Using A Statistics Formula To Establish Confidence Intervals of a
Poll
When political polls are conducted, a confidence interval is always
included. This confidence interval is referred to as a
margin of error. For example,
Rassmussun Poll Data showed Arnold Schwarzenegger trailing State
Treasurer Phil Angelides by one percentage point, 45% to 44% in the
California governors race, where the sampling error for the survey was
+/ 4.5 percentage points at the midpoint with a 95% level of
confidence. Given the 4.5% error, the 1% edge is insignificant. A
respected pollster like Rassmussun will use
proper survey methods to insure a good sample. The mathematics and
statistics used to establish the confidence interval is fairly
straightforward, but involves a somewhat large formula.
The Confidence Interval Formula
In this formula, each variable stands for a quantity obtained from
the survey or a calculated quantity. For example, n would be the
sample size, and z_{c} would be a zscored calculated using
another formula. You are basically substituting into several different
formulas to obtain the results. A student with little or no algebra
background would find this formula a bit intimidating. A student
completing several years of algebra would not have difficulty working
this formula however. An example where this formula is used
is given here.
I think you can see why the average first year introductory
statistics course in college
requires 1 or 2 years of algebra. And, statistics is
required for most degrees, even nonscience and nonmath degrees.
