17.01.2015 Views

LibraryPirate

LibraryPirate

LibraryPirate

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

13.9 THE FRIEDMAN TWO-WAY ANALYSIS OF VARIANCE BY RANKS 727<br />

4. Test statistic. By means of the Friedman test we will be able to determine<br />

if it is reasonable to assume that the columns of ranks have been<br />

drawn from the same population. If the null hypothesis is true we would<br />

expect the observed distribution of ranks within any column to be the<br />

result of chance factors and, hence, we would expect the numbers 1, 2,<br />

and 3 to occur with approximately the same frequency in each column.<br />

If, on the other hand, the null hypothesis is false (that is, the models are<br />

not equally preferred), we would expect a preponderance of relatively<br />

high (or low) ranks in at least one column. This condition would be<br />

reflected in the sums of the ranks. The Friedman test will tell us whether<br />

or not the observed sums of ranks are so discrepant that it is not likely<br />

they are a result of chance when H 0 is true.<br />

Since the data already consist of rankings within blocks (rows), our<br />

first step is to sum the ranks within each column (treatment). These sums<br />

are the R j shown in Table 13.9.1. A test statistic, denoted by Friedman<br />

as x 2 r, is computed as follows:<br />

x 2 r =<br />

k<br />

12<br />

nk1k + 12 a<br />

j =1<br />

1R j 2 2 - 3n1k + 12<br />

(13.9.1)<br />

where n = the number of rows (blocks) and k = the number of columns<br />

(treatments).<br />

5. Distribution of test statistic. Critical values for various values of n and<br />

k are given in Appendix Table O.<br />

6. Decision rule. Reject H 0 if the probability of obtaining (when H 0 is<br />

true) a value of x 2 r as large as or larger than actually computed is less<br />

than or equal to a.<br />

7. Calculation of test statistic. Using the data in Table 13.9.1 and Equations<br />

13.9.1, we compute<br />

x 2 r =<br />

12<br />

913213 + 12 311522 + 1252 2 + 1142 2 4 - 319213 + 12 = 8.222<br />

8. Statistical decision. When we consult Appendix Table Oa, we find that<br />

the probability of obtaining a value of x 2 r as large as 8.222 due to chance<br />

alone, when the null hypothesis is true, is .016. We are able, therefore,<br />

to reject the null hypothesis.<br />

9. Conclusion. We conclude that the three models of low-volt electrical<br />

stimulator are not equally preferred.<br />

10. p value. For this test, p = .016.<br />

■<br />

Ties When the original data consist of measurements on an interval or a ratio scale<br />

instead of ranks, the measurements are assigned ranks based on their relative magnitudes<br />

within blocks. If ties occur, each value is assigned the mean of the ranks for which<br />

it is tied.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!