Professional Documents
Culture Documents
Suppose that a large mobile phone company hires a business research company to
estimate the average number of text messages used per month by Canadians in the 35-
to-54-year age bracket.
If the mobile phone company uses the sample mean of 1,300 texts as an estimate for the
population mean, the same sample mean is used as a point estimate.
A point estimate vs interval estimate
Central Limit Theorem: if sample size is large, regardless of the shape of the
population distribution: (if is known)
Rearranging this:
Confidence Interval
Confidence Interval
Confidence Interval
• In most cases, if we don’t know the mean of the population and want to
estimate it, we don’t know the standard deviation of the population either.
• So, the sample standard deviation should be used instead.
• The z formulas are inappropriate when SD of the population is unknown
• T test: developed by a British statistician, William S. Gosset.
• Sample size is too small; n < 30
• No standard deviation
• Flatter and wider confidence level for same value
• Shape of T-distribution is based on degrees of freedom
The t distribution
• The t formula:
Let’s solve problem 8.13
90% CI
Let’s solve problem 8.35
8.35 Solution
Sample=200
Sample=114
Sample=57
Let’s solve problem 8.41
Let’s solve problem 8.41
Sample=2522
Sample=601
Sample=268
Sample=16577