In: Statistics and Probability
A teacher is trying to assign grades so that she has the following distribution:
Grade | % of Class |
A | 15 |
B | 30 |
C | 40 |
D | 10 |
F | 5 |
This is what she actually has in her grade book:
Grade | # Students |
A | 20 |
B | 40 |
C | 35 |
D | 8 |
F | 8 |
At a significance level of 0.05, are these two distributions different? Include either χ2 or the p-value to justify your answer.
Total Frequency = 20 + 40 + 35 + 8 + 8 = 111
Therefore the expected frequencies for each grade here is
computed as:
E(A) = 0.15*111 = 16.65
E(B) = 0.3*111 = 33.3
E(C) = 0.4*111 = 44.4
E(D) = 0.1*111 = 11.1
E(F) = 0.05*111 = 5.55
Therefore now the chi square test statistic here is computed as:
Grade | O_i | E_i | Chi-Sq |
A | 20 | 16.65 | 0.67402402 |
B | 40 | 33.3 | 1.34804805 |
C | 35 | 44.4 | 1.99009009 |
D | 8 | 11.1 | 0.86576577 |
F | 8 | 5.55 | 1.08153153 |
111 | 111 | 5.95945946 |
From the above bold value, we get here:
For k - 1 = 4 degrees of freedom, the p-value here is obtained from the chi square distribution tables as:
As the p-value here is 0.2022 > 0.05 which is the level of significance, therefore the test is not significant and we cannot reject the null hypothesis here. Therefore we have insufficent evidence here that the distribution is different here.