| If you're working with a small sample (less than about | | | | The hypothesis in this example is one-tailed; that is, |
| 30 or 40) in Microsoft Excel, you can use the Student's | | | | you're interested in finding probabilities of values less |
| t-test instead of the z-value or z-score to find the | | | | than 16. If instead you need to find the probabilities of |
| probability with which a value falls below a certain | | | | values both above and below x, your hypothesis is |
| number or to test how far an individual observation is | | | | two-tailed. |
| from the mean. To doso, you use the TINV function. | | | | Using the Excel TINV Function |
| Using the Excel TDIST Function | | | | If you know the probability and want to find the t-value, |
| You can use the TDIST function to make inferences | | | | use the TINV function. This function has the following |
| about the value of a population mean. | | | | syntax: |
| For example, if you randomly select 20 people from a | | | | =TINV (probability, degrees of freedom) |
| factory floor, ask them to try a new production | | | | If this is based on a one-tailed t distribution, multiply the |
| method, and then find that they can produce 17.25 units | | | | probability by 2. |
| an hour with a sample standard deviation of 3.3, you | | | | Using the Excel TTEST Function |
| can find the probability that the population mean takes | | | | To find the probability associated with a Student's |
| the value of 16 or less. To do so, you use Excel's | | | | t-test, use the TTEST function. The t-test is most |
| TDIST function. The function uses the following syntax: | | | | frequently used to test for a difference between two |
| =TDIST(x, degrees of freedom, tails) | | | | means. The TTEST function uses the following syntax: |
| For this example, the function takes the following form: | | | | =TTEST (data set 1,data set 2,tails,type) |
| =TDIST (16,19,1) | | | | Where type equals 1 for paired, 2 for two samples |
| Depending on your level of significance, you accept or | | | | with equal variance, or 3 for 2 samples with unequal |
| reject the hypothesis. | | | | variance. |