Suppose 36% of a university's alumni donated last year. What is the standard deviation of for random samples of 225 alumni? Give the exact value as a decimal.
The sampling distribution of a sample proportion
How behaves over repeated random samples: centered at , spread , and approximately normal once the success-failure condition holds.
Key ideas
- Sampling distribution
The distribution of a statistic over all possible random samples of the same size. Its observational units are samples: each dot is one sample's , not one person.
- Center (unbiasedness)
With random sampling the values of average out to . This holds for every sample size.
- Variability
The standard deviation of shrinks as grows (by a factor of ) and is largest when .
- Shape
Skewed for small samples when is near 0 or 1. Approximately normal once and (the success-failure condition).
- Sample versus sampling distribution
A sample distribution shows one sample's individual values. A sampling distribution shows a statistic across many samples. The number of simulated samples is not ; more simulations only sharpen the picture.
Formulas
Backyard gardens
In a large city, 30% of households have a backyard vegetable garden. Describe the sampling distribution of for random samples of 100 households, then explain what changes for samples of 400.
Center: for both sample sizes. Random sampling targets the parameter whatever is.
Spread at n = 100: .
Shape at n = 100: and are both at least 10, so approximately normal: .
At n = 400: the spread is , exactly half, because grew by a factor of 4 and . Center unchanged; shape still normal.
at and at : same center, half the spread.
Check yourself
A pollster switches from samples of 300 voters to samples of 1,200 voters from the same population. What happens to the standard deviation of the possible values?
For random samples from a population with , select every feature of the sampling distribution of that depends on the sample size.
About 4% of a manufacturer's circuit boards are defective. For random samples of 150 boards, the sampling distribution of is most likely:
Quantity A: the SD of for samples of 200 when . Quantity B: the SD of for samples of 200 when . Which is true?
A simulated sampling distribution of for random samples of 50 households is centered at 0.42. What is the best estimate of the population proportion ?