Location: MGSC 291

Discussion: Helpful Hints from Chapter 9Reported This is a featured thread

Showing 1 post

DayneLee
Helpful Hints from Chapter 9
Apr 20 2011, 10:33 PM EDT | Post edited: Apr 20 2011, 10:33 PM EDT
Test Statistic (z vs. t)
• When population standard deviation is known, solve for z and use the z-table.
• When population standard deviation is unknown (s will be given), solve for t and use the t-table

p-value Approach
1. Solve for the test statistic (z or t) using your sample mean (x-bar)
2. Solve for the p-value using z
a. If z is positive, subtract the % below the z-score from 1 to get the p-value
b. If z is negative, keep the % below the z-score from the table as the p-value
c. In either case, if it is a two-tail test, multiply the p-value by 2
3. Solve for the p-value using t
a. Find the 2 values for area in the upper tail that your t-value is between, those values compose the range for your p-value
b. If it is a two-tail test, multiply the 2 numbers of the range by 2
4. Compare p-value with alpha (level of significance)
a. If p-value is > alpha then do not reject the null
b. If p-value is < alpha then reject the null

Critical Value Approach
1. For a one-tail test, find the z-score that has the significance level below it.
2. If the z-value of your sample mean (x-bar) is positive, then compare to the positive form of the critical value
3. If the z-value of your sample mean (x-bar) is negative, then compare to the negative form of the critical value
4. For a two-tail test, divide the significance level by 2 and then determine the z-values that have that percent above and below
5. If the z-value for your sample mean (x-bar) is less than the negative critical value, reject the null
6. If the z-value for your sample mean (x-bar) is more than the positive critical value, reject the null
7. If the z-value for your sample mean (x-bar) is within/between the positive and negative forms of the critical value, do not reject the null

Do you find this valuable?    
Keyword tags: None

Sign in to be the first to reply.