An environmental scientist wants to determine whether temperature changes in the ocean near a nuclear power plant affect the growth of fish. The scientist randomly divides 25 newly hatched fish into four groups and places each group into a separate, simulated ocean environment. The simulated environments are identical except for temperature. Six months later, the scientist measures the weights of the fish. To determine whether the median weight of the fish differs among the four groups, the scientist uses Mood's median test.
 Open the sample data, FishWeights.MTW.
 Open the Mood’s Median Test dialog box.
 Select Responses are in one column for all factor levels.
 In Response, enter Weight.
 In Factor, enter Temp.
 Click OK.
Interpret the results
For each factor level, Minitab displays the median, the number of observations less than or equal to the median, the number of observations greater than the median, and the interquartile range.
Because the pvalue of 0.6965 is greater than the commonly used significance level of 0.05, there is insufficient evidence for the scientist to conclude that the medians are different. The differences between the median weights are not statistically significant.
The interval plot displays the 95% confidence interval for the population medians. You can be 95% confident that the population median for each group is within the corresponding interval.

38  19  4  3  4.00  42  19  3  3  9.50  46  22  2  4  7.25  50  18  4  2  4.25  Overall  19    



Null hypothesis  H₀: All medians are equal  Alternative hypothesis  H₁: At least one median is different 



