Confidence Levels: If the governor election were held today,

[rich]

I am looking for the 95% confidence interval estimate of the true population for the following:

If the governor election were held today, would you vote for Martinez or Pajcic?

Matinez : 43%
Pajcic: 40%
Undecided: 14%
Don't know: 8%

The sample population consists of 959 voters.

...................................................................

Here is what I have concluded so far:

The true value would be Martinez since he has the greater percentage of votes so far. And, since it is a 95% confidence level, the Z level must be 1.96.

0.43 (959) = 412 +/- 1.96 [(0.43)(1-0.43)/959]^2 = 412+/- 6.51

Therefore, the upper limit of the interval = 412 + 6.51
and, the lower limit of the interval = 412 - 6.51

.................................................................

Can anyone tell me if this accurate, please? Thank you in advance.

 

[JakeD]

Re: Confidence Levels: If the governor election were held to

I am looking for the 95% confidence interval estimate of the true population for the following:

If the governor election were held today, would you vote for Martinez or Pajcic?

Matinez : 43%
Pajcic: 40%
Undecided: 14%
Don't know: 8%

The sample population consists of 959 voters.

...................................................................

Here is what I have concluded so far:

The true value would be Martinez since he has the greater percentage of votes so far. And, since it is a 95% confidence level, the Z level must be 1.96.

0.43 (959) = 412 +/- 1.96 [(0.43)(1-0.43)/959]^2 = 412+/- 6.51

Therefore, the upper limit of the interval = 412 + 6.51
and, the lower limit of the interval = 412 - 6.51

.................................................................

Can anyone tell me if this accurate, please? Thank you in advance.

It's not accurate. With a sample size of 959, the margin of error should be around 3 percentage points (.03x959).The formula you show is not correct: there should be a square root, not square.

BTW, should you be counting the undecided/don't knows?

 

[rich]

What would the correct formula be?

I don't think I should be counting the undecided and don't knows because I am looking for the true population. But, I'm not sure.

 

[JakeD]

What would the correct formula be?

I don't think I should be counting the undecided and don't knows because I am looking for the true population. But, I'm not sure.

The margin of error appiled to proportions such as .43 is \(\displaystyle \L \pm 1.96\sqrt{p(1-p)/n .\)