Chi Square
Content
Chi-Square Test in SPSS (PASW)
Reference Department Albert S. Cook Library
Joyce Garczynski – [email protected] – 410-704-5168
Background: Chi-square is a statistical test that tests for the existence of a relationship between two variables. This test can be used with nominal, ordinal, or scale variables, so it is a very versatile test, but it is sensitive to sample sizes too. It is important to have at least a few cases in each of the values of both of the variables involved in this test or the results will be skewed.
1) Formulate a hypothesis about your variables What do you think is the relationship between the two variables? In this example, we want to test if men and women are significantly different in how they talk about politics. So the independent variable is gender (represented by gender in the dataset) and the dependent variable is frequency of political talk (represented by rsclaetalk in the dataset).
2) Select crosstabs Click on the “Analyze” tab at the top of the page Select “Descriptive Statistics” from the list Select “Crosstabs”
Page 1 of 4
3) Select your variables for the test
1) First select the dependent variable. Click on the variable name for the dependent variable in the left hand column so it is highlighted and then click the arrow next to “Row(s)” in between the two sections to move the variable to the rows box. Next select the independent variable. Click on the variable name for the independent variable in the left hand column so it is highlighted and then click the arrow next to “Column(s)” in between the two sections to move the variable to the columns box. 2) Click on the “Statistics” button. 3) Click on the “Cells” button.
2 3 1
4
4) Click on the “OK” button.
4) Select the chi-square statistic from the “Statistics” button
1
1) Select the “Chi-square” check box from the menu. 2) Click on the “Continue” button.
2
Page 2 of 4
5) Select column percentages from the “Cells” button
1) Select the “Column” check box from the menu.
1
2) Click on the “Continue” button.
2
6) Examine the results Look at the “Chi-Square Tests” table to see if there is a significant relationship
1) Look at the number in the “Asyp. Sig.” column for “Pearson Chi-Square row”. In communication, it is convention that if this value is less than .05, then the statistic is considered to be significant (meaning that the researcher can be 95% confident that the relationship between the two variables is not due to chance). In this example, since the Sig. value is .576 (which is greater than .05), we can say that there is not a significant relationship between gender and the frequency of political talk.
1
Look at the “Crosstabulation” table to see where a significant difference is
1) If your chi-square is significant, look across the column percentages for the values of the independent variable. You should see a definite trend in the percentages with the percentages for one value of the dependent variable trending one way, while the percentages for another value trend in the opposite direction. In this example, we do not see a trend because we do not have a significant chisquare.
1
Page 3 of 4
7) Write up the results in APA style When reporting the results of a chi-square test within the text of a paper, first write the Χ2 (chi-square) value with the degrees of freedom (located under the “df” column for “Pearson Chi-Square” row of the “Chi-Square Tests” table) and the sample size in parentheses. Then write the significance level. Note that the exact significance level should be reported unless it is less than .001 (that would be written p < .001). Also note that otherwise most statistics should be rounded to two decimal places. For example: A chi-square test was performed and no relationship was found between gender and the frequency of political talk, X2 (2, N = 170) = 1.10, p =.58.
Page 4 of 4
Reference Department Albert S. Cook Library
Joyce Garczynski – [email protected] – 410-704-5168
Background: Chi-square is a statistical test that tests for the existence of a relationship between two variables. This test can be used with nominal, ordinal, or scale variables, so it is a very versatile test, but it is sensitive to sample sizes too. It is important to have at least a few cases in each of the values of both of the variables involved in this test or the results will be skewed.
1) Formulate a hypothesis about your variables What do you think is the relationship between the two variables? In this example, we want to test if men and women are significantly different in how they talk about politics. So the independent variable is gender (represented by gender in the dataset) and the dependent variable is frequency of political talk (represented by rsclaetalk in the dataset).
2) Select crosstabs Click on the “Analyze” tab at the top of the page Select “Descriptive Statistics” from the list Select “Crosstabs”
Page 1 of 4
3) Select your variables for the test
1) First select the dependent variable. Click on the variable name for the dependent variable in the left hand column so it is highlighted and then click the arrow next to “Row(s)” in between the two sections to move the variable to the rows box. Next select the independent variable. Click on the variable name for the independent variable in the left hand column so it is highlighted and then click the arrow next to “Column(s)” in between the two sections to move the variable to the columns box. 2) Click on the “Statistics” button. 3) Click on the “Cells” button.
2 3 1
4
4) Click on the “OK” button.
4) Select the chi-square statistic from the “Statistics” button
1
1) Select the “Chi-square” check box from the menu. 2) Click on the “Continue” button.
2
Page 2 of 4
5) Select column percentages from the “Cells” button
1) Select the “Column” check box from the menu.
1
2) Click on the “Continue” button.
2
6) Examine the results Look at the “Chi-Square Tests” table to see if there is a significant relationship
1) Look at the number in the “Asyp. Sig.” column for “Pearson Chi-Square row”. In communication, it is convention that if this value is less than .05, then the statistic is considered to be significant (meaning that the researcher can be 95% confident that the relationship between the two variables is not due to chance). In this example, since the Sig. value is .576 (which is greater than .05), we can say that there is not a significant relationship between gender and the frequency of political talk.
1
Look at the “Crosstabulation” table to see where a significant difference is
1) If your chi-square is significant, look across the column percentages for the values of the independent variable. You should see a definite trend in the percentages with the percentages for one value of the dependent variable trending one way, while the percentages for another value trend in the opposite direction. In this example, we do not see a trend because we do not have a significant chisquare.
1
Page 3 of 4
7) Write up the results in APA style When reporting the results of a chi-square test within the text of a paper, first write the Χ2 (chi-square) value with the degrees of freedom (located under the “df” column for “Pearson Chi-Square” row of the “Chi-Square Tests” table) and the sample size in parentheses. Then write the significance level. Note that the exact significance level should be reported unless it is less than .001 (that would be written p < .001). Also note that otherwise most statistics should be rounded to two decimal places. For example: A chi-square test was performed and no relationship was found between gender and the frequency of political talk, X2 (2, N = 170) = 1.10, p =.58.
Page 4 of 4
