library(tidyverse)
theme_set(theme_minimal())

setwd("~/Desktop/Geneva workshop 2023/Lab 1/Data")

data <- read_csv("clean_data.csv")

colnames(data)
 # [1] "ParticipantID" "ListNumber"    "TrialNumber"   "ReactionTime" 
 # [5] "PictureType"   "TrialType"     "Scale"         "Condition"    
 # [9] "Response"      "Block"


# Block 1

data %>%
  filter(Block == 1, Condition == "Baseline") %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    )
  ) %>%
  group_by(Scale) %>%
  summarise(Mean = mean(Response_f))

# Block 2

data %>%
  filter(Block == 2, TrialType == "Target") %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    )
  ) %>%
  group_by(Scale, Condition) %>%
  summarise(Mean = mean(Response_f))


# Putting the two together

data %>%
  filter(Condition == "Baseline" | TrialType == "Target") %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    )
  ) %>%
  group_by(Scale, Condition) %>%
  summarise(Mean = mean(Response_f)) %>%
  print() %>%
  ggplot(aes(x = Condition, y = Mean, fill = Condition)) +
    geom_col() +
    geom_text(aes(y = Mean + 0.03, label = format(round(Mean, digits = 2), nsmall=2)), show.legend = FALSE) +
    facet_wrap(~Scale) +
    theme(legend.position = "none")

# Filtering targets preceded by wrong answers to primes
data <- data %>%
  mutate(
    PrimeAccuracy = case_when(
      TrialType == "Target" ~ 2,
      TrialType == "Prime1" & Scale == "AdHoc" & Condition == "Weak" & str_count(Response, "1") == 2 ~ 1,
      TrialType == "Prime1" & Scale == "AdHoc" & Condition == "Strong" & str_count(Response, "1") == 1 ~ 1,
      TrialType == "Prime1" & Scale == "Numeral" & Condition == "Weak" & str_count(Response, "6") == 1 ~ 1,
      TrialType == "Prime1" & Scale == "Numeral" & Condition == "Strong" & str_count(Response, "4") == 1 ~ 1,
      TrialType == "Prime1" & Scale == "Some" & Condition == "Weak" & str_count(Response, "9") == 1 ~ 1,
      TrialType == "Prime1" & Scale == "Some" & Condition == "Strong" & str_count(Response, "6") == 1 ~ 1,
      #
      TrialType == "Prime2" & Scale == "AdHoc" & Condition == "Weak" & str_count(Response, "1") == 2 ~ 1,
      TrialType == "Prime2" & Scale == "AdHoc" & Condition == "Strong" & str_count(Response, "1") == 1 ~ 1,
      TrialType == "Prime2" & Scale == "Numeral" & Condition == "Weak" & str_count(Response, "6") == 1 ~ 1,
      TrialType == "Prime2" & Scale == "Numeral" & Condition == "Strong" & str_count(Response, "4") == 1 ~ 1,    
      TrialType == "Prime2" & Scale == "Some" & Condition == "Weak" & str_count(Response, "9") == 1 ~ 1,  
      TrialType == "Prime2" & Scale == "Some" & Condition == "Strong" & str_count(Response, "6") == 1 ~ 1,
      .default = 0
    ),
    Include = 0
  )

for (r in 1:nrow(data)) {
  if (data[r, "TrialType"] == "Prime1" & data[r, "PrimeAccuracy"] == 1) {
  	if (data[r+1, "TrialType"] == "Prime2" & data[r, "PrimeAccuracy"] == 1) {
  	  data[r+2, "Include"] <- 1	
  	}
  }
}
  
data %>%
  filter((Condition == "Baseline") | (TrialType == "Target" & Include == 1)) %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    )
  ) %>%
  group_by(Scale, Condition) %>%
  summarise(Mean = mean(Response_f)) %>%
  print() %>%
  ggplot(aes(x = Condition, y = Mean, fill = Condition)) +
    geom_col() +
    geom_text(aes(y = Mean + 0.03, label = format(round(Mean, digits = 2), nsmall=2)), show.legend = FALSE) +
    facet_wrap(~Scale) +
    theme(legend.position = "none")


library(lme4)

data %>%
  filter(Scale == "AdHoc" & (Condition == "Baseline" | (TrialType == "Target" & Include == 1))) %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    ),
    Condition = factor(Condition, levels = c("Strong", "Weak", "Baseline"))
  ) %>%
  glmer(formula = Response_f ~ Condition + (1 | ParticipantID), family = "binomial") %>%
  summary()

data %>%
  filter(Scale == "AdHoc" & (Condition == "Baseline" | (TrialType == "Target" & Include == 1))) %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    ),
    Condition = factor(Condition, levels = c("Weak", "Strong", "Baseline"))
  ) %>%
  glmer(formula = Response_f ~ Condition + (1 | ParticipantID), family = "binomial") %>%
  summary()

data %>%
  filter(Scale == "Some" & (Condition == "Baseline" | (TrialType == "Target" & Include == 1))) %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    ),
    Condition = factor(Condition, levels = c("Strong", "Weak", "Baseline"))
  ) %>%
  glmer(formula = Response_f ~ Condition + (1 | ParticipantID), family = "binomial", control=glmerControl(optimizer="bobyqa")) %>%
  summary()

data %>%
  filter(Scale == "Some" & (Condition == "Baseline" | (TrialType == "Target" & Include == 1))) %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    ),
    Condition = factor(Condition, levels = c("Weak", "Strong", "Baseline"))
  ) %>%
  glmer(formula = Response_f ~ Condition + (1 | ParticipantID), family = "binomial", control=glmerControl(optimizer="bobyqa")) %>%
  summary()

data %>%
  filter(Scale == "Numeral" & (Condition == "Baseline" | (TrialType == "Target" & Include == 1))) %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    ),
    Condition = factor(Condition, levels = c("Strong", "Weak", "Baseline"))
  ) %>%
  glmer(formula = Response_f ~ Condition + (1 | ParticipantID), family = "binomial", control=glmerControl(optimizer="bobyqa")) %>%
  summary()
  
data %>%
  filter(Scale == "Numeral" & (Condition == "Baseline" | (TrialType == "Target" & Include == 1))) %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    ),
    Condition = factor(Condition, levels = c("Weak", "Strong", "Baseline"))
  ) %>%
  glmer(formula = Response_f ~ Condition + (1 | ParticipantID), family = "binomial", control=glmerControl(optimizer="bobyqa")) %>%
  summary()
  
# 9 comparisons in total: 3 comparisons x 3 scales
# Alpha level = 0.05/9 = 0.00555...


# By participant analysis

library(ggridges)

data %>%
  filter((Condition == "Baseline") | (TrialType == "Target" & Include == 1)) %>%
  mutate(
    Response_f = case_match(Response,
      "cover.png" ~ 1,
      .default = 0
    )
  ) %>%
  group_by(ParticipantID, Scale, Condition) %>%
  summarise(Mean = mean(Response_f)) %>%
  ggplot(aes(x = Mean, y = Condition, height = after_stat(density), fill = Condition, colour = Condition)) +
    geom_density_ridges(stat = "binline", scale = 0.95, show.legend = FALSE, alpha = 0.2, binwidth = 0.1) +
    facet_wrap(~Scale) +
    theme(legend.position = "none")