Journalism 614 Relationship testing Reading data files Survey

Journalism 614: Relationship testing

Reading data files ¨ Survey questionnaire (word): ¨ Survey data (excel): – Each row: a respondent – Each column: a variable (question/condition) ¨ For easy data management, copy your columns and the demographic variable columns and paste a clean excel file ¨ Keep an eye out on the correspondence between response categories and coded values

Experimental design Condition 1 Condition 2 Response

How to test the effects? ¨ Condition 1: The majority of public university funding is tax revenue from U. S. citizens. The U. S. government shouldn’t allow undocumented immigrants to access public education. ¨ Condition 2: The majority of public university funding is tax revenue from U. S. citizens. The U. S. government shouldn’t allow illegal immigrants to access public education. ¨ Response categories: – – – Strongly disagree (1) Somewhat disagree (2) Neither disagree nor agree (3) Somewhat agree (4) Strongly agree (5)

Step 1: Compare means ¨ Mean(undocumented immigrants): 3. 23 ¨ Mean(illegal immigrants): 3. 54 ¨ Statistical test: ONE-WAY ANOVA

Step 2: Compare distribution of responses strongly disagree undocumented immigrants illegal immigrants somewhat strongly disagree neither agree total 20 6 18 11 27 82 7 10 19 18 24 78 Figure 1:

Step 2: Compare distribution of responses strongly somewhat strongly disagree neither agree undocumented immigrants illegal immigrants total 24% 7% 22% 13% 33% 100% 9% 13% 24% 23% 31% 100% Figure 2:

Step 3: Compare responses within a question strongly somewhat strongly disagree neither agree undocumented immigrants illegal immigrants total 24% 7% 22% 13% 33% 100% 9% 13% 24% 23% 31% 100%

Step 3: Compare responses within a question strongly somewhat strongly disagree neither agree undocumented immigrants illegal immigrants total 24% 7% 22% 13% 33% 100% 9% 13% 24% 23% 31% 100% ¨ Undocumented immigrants condition: ¨ Strongly disagree (24%) vs Strongly agree (33%) ¨ Disagree (31%) vs Agree (46%)

Step 3: Compare responses within a question strongly somewhat strongly disagree neither agree undocumented immigrants illegal immigrants total 24% 7% 22% 13% 33% 100% 9% 13% 24% 23% 31% 100% ¨ Illegal immigrants condition: ¨ Strongly disagree (9%) vs Strongly agree (31%) ¨ Disagree (22%) vs Agree (54%)

Step 4: Compare the same response category of two conditions strongly somewhat strongly disagree neither agree undocumented immigrants illegal immigrants total 24% 7% 22% 13% 33% 100% 9% 13% 24% 23% 31% 100%

Step 4: Compare the same response category oftwo conditions strongly somewhat strongly disagree neither agree undocumented immigrants illegal immigrants total 24% 7% 22% 13% 33% 100% 9% 13% 24% 23% 31% 100% ¨ Do we have statistical significance in the difference of the percentages of respondents who strongly disagreed?

Step 4: Compare the same response category of two conditions strongly somewhat strongly disagree neither agree undocumented immigrants illegal immigrants total 24% 7% 22% 13% 33% 100% 9% 13% 24% 23% 31% 100% ¨ Do we have statistical significance in the difference of the percentages of respondents who strongly agreed?

Step 5: Demographics ¨ How do demographic characteristics affect responses? ¨ Party ID, news consumption, education, gender, age, race, etc. ¨ Statistical test: Multivariate Regression

Step 5: Demographics ¨ Party ID Figure 3:

Step 5: Demographics ¨ New consumption Figure 4:

Step 6: Interpret and draw conclusions ¨ Summarize the responses in two conditions (mean, within, between, relevant demographic variables) ¨ Highlight the differences in response that are statistically significant ¨ Draw conclusion: is your media effect hypothesis confirmed or disconfirmed? ¨ Any methodological notes about your data?