In: Statistics and Probability
Use the distribution below to test that Plain M&Ms follow the stated distribution. Discuss your choice of ?. Would a different ? have changed your conclusion?
M&M states the following distribution for Plain
M&Ms:
Red = 13%, Orange = 20%, Yellow = 13%, Green = 20%, Blue = 20%,
Brown = 14%
Total number of plain M&M's: 665 Total number of red plain M&M's: 76 Total number of plain brown M&M's: 66 Total number of blue plain M&M's: 179 Total number of orange plain M&M's: 140 Total number of green plain M&M's: 119 Total number of yellow plain M&M's: 85
Total number of peanut M&M's: 356 Total number of red peanut M&M's: 22 Total number of peanut brown M&M's: 34 Total number of blue peanut M&M's: 55 Total number of orange peanut M&M's: 62 Total number of green peanut M&M's: 72 Total number of yellow peanut M&M's: 111
Chi-square goodness of fit test:
Number of categories, n =6
Null Hypothesis(H0):
Plain M&Ms follow the stated distribution.
Alternative Hypothesis(H1):
Plain M&Ms do not follow the stated distribution.
Calculation of the test statistic:
Color of plain M&M | Observed frequencies: O | Expected frequencies: E | (O - E)2/E |
Red | 76 | 665*13% =86.45 | 1.2632 |
Orange | 140 | 665*20% =133 | 0.3684 |
Yellow | 85 | 665*13% =86.45 | 0.0243 |
Green | 119 | 665*20% =133 | 1.4737 |
Blue | 179 | 665*20% =133 | 15.9098 |
Brown | 66 | 665*14% =93.1 | 7.8884 |
Total | 665 | 665 | 26.9278 |
Test statistic, 26.9278
Comparing with the critical value:
Let the significance level, =0.05
It means, the probability of rejecting the true null hypothesis =P(type 1 error) =0.05
Degrees of freedom, df =n - 1 =6 - 1 =5
The critical value is: =11.0705
Decision criteria:
Reject H0 if . Otherwise, do not reject it.
Conclusion:
26.9278 > 11.0705
Since , reject the null hypothesis(H0) at 5% significance level.
Thus, there is a sufficient statistical evidence at 5% significance level to claim that the Plain M&Ms do not follow the stated distribution.
Let us take different significance level of 0.01.
At =0.01 and df =5, =15.0863
The conclusion remains the same at 0.01 significance level because 26.9278 > 15.0863