Chi-Square Goodness of Fit (GOF) tests if the data follows a multi-nominal distribution with certain proportions. If the proportions are not specified, the categories are assumed to have equal proportions. You can also specify assumed proportions to test. ProcessMA GOF can analyze raw or summarized data. In the case of raw data, the frequency count can be left empty (assumed = 1) and specified proportions for each category must be the same. Note that the chi-square may not be reliable if the expected frequency of any category is less than 5.
| Null Hypothesis, H0 | Alternate Hypothesis, H1 |
| Data follow a multi-nomial distribution with certain proportions | Data do not follow a multi-nomial distribution with certain proportions |
At the Excel Menu (For Excel 2007, go to Add-ins first)
Choose ProcessMA > Statistics > Basic Statistics > Chi-Square Goodness-of-Fit Test
In Category, select the column containing the categorical data (>= 2 distinct values)
In Frequency, select the column containing frequency counts (Positive integer)
In Proportions, select the column containing block levels (Optional, Numeric, >0 & <1)
Click OK
You want to assess if the new student’s interests in the different faculties are similar to the previous year. You have collected data on 300 new student’s interest.
Open data worksheet by choosing ProcessMA > Tools > Data
In Category, select BV - Interest
In Frequency, select BW - Count
In Proportions, select BX - Probability
Click OK
For a desired a = 0.05, since p = 0.003 < a, we can reject H0 and conclude that the interest in different faculties for the new students is different from the previous year.
Copyright © ProcessMA