# February 21

library(tidyverse)
library(ggeffects)

d <- read.csv("http://whlevine.hosted.uark.edu/psyc5143/PBmeat.csv")

# contrast code the meat factor & the PB factor & create a product term
d <- d %>% 
	mutate(meatC = ifelse(meat == "none", -1/2, 1/2),
				 PBC = ifelse(PB == "none", -1/2, 1/2),
				 int = meatC*PBC)

summary(lm(tastiness ~ meatC + PBC + int, d))

# creating a decent plot (still not great, but better?)
model.plot <- lm(tastiness ~ meat*PB, d)
plot1 <- ggemmeans(model.plot, terms = c("PB", "meat")) %>% 
	ggplot() + theme_bw() +
	geom_line(aes(x = x, y = predicted, group = group, color = group))

plot1

# dummy codes
d <- d %>% 
	mutate(meatD = ifelse(meat == "none", 0, 1),
				 PBD = ifelse(PB == "none", 0, 1))

summary(lm(tastiness ~ meatD*PBD, d), t = F)

# larger design

scrambled <- read.csv("http://whlevine.hosted.uark.edu/psyc5143/scrambled.csv")

# what's in there?
glimpse(scrambled)

# cell means
scrambled %>% group_by(text, wpm) %>%
	summarise(M = mean(dv),
						SD = sd(dv),
						n = length(dv))

# marginal means
scrambled %>% group_by(text) %>% summarise(M = mean(dv))
scrambled %>% group_by(wpm) %>% summarise(M = mean(dv))

# contrast codes
scrambled <- scrambled %>% 
	mutate(T = ifelse(text == "intact", 1/2, -1/2),
				 R1 = ifelse(wpm == "600 wpm", -2/3, 1/3),
				 R2 = case_when(wpm == "300 wpm" ~ 1/2,
				 							 wpm == "450 wpm" ~ -1/2,
				 							 wpm == "600 wpm" ~ 0),
				 TR1 = T*R1,
				 TR2 = T*R2)

model.contrasts <- lm(dv ~ T + R1 + R2 + TR1 + TR2, scrambled)
summary(model.contrasts)
anova(model.contrasts)

# plot!
model.default <- lm(dv ~ text*wpm, scrambled)
plot2 <- ggemmeans(model.default, terms = c("wpm", "text")) %>% 
	ggplot() + theme_bw() +
	geom_line(aes(x = x, y = predicted, group = group, color = group))

plot2