Scribd
Upload
44 views

Taking Perplexityoutof PValues

This document discusses how to interpret p-values from statistical tests in Minitab. It explains that a p-value indicates the probability of obtaining results at least as extreme as the observed results of a statistical test, given that the null hypothesis is true. P-values are compared to a predetermined significance level alpha to determine whether to reject the null hypothesis. Lower p-values provide more evidence against the null hypothesis. The document provides an example of interpreting the p-value from a t-test and explains how p-values are used and interpreted for other common statistical tests like regression, normality tests, and tests of independence.

Uploaded by

rohitas83
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
44 views

Taking Perplexityoutof PValues

This document discusses how to interpret p-values from statistical tests in Minitab. It explains that a p-value indicates the probability of obtaining results at least as extreme as the observed results of a statistical test, given that the null hypothesis is true. P-values are compared to a predetermined significance level alpha to determine whether to reject the null hypothesis. Lower p-values provide more evidence against the null hypothesis. The document provides an example of interpreting the p-value from a t-test and explains how p-values are used and interpreted for other common statistical tests like regression, normality tests, and tests of independence.

Uploaded by

rohitas83
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Technical Support Document

Taking the Perplexity out of P-Values


In Minitab, the conclusion for many statistical tests is reported in terms of the p-value, rather than in terms of rejecting or failing to reject the null hypothesis. Minitab technical support frequently receives calls from users who arent aware that they can use the p-value to decide whether or not to reject the null hypothesis. In a basic t-test, for example, the null hypothesis, H0, and the alternative hypotheses, HA, are usually H0: = 0 and HA: 0, where (mu) is the population mean for the group under study and where 0 is the hypothesized population mean. Lets say you want to determine if a new gasoline additive has an effect on gas mileage. If the known gas mileage for this particular class of cars is 25 miles per gallon (m.p.g.), the hypotheses for this study are H0: = 25 and HA: 25. You test 35 cars and find that the m.p.g.s range from 14.4 to 28.8. After putting the data in the column MPG, you run Minitabs t-test (the menu command Stat > Basic Statistics > 1-Sample t, or the session command TTEST) and get these results: Output from T-Test Test of mu = 25.000 vs mu not = 25.000 Variable N Mean StDev SE Mean MPG 35 23.754 3.156 0.533

T -2.34

P-Value 0.026

Now What? Interpreting the P-Value The results show that the mean of the 35-car sample is 23.754. But the mean m.p.g. of all cars of this type () may still be 25you need to know if there is enough sample evidence to reject H0. There are two usual ways to determine that, and both involve the significance level, (alpha). is the probability of rejecting H0 when H0 is true. In this case, thats the probability of concluding that the population mean is NOT 25 m.p.g., when in fact it is. The first way to determine if you have enough sample evidence to reject H0 is to compute the value of the test statistic and compare it to a value from the appropriate statistical table at a pre-specified value of . The second method is to compute the p-value, often called the attained significance level of the test, and compare it to . The p-value is an indication of how well the sample evidence supports the rejection of H0. In general, the smaller the p-value, the heavier the weight of the sample evidence for rejecting H0. More specifically, the p-value is the smallest value of that results in the rejection of H0. For any value of < pvalue, fail to reject H0, and for any value of p-value, reject H0.

Knowledgebase ID 127: http://www.minitab.com/support/answers/answer.aspx?ID=127

Page 1

Technical Support Document

Taking the Perplexity out of P-Values


The p-value is also the proportion of the test statistics from samples of the same size (taken from the same population) that would yield a more extreme value of the test statistic under the assumption that H0 is true. In our t-test example, the test statistic is a funcition of the mean, and the p-value is .026. This means that 2.6% of the means from samples of size 35, drawn from a population where = 25, will produce a mean which indicates that is not equal to 25. Ask yourself which is most likely: that you were lucky enough to draw one of the few samples that indicate that is not equal to 25; or that really is not equal to 25? The p-value is traditionally compared to values of less than .05 or .01, depending on the field of study (your mileage may vary!). Check journal articles in your field for acceptable values. In our example, lets assume a .05 value of . The p-value of .026 indicates that the mean m.p.g. of all cars of this type (and not just the mean of the 35 cars in the study) is probably not equal to 25. A more statistically correct way to state this is at a significance level of .05, the mean m.p.g. appears to be significantly different from 25. Using P-Values in Other Tests The null hypothesis varies from test to test. Interpreting the p-value depends on knowing what the null and alternative hypotheses are. The Minitab output for regression, for example, includes a hypothesis test for the estimates of0, 1, In the example below, the variable DIAMETER is very statistically significant, because the p-value is essentially 0. Output from Regression The regression equation is VOLUME = - 36.9 + 5.07 DIAMETER Predictor Constant DIAMETER s = 4.252 Coef -36.943 5.0659 Stdev 3.365 0.2474 t-ratio -10.98 20.48 p 0.000 0.000

R-sq = 93.5%

R-sq(adj) = 93.3%

In a test for normality, such as the Anderson-Darling test or the Kolmogorov-Smirnov test, the null hypothesis is that the data have a normal distribution. In a Chi-squared test of independence, the null hypothesis is that the two factors in the studyfor example, age and voting preferenceare independent of one another. If the test results in a small enough p-value, then there is enough sample evidence to reject the null hypothesis in favor of the alternative hypothesis (the data have a non-normal distribution or the factors are not independent). Using p-values, then, is simple if you know two key facts: the values of that are acceptable in your field, and the null and alternative hypotheses for the tests you are using.

Knowledgebase ID 127: http://www.minitab.com/support/answers/answer.aspx?ID=127

Page 2

You might also like