By Mary Mostert, Analyst, Original Sources
Sepbember 17, 1998
On Tuesday I analyzed two "polls"; - one the Los Angeles Times poll that claimed "Clinton's Leadership Ratings Strong Despite Starr Report" and the other the thousands of copies of e-mails to members of Congress that have come into my computer in the past few days.
The e-mail I am getting show a tidal wave of anger and demands from a wide variety of people, many of them who identify themselves as Democrats, that demand impeachment or resignation of President Clinton.
The Los Angeles Poll showed a majority still strongly supportive of the President. I also pointed out that the unreported aspects of this story have to do with the thousands of e-mails going to members of Congress. It's taken about a week, but one-by-one Democrats in Congress are beginning to call for the President's resignation. I suspect that is because they are being deluged with mail by Democrats who are demanding they take a stand. As one e-mail I read said, "If Bill Clinton is still in the White House in November, YOU won't be going back to the Congress in January!"
Those supporters appear to be in the Black Caucus, people like Patricia Ireland of the National Organization for Women, and among the abortion proponents. Was the LA Times poll rigged to question an unrepresentative sample, I wonders? Only 968 people nationwide were polled - supposedly via "scientific" methods. By the time I wrote the article on Monday evening, I already had seen 4000 irate e-mails demanding impeachment.
My column prompted one of my readers, T.A. Rodda, to do a bit of mathematical computation based on the reported results of the poll. It clearly showed a skewed sampling. So, I asked him to share how he did it. He responded, "Actually, the math I used for the poll calculations are simple algebra. I am certainly no whiz at math, but was intrigued at the prospect of a brain teaser (solving for three unknowns)."
Rodda lives in the Los Angeles area, and, he tells me, has "worked with computers my whole adult life." He and his wife have two boys (17 and 11) and "my 11 year-old, however, can be heard to shout "Impeach Clinton" on occasion (he listened seriously to the charges and decided for himself).
While the LA Times report gave no racial make-up of the sampling they took, it did give results by political affiliation, that is, the percentage of Republicans, Democrats and Independents responding to the poll. I asked him to send it to me - because I thought it showed quite a significant make-up. While we don't know how many of each race was in the poll - we do know, for instance, that blacks make up 13% of the American population. We also know that they are strongly supportive of the president. Almost 50% of the votes in Congress last Friday to deny public access to the Starr Report came from the Congressional Black Caucus. Was the Los Angeles poll weighted heavily with blacks? Or perhaps, with women who support abortion on demand? Or Democrats?
How can it be that 968 people came up with a tally so contrary to what we are all hearing among our friends and neighbors?
Rodda's calculations, which he was kind enough to share with me are as follows: "From your article, I found that 968 people were polled. From the poll numbers, I can create three equations that contain three unknowns. The first equation is the easiest, where the sum of Republicans, Democrats, and Independents equal all the people in the poll:
"968 = D + R + I (968 polled, D=Democrat, R=Republican, I=Iindependent)
"The second equation comes from the sixth poll question (Clinton's job approval). The percentage of Democrats who answered favorably was 80%, Republicans 43%, Independents 61%, and the total for all favorable answers was 64% (64% of 968 equals 620). These values form this equation:
"620 = .8D + .43R + .61I
"Two equations down, but we need a third. The unfavorable responses to the question came to 18% of democrats, 54% of republicans, 35% of independents, and the total unfavorable was 33% of 968, or 319. Thus the equation:
"319 = .18D + .54R + .35I
"Now, to solve for three unknowns, we must reduce the unknowns to two. Therefore, we solve one equation for I:
"I = 968 - D - R
"Substitute the value of I into the second equation:
"620 = .8D + .43R + .61(968 -D -R) Simplify:
620 = .8D + .43R + .590.48 -.61D - .61R
620 = .19D -.18R + 590.48
29.52 = .19D -.18R
"Now substitute the value of I into the third equation:
"319 = .18D + .54R +.35(968 -D -R) Simplify:
"319 = .18D + .54R + 338.8 -.35D -.35R
"319 = .19R - .17D + 338.8
"19.8 = .19R - .17D
"There are now two equations and two variables. We add the two equations, making sure one variable is eliminated in the process. I multiplied the first equation by ..17 (the coefficient of D in the second equation), and the second equation by .19 (the coefficient of D in the first equation) to achieve the following equations:
5.0184 = .0323D - .0306R
"-3.366 = -.0323D + .0361R
"After addition:
1.6524 = .0055R Simplify:
R = 300 (After rounding)
"Now plug this value into one of the final equations:
29.52 = .19D -.18(300) Simplify and solve for D:
29.52 = .19D -54.078
84 = .19D
D = 442 (after rounding).
"Back to the equation for I:
I = 968 - 442 - 300
I = 226
"Therefore, Republicans = 300, Democrats = 442, and Independents = 226
"For men (M) and women (W):
"968 = M + W (obvious)
W = 968 - M (again, obvious)
"From poll question 6:
620 = .67W + .60M (Substitute for W):
620 = .67(968 - M) + .60M
620 = 648.56 - .07M
28.56 = .07M
M = 408
"Therefore:
"W = 968-408 = 560
"So, the math indicates, based on the reported results and the percentages answering the questions, the LA Times Poll participants were:
"Women - 560 - 58% of the poll
"Men - 408 - 42% of the poll
"Democrats - 442 - 46%
"Republicans - 300 - 31%
"Independents - 226 - 23%
If Republicans represent a mere 31% of the people of America, one wonders how they managed to become a majority in Congress. As for the disparity between men and women, the Census figures show that there 51% of the population is female, and 49% is male -a two percentage point difference, not a sixteen percentage point difference.
It all reminds me of one of my Mother's favorite pithy sayings about questionable statistics: "Figures don't lie, but liars can figure."
The elite media have not told the American people the truth about Paula Jones, who actually created the government shut down in 1995, Welfare Reform, Newt Gingrich, or why the economy improved so much under Ronald Reagan. Why would we think now we are getting honest polls about Bill Clinton when the whole left wing governance of America appears to depend on his survival?
To comment: mmostert@originalsources.com