- 1 decade agoFavorite Answer
3.1 Chi-square test
3.2 No. it should be H0:age⊥gender vs H1:age (not)⊥ gender
3.3 The expectation of cell number. It is calculated from the corresponding row number multiplies the corresponding column number, then is devided by total number. For example. For cell(1,1), 4.3=102*9/214
3.4 Check the Chi-square table. The p-value of Pearson Chi-square is 0.726 which is larger than significance level 0.05, so you can conclude that it is not statistically significant. Meanwhile, we do not reject null hypothesis.
4.1 Pooled-t test
4.2 (1)H0:V(male)=V(female) vs H1: V(male) ne V(female)
(2)H0:mu_male=mu_female vs H1:mu_male ne mu_female
4.3 Check the t test table. First of all, you need to test the homogeneity of variance. The p-value of HOV is 0.940 which is larger than 0.05, so you can conclude both variances of two genders are equal. Then, go to the following columns and check the upper line. The p-value is 0.174 which is also larger than 0.05, so you can conclude two means are equal.
5.2 Based on cell mean coding:
H1:at least one beta_i is not equal to beta_j, where i ne j
5.3 First, check ANOVA table, the p-value is 0.000 which is smaller than 0.05. It means the ANOVA model is significant; meanwhile, beta is significant as well. Second, check the post hoc test, go to the column of p-value, and you can conclude them by the same principle above. It's too long, so I don't list details here.
6.1 Linear regression model
6.3 H0:beta_i=0 vs H1:beta_i ne 0, where i=1,2,3,4
6.4 The whole model is significant because of small p-value 0.000. All coefficients are also significant with small p-value besides EU because its p-value is larger than 0.05. The R^2 is 0.693, and the adjusted R^2 is 0.686, which means those factors can interpret Y around 70%.Source(s): me
- 阿偉Lv 71 decade ago