7.5 7.5 7.7 7.7 (a) (b) 7.7 (c) • Would you conclude that sales are up at all stores?
• What does Ha say?
• Does it say anything about ALL stores? 7.36 (a) 7.36 (a) 7.36 (b) 7.36 (c) • No outliers, but right skew in distribution of spending amounts
• Is it still OK to use the t-distribution?
• The MAJOR requirement for using T-test
is… 7.54 7.54 (b) 7.54 (b) • The interval (0.02,2.78)—What does it represent? • If we gave flat screen monitors to everyone, the mean amount that they would rate them, minus the mean rating of the regular monitors • In other words, it is our estimate for μ1-μ2 • We are 95% confident in this estimate • The true difference of the means could be higher or lower 7.54 (c) • Ho: μ1=μ2 • Ho: μ1≠μ2 • Alpha=0.05
• If we subtract μ2 from both sides of the equality and inequality
above, what do we get? • Ho: μ1-μ2=μ2-μ2 leads to Ho: μ1-μ2 =0 and • Ha: μ1-μ2≠μ2-μ2 leads to Ha: μ1- μ2 ≠ 0 • So we want to know if μ1- μ2 =0 and we want to be 95% confident in the answer • The estimate from the 95% confidence interval for μ1-μ2=(0.02,2.78) • So… Probability Set-Up is The Same
• Confidence Interval • Hypothesis test
1. Mu may be above 2.78 or below 1. We reject Ho for Ha if (xbar1- 0.02 xbar2) is either very big or very small 2. If Ho is true, then we will make a 2. So error could occur above or mistake if (xbar1-xbar2) is either below estimate of mu1-mu2 very big or very small 3. Risk of this mistake is 5%, so 95% 3. Confidence is 95% so risk is 5% confident 7.63 7.63 7.64 7.64 (a & b) 7.64 (c)
• If we reject Ho for Ha, what are we saying?
• What does that have to do with the statement above? 7.64 (d) • Say we reject Ho for Ha • What have we learned?
• What is a next natural question you would
ask if you were actually looking for an apartment?