Confidence Intervals for the population proportion (when n is large)

 

Suppose that you want to make a 95% confidence interval for the true proportion of male college students that binge drink.

A 2001 study by the Harvard School of Public Health showed that out of 3925 college males, 1908 of

them were binge drinkers.

 

A confidence interval with the proportion uses the z, standard normal distribution.

  1. To do this, enter your data into C1. Then go to Stat, Basic Statistics, 1 proportion.
  2. Your screen should look like this.

  1. Enter the summarized data. Enter n into the number of trials and enter the number of successes.


 

  1. Make sure that you entered the correct confidence level.
  2. N ext pick the type of confidence interval or bound.
    1. If you want a confidence interval select, "not equal" as your alternative.
    2. If you want a confidence bound, select "less than" for a upper bound or "greater than" for an lower bound.
  3. Make sure that the box is checked for "Use test and interval based on normal distribution".
  4. Your output may look like the following.

 Test and CI for One Proportion

 

Test of p = 0.5 vs p not = 0.5

 

 

Sample     X     N  Sample p         95% CI         Z-Value  P-Value

1       1908  3925  0.486115  (0.470478, 0.501751)    -1.74    0.082