Mass Communication Research
Steps in Statistical Hypothesis Testing
INDEX SYLLABUS SCHEDULE e-MEDIA COMM-STOP

The outline below provides a list of all the things you have to do to use inferential statistics to test hypothesis. It also gives some quick definitions, explanations, and things to look out for.
  1. State the hypothesis.
    • A hypothesis is a statement of a relationship between at least two variables.
    • Decide whether the hypothesis is one- or two-tailed. (One-tailed hypothesis indicates the direction of a difference or relationship; two-tailed hypothesis simply asserts that there is a difference or relationship.)

  2. Determine the independent and dependent variables.
    • The independent variables is the causal or predictor variable.
    • The dependent variable is the effect or variable the value of which depends on the independent variable.

  3. Determine the level of measurement for each variable.
    • Nominal: a variable for which numbers serve as the name of categories
    • Ordinal: a variable for which numbers indicate a rank, but the differences between adjacent numbers is not equal.
    • Interval: a variable for which numbers indicate rank and the differences between adjacent numbers is equal, but the there is no natural zero point at which none of the property being measures exists.
    • Ratio: a variable for which numbers indicate rank, differences between adjacent numbers are equal and there is a natural zero point.

  4. Choose the inferential statistics (four are considered in Media Research)
    • t test: Compares two means.
      Appropriate when the independent variable is nominal -- with two categories -- and the dependent variable is interval or ratio.
    • Chi-square test: Compares proportions or percentages across two or more groups.
      Appropriate when both the independent and dependent variables are nominal.
      There are two kinds of chi-square tests
      1. Goodness of Fit
      2. Test for Independence
    • F-test ( Analysis of Variance , ANOVA): Compares three or more means
      Appropriate when the independent variable is nominal and the dependent variable is interval or ratio.
      There are many kinds of ANOVA, but we will consider only the simple one-way case in Media Research
    • Pearson's r: tests the relationship between to interval or ratio variables.
    • Spearman's Rho: test the relationship between two ordinal variables.

    Dependent Variable
    NominalInterval or Ratio
    StandardChi-Square "Goodness of Fit"One-sample t-Test
    Difference between groupsChi-Square Test for IndependenceIndependent Samples t-Test
    or  ANOVA 1
    Interdependent Variables
    Comparing relationships between ordinal ratingsSpearman Rho Correlation
    Comparing relationships between interval or ratio ratingsPearson Correlation
    Comparing differences between interval or ratio ratingsPaired Sample t-Test
    1. The independent variable has more than two groups.
  5. Calculate the inferential statistic. Use a Computer to do this.

  6. Interpret the significance level.
    • Generally in the social sciences including communication we use the .05 significance level.
    • Remember to keep track of one- and two-tailed hypotheses. If the computer reports a two-tailed significance level and the hypothesis is one tailed, divide the significance level by two.

  7. Interpret the results.
    • A statement about differences between means for t-test and ANOVA
    • A statement about differences in percentages for chi-square
    • A statement about relationships or slopes for Correlation's/Regression.
If you don't understand something in this Web note, please e-mail Dr. Sitton.
INDEX SYLLABUS SCHEDULE e-MEDIA COMM-STOP
©M. Mark Miller & Ronald W. Sitton 2009
Revised 092811 — http://www.uamont.edu/FacultyWeb/sitton/crz/mrea/hypo.html