Calculate the total for each column and each row of the table.
Calculate the grand total in the bottom right-hand corner of the table.
Calculate the expected frequency using the following equation: (column total/grand total)*row total.
Determine if the observed values deviate from the expected values using the following equation: (expected value - observed value)^2) / expected value. In other words, "the chi square is the sum of the squared difference between observed and expected data, divided by the expected data in all possible categories," according to information from Penn State University.
Evaluate your results to determine if a significant difference between the expected and observed results exists. If no significant difference exists, it is called a null hypothesis.
Calculate the total for each column and each row of the table.
Calculate the grand total in the bottom right-hand corner of the table.
Calculate the expected frequency using the following equation: (column total/grand total)*row total.
Determine the G-value by multiplying the observed value by the natural log of the value determined from the observed value ("O") divided by the expected value ("E") and then multiplying the entire sum by 2. In other words, the G-value ("G") can be calculated as 2* [O * ln (O/E)].
Evaluate the value to understand its correlation to the hypothesis or expected value. A G-value of zero means that the observed numbers are as expected. A larger G-value means there is a larger difference between the observed and the expected values.
Calculate the degrees of freedom by determining the minimum number of categories whose value need to be known before the remaining values can be calculated. For example, if you surveyed 100 people and you know from the first two values that 79 people have been surveyed, then you can infer the results of the third. Therefore, the degrees of freedom is two.
Determine the critical G-value based upon the number of degrees of freedom. The critical G-value should be based upon a confidence interval, or P-value, which is typically 0.05. For two degrees of freedom and a P-value of 0.05, the critical G-value is 6.0.
Compare the critical G-value to the calculated G-value. If the calculated G-value is not greater than the G-value, then the data does not significantly differ from the expected value.