or log-linear regression might have better predictive power than a straight linear regression. But I don't remember techniques for calculating either of the former two. Need to get out the old statistics 201 textbook and study up!

>> The Sunday numbers have shown a ‘Bama game almost every week. <<

Sorry, but your opinion seems not to be supported by the data.

In order to analyse the matter, let’s remember first that what Rasmussen gives out every morning is a three-day moving average incorporating “raw data” from the previous three nights’ polling. So if there’s indeed a weekend effect, it should show up most strongly in the Monday morning reports.

Therefore, to seek evidence of weekend bias in Rasmussen’s Index, I’ve looked at numbers covering the past 13 weeks’ worth of Monday morning reports. I’ve compared these reports both with (a) the ten-day moving average on each Monday morning and with (b) the previous Friday mornings’ reports — the latter consisting of raw data from three non-weekend nights.

Findings:

1. In seven cases out of 13, the Monday morning report showed an Index value BELOW that morning’s ten-day average, that is, a movement against 0bama.

2. In eight cases out of 13, the Monday morning report also showed an Index value BELOW the preceding Friday morning’s report, another series of movements unfavorable to 0bama.

Conclusion:

By both of my measurement standards, during the most recent quarter there has been nothing approaching a statistically significant weekend effect.

Now there’s always more than one way to skin a cat. So if you or anybody else has a statistical analysis with a contrary outcome, please let’s compare notes.

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.