D10: Relationships – Chi-Square and related stats

22 unread replies.22 replies.

Start this discussion as soon as you’ve read the text/module info on the topic and wrapped up the last discussion. Try not to be working in two discussions at the same time – that’s a recipe for confusion!

Three posts are the minimum – one in response to my prompt below and at least two responses to the posts of other students. Use the discussion like a study group: Use your knowledge, practice, ask questions… Wrap up your discussion after participating for a few days -while you’re working on the worksheet and homework related to the topic – don’t wait too long since the quiz is coming. All discussion for this section close with the quiz.

I have completed a number of Chi-Square tests using the following variables from the GSS – a national random sample done by a reputable research entity. (For more info on the GSS, visit their website: gss.norc.org (Links to an external site.))

Each of the discussion topics has one main nominal (or ordinal) variable – and the output from a number of tests with different nominal (or ordinal) variables.

• Choose ONE of these sets of tests to discuss. You will choose a Chi-Square test from this set to interpret (your first post) and then discuss (with at least two other posts) those and other findings.
  • Select a nominal/ordinal variable (HealthActions, HappinessActions, LifeActions, Same-Sex MarriageActions) and open the file with that name (click the name). You will see inside that file two sets of output. Your variable will be compared with one or more of the following variables (Born in the U.S. – Social Class – Degree – Race – Marital Status – Sex – Spiritual Person).
    • The first box of output (first page) has the general statistics for each set of variables – the mean, N, and standard deviation.
    • The second box (second page) has the Chi-Square test results…
    • Find the results for your main variable and its comparison to the other variable you chose.
    • Do only one test!
• Figure the df [(#rows-1)*(#columns-1)] and, using an alpha of 0.05, find the critical value of the test statistic (Use the correct table in the book – Critical Value of Chi-Square).
• Make a decision about the null hypothesis – accept or reject. (Identify the Chi-Square for your variable’s test on the printout.)
• Interpret your findings – is there a relationship between these two variables? If so, what is the nature of that relationship? (Refer to the means or percents and the relevant correlation coefficient).
• Speculate on why these findings might be as they are. Is it logical? Might there be an error in your result? If so, why? (sample size? bias? Type I or II error?)

---

Nominal/Ordinal (dependent) variables: Health, Happiness, Life, Same-Sex Marriage

Nominal/Ordinal (independent) variables: Born in the U.S. – Social Class – Degree – Race – Marital Status – Sex – Spiritual Person

---

Note: R = Respondent (the person who answered the survey)