diff --git a/book/_build/.doctrees/environment.pickle b/book/_build/.doctrees/environment.pickle index 55abbd9..c48b259 100644 Binary files a/book/_build/.doctrees/environment.pickle and b/book/_build/.doctrees/environment.pickle differ diff --git a/book/_build/.doctrees/sampling.doctree b/book/_build/.doctrees/sampling.doctree index c41e062..4b77fe0 100644 Binary files a/book/_build/.doctrees/sampling.doctree and b/book/_build/.doctrees/sampling.doctree differ diff --git a/book/_build/html/_sources/sampling.md b/book/_build/html/_sources/sampling.md index 2a939e6..cecb0bf 100644 --- a/book/_build/html/_sources/sampling.md +++ b/book/_build/html/_sources/sampling.md @@ -188,7 +188,7 @@ Thus, the SE for the percentage of members with an Apple Watch based on the samp **Adjustment 2**. The SE formulas we first learned assume draws are made *with replacement*. Simple random samples are done without replacement. While we found SE to be about 4.9%, we should note that if a sample of 100 members were done without replacement, we'd have sampled the entire population and there would be no variability in the sample percentage. We'd always find 40% of members have an Apple Watch. This reveals that sampling without replacement actually has a lower associated standard error. The is corrected by a correction factor: -$$\text{SE drawing without replacement = correction factor \times SE drawing with replacement}.$$ +$$\text{SE drawing without replacement = correction factor} \times \text{SE drawing with replacement}.$$ And the correction factor is diff --git a/book/_build/html/sampling.html b/book/_build/html/sampling.html index fe5080d..a112038 100644 --- a/book/_build/html/sampling.html +++ b/book/_build/html/sampling.html @@ -631,7 +631,7 @@
Adjustment 2. The SE formulas we first learned assume draws are made with replacement. Simple random samples are done without replacement. While we found SE to be about 4.9%, we should note that if a sample of 100 members were done without replacement, we’d have sampled the entire population and there would be no variability in the sample percentage. We’d always find 40% of members have an Apple Watch. This reveals that sampling without replacement actually has a lower associated standard error. The is corrected by a correction factor:
And the correction factor is