Nonparametric Tests: Chi Square Lesson 16 Parametric vs. Nonparametric Tests Parametric hypothesis test about population parameter (m or s ) z, t, F.
Download ReportTranscript Nonparametric Tests: Chi Square Lesson 16 Parametric vs. Nonparametric Tests Parametric hypothesis test about population parameter (m or s ) z, t, F.
Nonparametric Tests: Chi Square 2 Lesson 16 Parametric vs. Nonparametric Tests Parametric hypothesis test 2 about population parameter (m or s ) z, t, F tests interval/ratio data Nonparametric tests do not test a specific parameter nominal & ordinal data frequency data ~ Chi-square ( 2) Nonparametric tests same 4 steps as parametric tests Chi-square test for goodness of fit single variable Chi-square test for independence two variables Same formula for both degrees of freedom different fe calculated differently ~ Sample Data: 2 Frequency Expected frequency (fe) fe = pn Observed frequency (fo) S fo = n Degrees of freedom:Goodness of fit C-1 C = number of cells (categories) 2 cv from table B.5, page 364 ~ Chi-square (2) 2 fo fe 2 fe Assumptions & Restrictions Independence of observations any score may be counted in only 1 category Size of expected frequencies 2 If fe < 5 for any cell cannot use More likely to make Type I error Solution: use larger sample ~ 2 Test for Goodness of Fit Test about proportions (p) in distribution 2 different forms of H0 No preference category proportions are equal No difference from comparison population e.g., student population 55% female and 45% male? H1: the proportions are different ~ Null Hypotheses: 2 No preference: H0 No difference: H0 Coke Pepsi ½ ½ Female Male 55% 45% SPSS: No Preference Data in 1 column Analyze Nonparametric Legacy Dialogs Chi square Dialogue box Test Variable List Expected Values All categories Equal Options Descriptives (frequencies) ~ SPSS: No Difference Same menus as No Preference But must specify proportions or frequencies Dialogue box Expected Values Values Specify & Add vales one at time In same order as defined values for variable in variable view ~ *Effect Size: 1 Variable EffectSize 2 N (df ) N = total sample size across all categories df = #categories – 1 zero = no difference 1 = large difference ~ 2 Test for Independence 2 variables are they related or independent H0: distribution of 1 variable is the same for the categories of other no difference Same formula as Goodness of Fit different df ~ 2 Test for Independence Differences from Goodness of Fit df = (R-1)(C-1) R = rows C = columns Expected frequency for each cell fC f R fe N Example Does watching violent TV programs cause children to be more aggressive on the playground? Data: frequency data Violent program: yes or no Aggressive: yes or no ~ 2 Test for Independence Aggressive Yes No Violent TV Yes No 41 9 17 33 SPSS: Test for Independence Two variables Two-Way Contingency Table Analysis Data: 1 column for each variable Analyze Descriptives Crosstabs Dialogue Box Variables Rows or Columns Statistics Chi Square, *Phi ~ Effect Size (): 2 Variables 2 N N = total sample size across all categories Phi values: 0-1 Interpret similar to Pearson’s r Small = .1; medium = .3, large = .5 ~