NA1C0012 Jenny 11.5.2 ScentenceAccent with sex ScentenceAccent with status 5 6 6 38 5 1 1 2 2 3 4 sentenceaccent 4 3 sentenceaccent 36 Female Male Non sex Late Early status The boxplot of scentenceaccent and sex shows outliers both in female and male. That seems variances are very different for males and females. The boxplot of sentenceaccent and status shows that the data is normally distributed. AnovaModel.1 <- (lm(sentenceaccent ~ sex*status, data=forget)) > Anova(AnovaModel.1) Anova Table (Type II tests) Response: sentenceaccent Sum Sq sex 6.071 status 20.270 sex:status 1.464 Residuals 35.186 38 Df 1 2 2 F value 6.5563 10.9452 0.7908 Pr(>F) 0.0145480 * 0.0001762 *** 0.4608118 --Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 According to the result of the factorial ANOVA testing for the intersation of two factors( sex and status) within the response variable (sentenceaccent), the factors of “sex” (F value =6.55; p-value=0.01)and “status” (F value =10.945; p-value=0.0001) are statistical, however, the interaction between sex and status(F value =0.79; p-value=0.46) is not statistical. > numSummary(forget[,"sentenceaccent"], groups=forget$sex, statistics=c("mean", "sd", "IQR", "quantiles"), quantiles=c(0,.25,.5,.75,1)) mean sd IQR 0% 25% 50% 75% 100% data:n Female 3.750000 1.207884 1.5000 Male 2.238636 1.011581 0.9375 1.375 2.875 3.625 1.000 1.500 2.125 4.3750 2.4375 6.75 4.75 33 11 > numSummary(forget[,"sentenceaccent"], groups=forget$status, statistics=c("mean", "sd", "IQR", "quantiles"), quantiles=c(0,.25,.5,.75,1)) mean sd IQR 0% 25% 50% 75% 100% data:n Non 2.508333 1.0116141 1.7500 Late 3.100000 0.9344594 Early 4.589286 1.1152662 1.000 1.50000 2.625 3.25000 4.375 1.625 2.31250 2.875 3.75000 4.875 15 .2500 2.875 3.90625 4.750 5.15625 6.750 14 1.4375 15 From the mean scores of gender, we can see the females performed statistically better than males. But for the result of status, we only can see the Early group is the best of three groups , the Late group is in the middle and the Non group is the worst. ※posthoc summary(.Pairs) # pairwise tests Simultaneous Tests for General Linear Hypotheses Multiple Comparisons of Means: Tukey Contrasts Fit: aov(formula = sentenceaccent ~ status, data = forget) Linear Hypotheses: Estimate Std. Error t value Late - Non == 0 0.5917 0.3727 1.587 Early - Non == 0 2.0810 0.3793 5.486 Early - Late == 0 1.4893 0.3793 3.926 --Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 Pr(>|t|) 0.263 <0.001 *** <0.001 *** (Adjusted p values reported -- single-step method) The posthoc test shows that the group of “Late-Non” is not statistic, but the other two groups”Early-Non” and “Early-Late” are statistic. Hence, we can say that Early group did the best of all. ※ggplot attach(forget) forget ggplot(forget,aes(x=status,y=sentenceaccent,fill=sex))+ geom_bar(stat="identity",width=0.5,color="black",position="dodge") From the ggplot of the data”forget”, we can see no matter in “Non”,”Late”or “Early”, females have the better performance than males. Furthermore, foucsed on the result of status visually, here the overall performance of “Early Group” is better than the orther two groups. 11.5.3 Anova Table (Type II tests) Response: score Sum Sq 272.25 1599.74 236.30 806.30 90 Df 1 4 4 F value 30.389 44.641 6.594 Pr(>F) 3.342e-07 *** < 2.2e-16 *** 0.0001062 *** agegroup task agegroup:task Residuals --Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 Pairwise comparisons using t tests with pooled SD data: score and task Rhyming Adjective Imagery Intention Counting 0.6718 2.0e-06 1.5e-10 3.2e-11 Counting vs. Intention most significant Rhyming Adjective 9.7e-06 8.5e-10 0.0369 1.2e-10 0.0098 Imagery 0.6466 Counting vs. Rhymingnot significant Imagery vs. Intentionnot significant Old:Rhyming Old:Adjective Old:Imagery Old:Counting Old:Rhyming Old:Adjective Old:Imagery Old:Intention 0.94061 0.00542 0.00449 1.6e-05 1.3e-05 0.09542 - Old:Intention Young:Counting Young:Rhyming Young:Adjective Young:Imagery Young:Intention 0.00059 0.74264 0.70185 3.0e-07 3.9e-11 5.8e-15 Old:Rhyming Old:Adjective Young:Counting Young:Rhyming Young:Adjective Young:Imagery Adjective vs. Intention or Imagery is originally significant - Old:Imagery Old:Intention Young:Counting Young:Rhyming Young:Adjective Young:Imagery Young:Intention 0.47698 6.8e-08 9.1e-12 2.9e-15 0.00048 0.78317 0.66106 2.3e-07 3.2e-11 5.8e-15 0.51409 0.00197 0.01800 0.00812 9.9e-06 2.8e-09 Applying post hoc 0.35338 4.2e-06 8.6e-05 0.35338 0.00370 4.2e-06 the intention they will have) Adjective vs. Imagerynot significant 1.9e-06 2.8e-10 0.05198 3.7e-14 0.00017 0.00017 0.00241 0.05198 0.00014 6.8e-08 - Adjective vs. Old:Intentionnot significant (the more the age, the less 0.06000