aliens <- read.csv ("aliens.csv", header = TRUE, stringsAsFactors = TRUE)
source('special_functions.R')
my_sample <- make.my.sample(33002176, 100, aliens)
## Warning in RNGkind("Mersenne-Twister", "Inversion", "Rounding"): non-uniform
## 'Rounding' sampler used
library(lsr)
library(sciplot)
bargraph.CI(x.factor = my_sample$island,
response = my_sample$control,
ci.fun = ciMean, ylim = c(0, 100),
legend = T)
bargraph.CI(x.factor = my_sample$college,
response = my_sample$control,
ci.fun = ciMean, ylim = c(0, 100),
legend = T)
Based on the graphs, I do not think intelligence differs between islands
and I do not think intelligence differs between colleges? I really do
not see any differences among the islands or colleges in the topic of
intelligence.
cp.anova <- aov(my_sample$control ~ my_sample$island)
summary(cp.anova)
## Df Sum Sq Mean Sq F value Pr(>F)
## my_sample$island 2 353 176.49 1.827 0.166
## Residuals 97 9371 96.61
cp.anova <- aov(my_sample$control ~ my_sample$college)
summary(cp.anova)
## Df Sum Sq Mean Sq F value Pr(>F)
## my_sample$college 3 458 152.55 1.58 0.199
## Residuals 96 9267 96.53
The Df different between the two analyses are 2 for island and 3 for colleges. The p-value in each case is 0.146 for islands and 0.644 for colleges. The null hypothesis in each case is there is no difference in means and because there is a difference in means I choose to reject the null hypothesis.
TukeyHSD(cp.anova)
## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = my_sample$control ~ my_sample$college)
##
## $`my_sample$college`
## diff lwr upr p adj
## Europa-Callisto -1.8066667 -8.763063 5.149730 0.9048137
## Ganymede-Callisto -5.5380952 -12.288161 1.211970 0.1464020
## Io-Callisto -2.2607843 -10.059040 5.537471 0.8730226
## Ganymede-Europa -3.7314286 -10.799852 3.336995 0.5147432
## Io-Europa -0.4541176 -8.529513 7.621278 0.9988590
## Io-Ganymede 3.2773109 -4.621040 11.175661 0.6995706
After interpreting these results, it is shown that the aliens from Io who attend Ganymede have the most significant difference from the rest of the islands and colleges.
bargraph.CI(x.factor = my_sample$antennae,
group = my_sample$color,
response = my_sample$food1,
ci.fun = ciMean, ylim = c(0, 20),
legend = T)
The patterns I see in this graph show no significant difference in the
consumption of food based on antennae and color.
sci.anova <- aov(my_sample$food1 ~ my_sample$color*my_sample$antennae)
summary(sci.anova)
## Df Sum Sq Mean Sq F value Pr(>F)
## my_sample$color 1 28.0 28.025 6.677 0.0113 *
## my_sample$antennae 1 2.8 2.843 0.677 0.4125
## my_sample$color:my_sample$antennae 1 19.0 18.956 4.516 0.0361 *
## Residuals 96 402.9 4.197
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
The Df in each row is 1. This proves these three factors do not seem to have much effect on one another. The null hypotheses is that in each case is there is no difference in means and because there is a difference in means I choose to reject the null hypothesis.
bargraph.CI(x.factor = my_sample$college,
group = my_sample$politics,
response = my_sample$intelligence,
ci.fun = ciMean, ylim = c(0, 300),
legend = T)
sci.anova <- aov(my_sample$intelligence ~ my_sample$college*my_sample$politics)
summary(sci.anova)
## Df Sum Sq Mean Sq F value Pr(>F)
## my_sample$college 3 5015 1671.6 48.392 <2e-16 ***
## my_sample$politics 2 108 54.0 1.563 0.2152
## my_sample$college:my_sample$politics 6 464 77.4 2.240 0.0466 *
## Residuals 88 3040 34.5
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
The degrees of freedom in the ANOVA table seem to be different for each category.The null hypotheses is that in each case is there is no difference in means and because there is a difference in means I choose to reject the null hypothesis.