Online Ellenbogen replication attempt (total n=110 after clearning)

group learningCriteria subs
sleep 66% 32
sleep 75% 24
wake 66% 31
wake 75% 23

Session 1 preprocessing

details.ellrep = file.info(list.files(path = "C:\\Users\\saptaf1\\Documents local\\ellrep-ti-task\\session1\\data", pattern = "*.csv", full.names = TRUE)) 
details.filtered.ellrep <- details.ellrep

data_all <- rownames(details.filtered.ellrep) %>%
  map_df(~read_csv(.x, col_types = cols(slider.response = col_character(),slider.rt = col_character(),fixation_cross.stopped = col_character(),stopFeedback.started =  col_character(),blockPerformance.started = col_character(), block_avg=col_number(),session=col_character(), block_avg_middlepair=col_number())))
## New names:
## • `` -> `...51`
session1.ellrep <- bind_rows(data_all) %>% mutate(stimCategory = case_when(
                                                                                          grepl(pattern = "faces", x = Image1) ~ "faces",
                                                                                          grepl(pattern = "objects", x = Image1) ~ "objects",
                                                                                          grepl(pattern = "scenes", x = Image1) ~ "scenes")) %>% filter(!is.na(Rank1))
## New names:
## • `...51` -> `...130`
session1.attention <- bind_rows(data_all) %>% mutate(stimCategory = case_when(
                                                                                          grepl(pattern = "faces", x = Image1) ~ "faces",
                                                                                          grepl(pattern = "objects", x = Image1) ~ "objects",
                                                                                          grepl(pattern = "scenes", x = Image1) ~ "scenes")) %>% filter(!is.na(num1))
## New names:
## • `...51` -> `...130`
session1.ellrep %>% group_by(participant) %>% summarise(blockcount = ifelse(sum(!is.na(key_resp_learning.keys))/10>10,10,sum(!is.na(key_resp_learning.keys)/10))) %>% filter(blockcount>=30) -> avgBlocks
hist(avgBlocks$blockcount)

session1.ellrep %>% group_by(participant) %>% summarise(blockcount = sum(!is.na(key_resp_learning.keys))/10) %>% filter(blockcount<=10)-> tmp

session1.avgs.ellrep <- session1.ellrep %>% select(participant,group, session, stimCategory, key_resp_im.corr, Rank1, Rank2) %>% mutate(cRank= paste0(pmin(Rank1,Rank2),pmax(Rank1,Rank2))) %>% group_by(participant,group, session,cRank,stimCategory) %>% summarise_each(funs(mean(key_resp_im.corr, na.rm = TRUE))) %>% filter(!is.na(session)) %>% select(-Rank1,-Rank2)

# for output

session1.ellrep %>% select(participant,group, session, stimCategory, key_resp_im.corr,key_resp_im.rt, Rank1, Rank2,date) %>% mutate(cRank= paste0(pmin(Rank1,Rank2),pmax(Rank1,Rank2))) %>% na.omit() %>% rename(corr=key_resp_im.corr, rt=key_resp_im.rt) %>% mutate(part="immediateTest") %>% mutate(date = lubridate::parse_date_time(date,'ymd HMS!')) %>% mutate(learningCriteria = ifelse(date>=as.Date('2022-11-23 10:30'),"75%","66%")) -> s1.im
s1.im %>% saveRDS("ellrep_tamas/session1_ellrep_trials_immediate_design-bws.RDS")

# NOTE!! 4xAFERQVTO not okay // 136 people reached criteria
s1.im %>% group_by(participant,group,date) %>% count() %>% nrow()
## [1] 136
session1.ellrep %>% select(participant,group, session, stimCategory, key_resp_learning.corr,key_resp_learning.rt, Rank1, Rank2,date) %>% mutate(part="learning") %>% mutate(date = lubridate::parse_date_time(date,'ymd HMS!')) %>% mutate(learningCriteria = ifelse(date>=as.Date('2022-11-23 10:30'),"75%","66%")) %>%  mutate(cRank= paste0(pmin(Rank1,Rank2),pmax(Rank1,Rank2))) %>% na.omit() %>% filter(date %in% s1.im$date ) ->tmp.learn 

tmp.learn  %>% saveRDS("ellrep_tamas/session1_ellrep_trials_learning_design-bws.RDS")

# NOTE!! 4xAFERQVTO not okay // 134 people reached criteria -weird discrepancy
tmp.learn %>% group_by(participant,group,date) %>% count() %>% nrow()
## [1] 134
#attention checks

session1.attention %>% select(participant,group, session, date,slider.response,desc,num1,num2,  totalsum)  %>% mutate(date = lubridate::parse_date_time(date,'ymd HMS!')) %>% na.omit() %>% filter(date %in% s1.im$date ) %>% saveRDS("ellrep_tamas/session1_ellrep_attentioncheck_design-bws.RDS")

#immediate recall performance
print(head(session1.avgs.ellrep))
## # A tibble: 6 × 6
## # Groups:   participant, group, session, cRank [6]
##   participant group session cRank stimCategory key_resp_im.corr
##   <chr>       <chr> <chr>   <chr> <chr>                   <dbl>
## 1 AFERQVTO    sleep 1       12    faces                  NaN   
## 2 AFERQVTO    sleep 1       23    faces                  NaN   
## 3 AFERQVTO    sleep 1       34    faces                  NaN   
## 4 AFERQVTO    sleep 1       45    faces                  NaN   
## 5 AFERQVTO    sleep 1       56    faces                  NaN   
## 6 AFERQVTO    wake  1       12    faces                    0.52
immediatePerformance.s1 <- session1.avgs.ellrep %>% mutate(pairType= ifelse(cRank %in% c("12","23","34","45","56"),"premise", ifelse(cRank %in% c("24","25","35"), "inference","anchor"))) %>% na.omit() %>% group_by(participant,group, pairType) %>% summarise(meanPerf=mean(key_resp_im.corr)) 
## `summarise()` has grouped output by 'participant', 'group'. You can override
## using the `.groups` argument.
grouped_ggbetweenstats(
  data = immediatePerformance.s1,
  x = group,
  y = meanPerf,
  grouping.var = pairType,
  type="p",
  bf.messsage = FALSE
)

s1.im %>% group_by(group,date,participant,learningCriteria) %>% summarise(meanPer=mean(corr))
## `summarise()` has grouped output by 'group', 'date', 'participant'. You can
## override using the `.groups` argument.
## # A tibble: 136 × 5
## # Groups:   group, date, participant [136]
##    group date                participant learningCriteria meanPer
##    <chr> <dttm>              <chr>       <chr>              <dbl>
##  1 sleep 2022-02-16 21:28:45 EYHPBRAT    66%                0.8  
##  2 sleep 2022-02-23 20:59:31 XUWCSLAG    66%                0.25 
##  3 sleep 2022-02-23 21:49:07 LWAHPZSK    66%                0.725
##  4 sleep 2022-02-24 21:05:51 ZDQXYIOU    66%                0.375
##  5 sleep 2022-02-24 21:11:05 OIEJGADV    66%                0.55 
##  6 sleep 2022-02-25 21:01:06 WLZGFQUD    66%                0.85 
##  7 sleep 2022-02-25 21:08:17 NCBXYVKS    66%                0.775
##  8 sleep 2022-02-25 21:09:00 LJXNTYCF    66%                0.5  
##  9 sleep 2022-02-26 21:20:59 USTGPEMF    66%                0.75 
## 10 sleep 2022-02-28 21:05:46 SJXWYEKQ    66%                0.575
## # … with 126 more rows
grouped_ggbetweenstats(
  data = s1.im %>% group_by(group,date,participant,learningCriteria) %>% summarise(meanPerf=mean(corr)),
  x = group,
  y = meanPerf,
  grouping.var = learningCriteria,
  type="p",
  bf.messsage = FALSE
)
## `summarise()` has grouped output by 'group', 'date', 'participant'. You can
## override using the `.groups` argument.

ggbetweenstats(
  data = s1.im %>% group_by(group,date,participant,learningCriteria) %>% summarise(meanPerf=mean(corr)) %>% mutate(condition=paste(group,learningCriteria)),
  x = condition,
  y = meanPerf,
  type="np",
  pairwise.display = "all",
  bf.messsage = FALSE,
  p.adjust.method   = "none"
)
## `summarise()` has grouped output by 'group', 'date', 'participant'. You can
## override using the `.groups` argument.
## Adding missing grouping variables: `group`, `date`, `participant`

Session 2 preprocessing

details.ellrep = file.info(list.files(path = "C:\\Users\\saptaf1\\Documents local\\ellrep-ti-task\\session2\\data",     
                       pattern = "*.csv", full.names = TRUE))
details.filtered.ellrep <- details.ellrep

data_all <- rownames(details.filtered.ellrep) %>%
  map_df(~read_csv(.x, col_types = cols(slider.response = col_character(),slider.rt = col_character(),fixation_cross.stopped = col_character(),stopFeedback.started =  col_character(),blockPerformance.started = col_character(), block_avg=col_number(), block_avg_middlepair=col_number())))

session2.ellrep <- bind_rows(data_all) %>% mutate(stimCategory = case_when(
                                                                                          grepl(pattern = "faces", x = Image1) ~ "faces",
                                                                                          grepl(pattern = "objects", x = Image1) ~ "objects",
                                                                                          grepl(pattern = "scenes", x = Image1) ~ "scenes")) %>% filter(!is.na(Rank1))

# session2.ellrep %>% group_by(participant) %>% summarise(blockcount = ifelse(sum(!is.na(key_resp_learning.keys))/10>10,10,sum(!is.na(key_resp_learning.keys)/10))) %>% filter(blockcount>=30) -> avgBlocks
# hist(avgBlocks$blockcount)

# session2.ellrep %>% group_by(participant) %>% summarise(blockcount = sum(!is.na(key_resp_learning.keys))/10) %>% filter(blockcount<=10)-> tmp

session2.ellrep %>% select(participant,group, session, stimCategory, key_resp_im.corr,key_resp_im.rt, Rank1, Rank2,date) %>% mutate(cRank= paste0(pmin(Rank1,Rank2),pmax(Rank1,Rank2))) %>% na.omit() %>% rename(corr=key_resp_im.corr, rt=key_resp_im.rt) %>% mutate(part="delayedTest") %>% mutate(date = lubridate::parse_date_time(date,'ymd HMS!')) %>%  mutate(learningCriteria = ifelse(date>=as.Date('2022-11-24'),"75%","66%"), session=2) %>% saveRDS("ellrep_tamas/session2_ellrep_trials_delayed_design-bws.RDS")

session2.ellrep %>% select(participant,group, session, stimCategory, key_resp_im.corr,key_resp_im.rt, Rank1, Rank2,date) %>% mutate(cRank= paste0(pmin(Rank1,Rank2),pmax(Rank1,Rank2))) %>% na.omit() %>% rename(corr=key_resp_im.corr, rt=key_resp_im.rt) %>% mutate(part="delayedTest") %>% mutate(date = lubridate::parse_date_time(date,'ymd HMS!')) %>%  mutate(learningCriteria = ifelse(date>=as.Date('2022-11-24'),"75%","66%"), session=2) %>% mutate(participant_u_id =paste(participant,group,date, sep ="_" ), distance=abs(Rank1-Rank2), pairType=ifelse(distance==1,"premise",ifelse(distance %in% c(2,3), "inference","anchor"))) %>% filter(group %in% c("wake","sleep"))-> session2.ellrep.tidy

session2.avgs.ellrep <- session2.ellrep %>% select(participant,group, session, stimCategory, key_resp_im.corr, Rank1, Rank2,date) %>% mutate(cRank= paste0(pmin(Rank1,Rank2),pmax(Rank1,Rank2))) %>% group_by(participant,group, session,cRank,stimCategory,date) %>% summarise_each(funs(mean(key_resp_im.corr, na.rm = TRUE))) %>% filter(!is.na(session)) %>% select(-Rank1,-Rank2) %>% mutate(date = lubridate::parse_date_time(date,'ymd HMS!')) %>% mutate(learningCriteria = ifelse(date>=as.Date('2022-11-24'),"75%","66%")) %>% filter(group != "ws")

# how many people per group
# time of when the raise to 75% happens: 2022-11-23_10h49.00.768.csv
session2.avgs.ellrep %>% group_by(participant,group,date,learningCriteria) %>% count() %>% ungroup() %>% group_by(group, learningCriteria) %>% count() %>% arrange(learningCriteria)
## # A tibble: 4 × 3
## # Groups:   group, learningCriteria [4]
##   group learningCriteria     n
##   <chr> <chr>            <int>
## 1 sleep 66%                 38
## 2 wake  66%                 34
## 3 sleep 75%                 29
## 4 wake  75%                 34
#for some reason there are 2s and 3s in this column.. investigate
session2.avgs.ellrep$session <-2

#immediate recall performance
print(length(unique(session2.avgs.ellrep$participant)))
## [1] 133
#immediate recall performance
print(head(session2.avgs.ellrep))
## # A tibble: 6 × 8
## # Groups:   participant, group, session, cRank, stimCategory [6]
##   participant group session cRank stimCate…¹ date                key_r…² learn…³
##   <chr>       <chr>   <dbl> <chr> <chr>      <dttm>                <dbl> <chr>  
## 1 AFERQVTO    sleep       2 12    faces      2022-11-23 18:43:13   0.375 66%    
## 2 AFERQVTO    sleep       2 16    faces      2022-11-23 18:43:13   0.875 66%    
## 3 AFERQVTO    sleep       2 23    faces      2022-11-23 18:43:13   0.625 66%    
## 4 AFERQVTO    sleep       2 24    faces      2022-11-23 18:43:13   0.5   66%    
## 5 AFERQVTO    sleep       2 25    faces      2022-11-23 18:43:13   0.375 66%    
## 6 AFERQVTO    sleep       2 34    faces      2022-11-23 18:43:13   0.5   66%    
## # … with abbreviated variable names ¹​stimCategory, ²​key_resp_im.corr,
## #   ³​learningCriteria
grouped_ggbetweenstats(
  data = session2.avgs.ellrep %>% mutate(pairType= ifelse(cRank %in% c("12","23","34","45","56"),"premise", ifelse(cRank %in% c("24","25","35"), "inference","anchor"))) %>% group_by(participant,group,learningCriteria, pairType) %>% summarise(meanPerf=mean(key_resp_im.corr)) %>% ungroup(),
  x = group,
  y = meanPerf,
  grouping.var = pairType,
  type="p",
  bf.messsage = FALSE
)
## `summarise()` has grouped output by 'participant', 'group', 'learningCriteria'.
## You can override using the `.groups` argument.

grouped_ggbetweenstats(
  data = session2.avgs.ellrep %>% mutate(pairType= ifelse(cRank %in% c("12","23","34","45","56"),"premise", ifelse(cRank %in% c("24","25","35"), "inference","anchor"))) %>% group_by(participant,group,learningCriteria, pairType) %>% summarise(meanPerf=mean(key_resp_im.corr)) %>% ungroup() %>% filter(learningCriteria== "66%"),
  x = group,
  y = meanPerf,
  grouping.var = pairType,
  type="np",
  bf.messsage = FALSE
)
## `summarise()` has grouped output by 'participant', 'group', 'learningCriteria'.
## You can override using the `.groups` argument.

grouped_ggbetweenstats(
  data = session2.avgs.ellrep %>% mutate(pairType= ifelse(cRank %in% c("12","23","34","45","56"),"premise", ifelse(cRank %in% c("24","25","35"), "inference","anchor"))) %>% group_by(participant,group,learningCriteria, pairType) %>% summarise(meanPerf=mean(key_resp_im.corr)) %>% ungroup() %>% filter(learningCriteria== "75%"),
  x = group,
  y = meanPerf,
  grouping.var = pairType,
  type="p",
  bf.messsage = FALSE
)
## `summarise()` has grouped output by 'participant', 'group', 'learningCriteria'.
## You can override using the `.groups` argument.

# proper filtering
s1.im %>% group_by(participant,group,date)%>% summarise(meanIm=mean(corr)) %>% filter(participant !="AFERQVTO") -> s1.im.aggr
## `summarise()` has grouped output by 'participant', 'group'. You can override
## using the `.groups` argument.
session2.ellrep.tidy  %>% group_by(participant,learningCriteria,group,date)%>%filter(pairType=="premise") %>%  summarise(meanDelayed=mean(corr)) %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% pivot_longer(c("meanIm","meanDelayed"), names_to = "session", values_to = "meanPerf")
## `summarise()` has grouped output by 'participant', 'learningCriteria', 'group'.
## You can override using the `.groups` argument.
## # A tibble: 270 × 7
## # Groups:   participant, learningCriteria, group [134]
##    parti…¹ learn…² group date.x              date.y              session meanP…³
##    <chr>   <chr>   <chr> <dttm>              <dttm>              <chr>     <dbl>
##  1 AFERQV… 66%     sleep 2022-11-23 18:43:13 NA                  meanIm   NA    
##  2 AFERQV… 66%     sleep 2022-11-23 18:43:13 NA                  meanDe…   0.525
##  3 AFERQV… 66%     wake  2022-02-08 17:41:31 NA                  meanIm   NA    
##  4 AFERQV… 66%     wake  2022-02-08 17:41:31 NA                  meanDe…   0.5  
##  5 AHSRJU… 66%     wake  2022-11-15 22:43:18 2022-11-15 10:27:37 meanIm    0.6  
##  6 AHSRJU… 66%     wake  2022-11-15 22:43:18 2022-11-15 10:27:37 meanDe…   0.375
##  7 AKPJDW… 75%     wake  2022-12-06 04:47:34 2022-12-05 09:07:45 meanIm    0.95 
##  8 AKPJDW… 75%     wake  2022-12-06 04:47:34 2022-12-05 09:07:45 meanDe…   0.775
##  9 AKPJDW… 75%     wake  2022-12-09 22:45:00 2022-12-09 09:08:27 meanIm    0.35 
## 10 AKPJDW… 75%     wake  2022-12-09 22:45:00 2022-12-09 09:08:27 meanDe…   0.475
## # … with 260 more rows, and abbreviated variable names ¹​participant,
## #   ²​learningCriteria, ³​meanPerf
grouped_ggwithinstats(
  data = session2.ellrep.tidy  %>% group_by(participant,learningCriteria,group,date)%>%filter(pairType=="premise") %>%  summarise(meanDelayed=mean(corr)) %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% pivot_longer(c("meanIm","meanDelayed"), names_to = "session", values_to = "meanPerf"),
  x = session,
  y = meanPerf,
  grouping.var = learningCriteria,
  bf.message=FALSE,
  type="p"
)
## `summarise()` has grouped output by 'participant', 'learningCriteria', 'group'.
## You can override using the `.groups` argument.

session2.ellrep.tidy  %>% group_by(participant,learningCriteria,group,date)%>%filter(pairType=="premise") %>%  summarise(meanDelayed=mean(corr)) %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% pivot_longer(c("meanIm","meanDelayed"), names_to = "session", values_to = "meanPerf") %>% ungroup() %>% filter(participant !="AFERQVTO")-> anova_dat
## `summarise()` has grouped output by 'participant', 'learningCriteria', 'group'.
## You can override using the `.groups` argument.
# library(rstatix)
# res.aov <- anova_test(
#   data = session2.ellrep.tidy  %>% group_by(participant,learningCriteria,group,date)%>%filter(pairType=="premise") %>%  summarise(meanDelayed=mean(corr)) %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% pivot_longer(c("meanIm","meanDelayed"), names_to = "session", values_to = "meanPerf") %>% ungroup() %>% filter(participant %in% subs.normalrange) %>% mutate(participant=paste0(participant,date.x)), dv = meanPerf, wid = participant,
#   between = c(group, learningCriteria), within = session)
# get_anova_table(res.aov)

library(afex)
## Loading required package: lme4
## Loading required package: Matrix
## 
## Attaching package: 'Matrix'
## 
## The following objects are masked from 'package:tidyr':
## 
##     expand, pack, unpack
## 
## ************
## Welcome to afex. For support visit: http://afex.singmann.science/
## - Functions for ANOVAs: aov_car(), aov_ez(), and aov_4()
## - Methods for calculating p-values with mixed(): 'S', 'KR', 'LRT', and 'PB'
## - 'afex_aov' and 'mixed' objects can be passed to emmeans() for follow-up tests
## - NEWS: emmeans() for ANOVA models now uses model = 'multivariate' as default.
## - Get and set global package options with: afex_options()
## - Set orthogonal sum-to-zero contrasts globally: set_sum_contrasts()
## - For example analyses see: browseVignettes("afex")
## ************
## 
## Attaching package: 'afex'
## 
## The following object is masked from 'package:lme4':
## 
##     lmer
(aw <- aov_ez("participant", "meanPerf", anova_dat, within = c("session"), between=c("group", "learningCriteria")))
## Converting to factor: group, learningCriteria
## Contrasts set to contr.sum for the following variables: group, learningCriteria
## Anova Table (Type 3 tests)
## 
## Response: meanPerf
##                           Effect     df  MSE       F   ges p.value
## 1                          group 1, 125 0.06    0.17  .001    .685
## 2               learningCriteria 1, 125 0.06    1.43  .010    .234
## 3         group:learningCriteria 1, 125 0.06    2.69  .019    .103
## 4                        session 1, 125 0.01 9.47 **  .006    .003
## 5                  group:session 1, 125 0.01  3.58 +  .002    .061
## 6       learningCriteria:session 1, 125 0.01    0.02 <.001    .893
## 7 group:learningCriteria:session 1, 125 0.01    2.41  .002    .123
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1
afex_plot(aw, x = "session", trace = c("learningCriteria", "group")) 

session2.ellrep.tidy.ext <- session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant!="AFERQVTO") %>% mutate(timeelapsed=date.x-date.y)

session2.ellrep.tidy.ext %>% group_by(participant,group,date.x) %>% summarise(timeelapsed=unique(timeelapsed)) %>% ggplot(., aes(timeelapsed)) +
  geom_histogram()
## `summarise()` has grouped output by 'participant', 'group'. You can override
## using the `.groups` argument.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

minTime<-480
maxTime<-840
session2.ellrep.tidy.ext %>% group_by(participant,group,date.x) %>% summarise(timeelapsed=unique(timeelapsed)) %>%  filter(timeelapsed>minTime,timeelapsed<maxTime)  %>% ggplot(., aes(timeelapsed)) +
  geom_histogram()
## `summarise()` has grouped output by 'participant', 'group'. You can override
## using the `.groups` argument.
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

session2.ellrep.tidy.ext %>% group_by(participant,group,date.x) %>% summarise(timeelapsed=unique(timeelapsed)) %>%  filter(timeelapsed>minTime,timeelapsed<maxTime) -> subs.normalrange
## `summarise()` has grouped output by 'participant', 'group'. You can override
## using the `.groups` argument.
grouped_ggbetweenstats(
  data = session2.avgs.ellrep %>% mutate(pairType= ifelse(cRank %in% c("12","23","34","45","56"),"premise", ifelse(cRank %in% c("24","25","35"), "inference","anchor"))) %>% group_by(participant,group,learningCriteria, pairType) %>% summarise(meanPerf=mean(key_resp_im.corr)) %>% ungroup() %>% filter(participant %in% subs.normalrange$participant),
  x = group,
  y = meanPerf,
  grouping.var = pairType,
  type="r",
  bf.messsage = FALSE
)
## `summarise()` has grouped output by 'participant', 'group', 'learningCriteria'.
## You can override using the `.groups` argument.

grouped_ggbetweenstats(
  data = session2.avgs.ellrep %>% mutate(pairType= ifelse(cRank %in% c("12","23","34","45","56"),"premise", ifelse(cRank %in% c("24","25","35"), "inference","anchor"))) %>% group_by(participant,group,learningCriteria, pairType) %>% summarise(meanPerf=mean(key_resp_im.corr)) %>% ungroup() %>% filter(learningCriteria== "66%")  %>% filter(participant %in% subs.normalrange$participant),
  x = group,
  y = meanPerf,
  grouping.var = pairType,
  type="p",
  bf.messsage = FALSE
)
## `summarise()` has grouped output by 'participant', 'group', 'learningCriteria'.
## You can override using the `.groups` argument.

grouped_ggbetweenstats(
  data = session2.avgs.ellrep %>% mutate(pairType= ifelse(cRank %in% c("12","23","34","45","56"),"premise", ifelse(cRank %in% c("24","25","35"), "inference","anchor"))) %>% group_by(participant,group,learningCriteria, pairType) %>% summarise(meanPerf=mean(key_resp_im.corr)) %>% ungroup() %>% filter(learningCriteria== "75%")  %>% filter(participant %in% subs.normalrange$participant),
  x = group,
  y = meanPerf,
  grouping.var = pairType,
  type="p",
  bf.messsage = FALSE
)
## `summarise()` has grouped output by 'participant', 'group', 'learningCriteria'.
## You can override using the `.groups` argument.

Qualtrics

library(qualtRics)

qualtrics_api_credentials(api_key = "ysZ6HezYt3Wc43hoFcwSG82LoMSbFHTqb0x1FWdW",
                          base_url = "cardiffunipsych.fra1.qualtrics.com",
                          install = TRUE)


awareness_survey.ellrep <-fetch_survey(surveyID = "SV_7armt1volf3zVfE", 
                         verbose = TRUE)

awareness_survey.ellrep %>% saveRDS("ellrep_tamas/session2_ellrep_awarenesssurvey_design-bws.RDS")

s1start.ellrep <-fetch_survey(surveyID = "SV_6lJ6YCFlcjdao8S", 
                         verbose = TRUE)

s1start.starttime.ellrep <- s1start.ellrep %>% group_by(code) %>% summarise(starttime=max(EndDate))

s1finish.ellrep <-fetch_survey(surveyID = "SV_2rTYD8fdqbPRWJM", 
                         verbose = TRUE)
s1finish.starttime.ellrep <- s1finish.ellrep %>% group_by(code) %>% summarise(endtime=max(EndDate))


s1start.starttime.ellrep %>% left_join(s1finish.starttime.ellrep) %>% mutate(total = endtime-starttime) %>% filter(code %in% awareness_survey.ellrep$code ) -> totalTimes

s2start.ellrep <-fetch_survey(surveyID = "SV_9phc9enBGE9abbg", 
                         verbose = TRUE)

s2start.starttime.ellrep <- s2start.ellrep %>% group_by(code) %>% summarise(starttime=max(EndDate))

s2finish.ellrep <-fetch_survey(surveyID = "SV_7armt1volf3zVfE", 
                         verbose = TRUE)
s2finish.starttime.ellrep <- s2finish.ellrep %>% group_by(code) %>% summarise(endtime=max(EndDate))

s2start.starttime.ellrep %>% left_join(s2finish.starttime.ellrep) %>% mutate(total = endtime-starttime) %>% filter(code %in% awareness_survey.ellrep$code ) -> totalTimes

#check whether time between S1.end and S2.start is approx 12h
s2start.starttime.ellrep %>% left_join(s1finish.starttime.ellrep) %>% mutate(total = difftime(starttime, endtime, units='hours')) %>% filter(code %in% awareness_survey.ellrep$code ) -> totaltimebetweensessions

hist(as.numeric(totaltimebetweensessions$total))

awareness_survey.ellrep %>% select(code,SC0) %>% rename("awareness"="SC0") -> awareness_score.ellrep

hist(awareness_score.ellrep$awareness)

Stats

library(tidyverse) # data wrangling and visualization
library(sjPlot)    # to visualizing mixed-effects models
# library(effects)   # to visualizing mixed-effects models
library(lme4)      # "golden standard" for mixed-effects modelling in R (no p-values)
library(lmerTest)  # p-values for MEMs based on the Satterthwaite approximation
## 
## Attaching package: 'lmerTest'
## The following object is masked from 'package:lme4':
## 
##     lmer
## The following object is masked from 'package:stats':
## 
##     step
library(report)    # mainly for an "report" function
# library(emmeans)   # post-hoc analysis
# library(knitr)     # beautifying tables
# library(sjstats)   # ICC - intraclass-correlation coefficient
# library(caret)     # ML, model comparison & utility functions

str(session2.ellrep.tidy)
## tibble [9,720 × 15] (S3: tbl_df/tbl/data.frame)
##  $ participant     : chr [1:9720] "AFERQVTO" "AFERQVTO" "AFERQVTO" "AFERQVTO" ...
##  $ group           : chr [1:9720] "wake" "wake" "wake" "wake" ...
##  $ session         : num [1:9720] 2 2 2 2 2 2 2 2 2 2 ...
##  $ stimCategory    : chr [1:9720] "faces" "faces" "faces" "faces" ...
##  $ corr            : num [1:9720] 0 0 1 1 0 0 0 1 0 0 ...
##  $ rt              : num [1:9720] 1.168 0.244 0.154 0.177 0.134 ...
##  $ Rank1           : num [1:9720] 5 5 2 1 6 4 3 2 2 5 ...
##  $ Rank2           : num [1:9720] 3 2 3 2 1 3 2 5 1 4 ...
##  $ date            : POSIXct[1:9720], format: "2022-02-08 17:41:31" "2022-02-08 17:41:31" ...
##  $ cRank           : chr [1:9720] "35" "25" "23" "12" ...
##  $ part            : chr [1:9720] "delayedTest" "delayedTest" "delayedTest" "delayedTest" ...
##  $ learningCriteria: chr [1:9720] "66%" "66%" "66%" "66%" ...
##  $ participant_u_id: chr [1:9720] "AFERQVTO_wake_2022-02-08 17:41:31" "AFERQVTO_wake_2022-02-08 17:41:31" "AFERQVTO_wake_2022-02-08 17:41:31" "AFERQVTO_wake_2022-02-08 17:41:31" ...
##  $ distance        : num [1:9720] 2 3 1 1 5 1 1 3 1 1 ...
##  $ pairType        : chr [1:9720] "inference" "inference" "premise" "premise" ...
session2.ellrep.tidy %>% group_by(group, learningCriteria) %>% summarise(subs = n_distinct(participant_u_id))
## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.
## # A tibble: 4 × 3
## # Groups:   group [2]
##   group learningCriteria  subs
##   <chr> <chr>            <int>
## 1 sleep 66%                 38
## 2 sleep 75%                 29
## 3 wake  66%                 34
## 4 wake  75%                 34
#good plot
ggplot(aes(x=group, y=corr, group=interaction(learningCriteria,pairType), color=pairType, linetype=learningCriteria), data=session2.ellrep.tidy) + 
  stat_summary(fun.data="mean_cl_boot", geom='line') + ggtitle("All data..") +   stat_summary(fun.data=mean_cl_boot,fun.args=list(conf.int=.68), geom='errorbar', width=0.1) + ylab("Inference %")

# grouped_ggbetweenstats(
#   data = session2.ellrep.tidy %>% filter(learningCriteria=="75%",participant %in% subs.normalrange$participant)%>% group_by(group, pairType,participant_u_id) %>% summarise(meanPerf= mean(corr)),
#   x = group,
#   y = meanPerf,
#   grouping.var = pairType,
#   type="r",
#   bf.messsage = FALSE
# )

ggplot(aes(x=group, y=corr, group=interaction(learningCriteria,pairType), color=pairType, linetype=learningCriteria), data=session2.ellrep.tidy%>% filter(participant %in% subs.normalrange$participant)) + 
  stat_summary(fun.data="mean_cl_boot", geom='line') +   stat_summary(fun.data=mean_cl_boot,fun.args=list(conf.int=.68), geom='errorbar', width=0.1) + ggtitle("Filtered data..") + ylab("Inference %")

#p=0.064 groupXpairType interaction
m1 <- glmer(corr ~ group*pairType  + (1 | participant_u_id ), data = session2.ellrep.tidy%>% filter(participant %in% subs.normalrange$participant, pairType %in% c("premise","inference")), family = binomial, control = glmerControl(optimizer = "bobyqa"))



m2 <- glmer(corr ~ group*pairType  + meanIm + (1 | participant_u_id ), data = session2.ellrep.tidy%>% filter(participant %in% subs.normalrange$participant,pairType %in% c("premise","inference")) %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")), family = binomial, control = glmerControl(optimizer = "bobyqa"))
tab_model(m1,m2)
  corr corr
Predictors Odds Ratios CI p Odds Ratios CI p
(Intercept) 1.45 1.12 – 1.87 0.005 0.12 0.08 – 0.20 <0.001
group [wake] 0.82 0.57 – 1.18 0.288 0.82 0.63 – 1.07 0.137
pairType [premise] 1.81 1.56 – 2.11 <0.001 1.81 1.56 – 2.10 <0.001
group [wake] × pairType
[premise] 		1.09 0.88 – 1.35 0.426 1.09 0.88 – 1.35 0.412
meanIm 32.08 16.81 – 61.22 <0.001
Random Effects
σ2 3.29 3.29
Ï„00 0.77 participant_u_id 0.30 participant_u_id
ICC 0.19 0.08
N 110 participant_u_id 110 participant_u_id
Observations 7040 7040
Marginal R2 / Conditional R2 0.024 / 0.209 0.123 / 0.196 
#mean inf sig above chance
m1.inf <- glmer(corr ~ group + learningCriteria + (1 | participant_u_id ), data = session2.ellrep.tidy%>% filter(participant %in% subs.normalrange$participant) %>% filter(pairType == "inference"), family = binomial, control = glmerControl(optimizer = "bobyqa"))
tab_model(m1.inf)
  corr
Predictors Odds Ratios CI p
(Intercept) 1.57 1.05 – 2.36 0.027
group [wake] 0.80 0.49 – 1.31 0.379
learningCriteria [75%] 0.90 0.55 – 1.48 0.684
Random Effects
σ2 3.29
Ï„00 participant_u_id 1.47
ICC 0.31
N participant_u_id 110
Observations 2640
Marginal R2 / Conditional R2 0.003 / 0.311 
# relevant

m1.inf.subset <- glmer(corr ~ group+ (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference", learningCriteria =="75%") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m2.inf.subset <- glmer(corr ~ group+meanIm+ (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference", learningCriteria =="75%") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m3.inf.subset <- glmer(corr ~ group*meanIm+ (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference", learningCriteria =="75%") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

tab_model(m1.inf.subset,m2.inf.subset,m3.inf.subset)
  corr corr corr
Predictors Odds Ratios CI p Odds Ratios CI p Odds Ratios CI p
(Intercept) 1.60 0.87 – 2.95 0.132 0.25 0.05 – 1.36 0.109 0.14 0.01 – 1.96 0.143
group [wake] 0.66 0.27 – 1.57 0.345 0.71 0.31 – 1.63 0.418 1.82 0.07 – 50.29 0.724
meanIm 12.33 1.42 – 107.41 0.023 27.95 0.81 – 958.91 0.065
group [wake] × meanIm 0.27 0.00 – 23.38 0.566
Random Effects
σ2 3.29 3.29 3.29
Ï„00 2.07 participant_u_id 1.82 participant_u_id 1.82 participant_u_id
ICC 0.39 0.36 0.36
N 47 participant_u_id 47 participant_u_id 47 participant_u_id
Observations 1128 1128 1128
Marginal R2 / Conditional R2 0.008 / 0.391 0.051 / 0.389 0.054 / 0.391 
# relevant

m1.inf <- glmer(corr ~ group+ (1 | participant_u_id ) + (1 |stimCategory), data = session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m2.inf <- glmer(corr ~ group+meanIm+ (1 | participant_u_id ) + (1 |stimCategory), data = session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m3.inf <- glmer(corr ~ group*meanIm+ (1 | participant_u_id )+ (1 |stimCategory), data = session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

tab_model(m1.inf,m2.inf,m3.inf)
  corr corr corr
Predictors Odds Ratios CI p Odds Ratios CI p Odds Ratios CI p
(Intercept) 1.48 0.98 – 2.24 0.064 0.32 0.12 – 0.84 0.021 0.25 0.06 – 1.01 0.052
group [wake] 0.80 0.49 – 1.30 0.368 0.80 0.50 – 1.27 0.345 1.25 0.19 – 8.21 0.816
meanIm 9.07 2.48 – 33.14 0.001 13.06 1.83 – 93.30 0.010
group [wake] × meanIm 0.53 0.04 – 7.15 0.630
Random Effects
σ2 3.29 3.29 3.29
Ï„00 1.44 participant_u_id 1.30 participant_u_id 1.31 participant_u_id
0.04 stimCategory 0.01 stimCategory 0.01 stimCategory
ICC 0.31 0.29 0.29
N 110 participant_u_id 110 participant_u_id 110 participant_u_id
3 stimCategory 3 stimCategory 3 stimCategory
Observations 2640 2640 2640
Marginal R2 / Conditional R2 0.003 / 0.312 0.037 / 0.312 0.038 / 0.313 
m1.inf.ext <- glmer(corr ~ group*meanIm + (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant) %>% na.omit(), family = binomial, control = glmerControl(optimizer = "bobyqa"))
tab_model(m1.inf.ext)
  corr
Predictors Odds Ratios CI p
(Intercept) 0.24 0.06 – 0.97 0.045
group [wake] 1.30 0.20 – 8.42 0.783
meanIm 13.58 1.96 – 94.12 0.008
group [wake] × meanIm 0.50 0.04 – 6.57 0.595
Random Effects
σ2 3.29
Ï„00 participant_u_id 1.32
ICC 0.29
N participant_u_id 110
Observations 2640
Marginal R2 / Conditional R2 0.038 / 0.314 
m1.inf.ext.sleep <- glmer(corr ~ learningCriteria+meanIm + (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference", group=="sleep") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant) %>% na.omit(), family = binomial, control = glmerControl(optimizer = "bobyqa"))
tab_model(m1.inf.ext.sleep)
  corr
Predictors Odds Ratios CI p
(Intercept) 0.25 0.06 – 0.97 0.046
learningCriteria [75%] 0.90 0.46 – 1.75 0.761
meanIm 14.07 2.01 – 98.68 0.008
Random Effects
σ2 3.29
Ï„00 participant_u_id 1.28
ICC 0.28
N participant_u_id 56
Observations 1344
Marginal R2 / Conditional R2 0.043 / 0.311 
session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant")) %>% filter(participant %in% subs.normalrange$participant) %>% na.omit() %>% group_by(group.x,learningCriteria) %>% summarise(subs = n_distinct(participant_u_id))
## `summarise()` has grouped output by 'group.x'. You can override using the
## `.groups` argument.
## # A tibble: 4 × 3
## # Groups:   group.x [2]
##   group.x learningCriteria  subs
##   <chr>   <chr>            <int>
## 1 sleep   66%                 32
## 2 sleep   75%                 24
## 3 wake    66%                 31
## 4 wake    75%                 23
plot_model(m1.inf.ext,"pred", terms = c("meanIm[all]","group"))

plot_model(m1.inf.ext,"int", terms = c("group"))

Symbolic distance

# relevant

m0.inf.subset.d <- glmer(corr ~ distance + (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference", learningCriteria =="75%") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m1.inf.subset.d <- glmer(corr ~ group+distance+ (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference", learningCriteria =="75%") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m2.inf.subset.d <- glmer(corr ~ group*distance+meanIm+ (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference", learningCriteria =="75%") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

tab_model(m0.inf.subset.d,m1.inf.subset.d,m2.inf.subset.d)
  corr corr corr
Predictors Odds Ratios CI p Odds Ratios CI p Odds Ratios CI p
(Intercept) 1.52 0.68 – 3.39 0.307 1.86 0.75 – 4.61 0.179 0.42 0.06 – 2.91 0.382
distance 0.94 0.70 – 1.25 0.656 0.94 0.70 – 1.25 0.656 0.80 0.54 – 1.19 0.271
group [wake] 0.66 0.27 – 1.57 0.345 0.33 0.07 – 1.59 0.167
meanIm 12.37 1.41 – 108.22 0.023
group [wake] × distance 1.39 0.78 – 2.46 0.261
Random Effects
σ2 3.29 3.29 3.29
Ï„00 2.12 participant_u_id 2.07 participant_u_id 1.83 participant_u_id
ICC 0.39 0.39 0.36
N 47 participant_u_id 47 participant_u_id 47 participant_u_id
Observations 1128 1128 1128
Marginal R2 / Conditional R2 0.000 / 0.392 0.008 / 0.391 0.052 / 0.391 
# relevant

m1.inf.d <- glmer(corr ~ group*distance+ (1 | participant_u_id ) + (1 | stimCategory), data = session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m2.inf.d <- glmer(corr ~ group*distance+meanIm+ (1 | participant_u_id ) + (1 | stimCategory), data = session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m3.inf.d <- glmer(corr ~ group*meanIm*distance+ (1 | participant_u_id ), data = session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

m4.inf.d <- glmer(corr ~ group*distance+ meanIm + (1 | participant_u_id ) + (1 | stimCategory), data = session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant), family = binomial, control = glmerControl(optimizer = "bobyqa"))

tab_model(m1.inf.d,m2.inf.d,m3.inf.d)
  corr corr corr
Predictors Odds Ratios CI p Odds Ratios CI p Odds Ratios CI p
(Intercept) 1.71 0.83 – 3.54 0.145 0.37 0.12 – 1.15 0.086 0.45 0.03 – 7.90 0.586
group [wake] 0.40 0.15 – 1.07 0.069 0.40 0.15 – 1.06 0.065 5.56 0.12 – 255.84 0.380
distance 0.94 0.73 – 1.21 0.629 0.94 0.73 – 1.21 0.630 0.77 0.26 – 2.26 0.628
group [wake] × distance 1.34 0.93 – 1.93 0.113 1.34 0.93 – 1.93 0.113 0.53 0.13 – 2.26 0.394
meanIm 9.10 2.48 – 33.33 0.001 6.78 0.12 – 391.49 0.355
group [wake] × meanIm 0.02 0.00 – 4.88 0.167
meanIm × distance 1.35 0.29 – 6.29 0.702
(group [wake] × meanIm) ×
distance 		3.79 0.49 – 29.13 0.200
Random Effects
σ2 3.29 3.29 3.29
Ï„00 1.44 participant_u_id 1.31 participant_u_id 1.33 participant_u_id
0.04 stimCategory 0.01 stimCategory  
ICC 0.31 0.29 0.29
N 110 participant_u_id 110 participant_u_id 110 participant_u_id
3 stimCategory 3 stimCategory  
Observations 2640 2640 2640
Marginal R2 / Conditional R2 0.004 / 0.313 0.038 / 0.314 0.042 / 0.318 
anova(m1.inf.d,m2.inf.d,m3.inf.d)
## Data: session2.ellrep.tidy %>% filter(pairType == "inference") %>%  ...
## Models:
## m1.inf.d: corr ~ group * distance + (1 | participant_u_id) + (1 | stimCategory)
## m2.inf.d: corr ~ group * distance + meanIm + (1 | participant_u_id) + (1 | stimCategory)
## m3.inf.d: corr ~ group * meanIm * distance + (1 | participant_u_id)
##          npar    AIC    BIC  logLik deviance   Chisq Df Pr(>Chisq)   
## m1.inf.d    6 3249.6 3284.9 -1618.8   3237.6                         
## m2.inf.d    7 3241.0 3282.1 -1613.5   3227.0 10.6020  1    0.00113 **
## m3.inf.d    9 3239.0 3291.9 -1610.5   3221.0  5.9872  2    0.05011 . 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

More distance

session2.ellrep.tidy %>% filter(pairType == "inference", learningCriteria =="75%") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant) %>% group_by(group,participant_u_id,learningCriteria,cRank) %>% summarise(meanPerf = mean(corr)) -> tmp
## `summarise()` has grouped output by 'group', 'participant_u_id',
## 'learningCriteria'. You can override using the `.groups` argument.
ggstatsplot::grouped_ggbetweenstats(data = tmp, x=group,y=meanPerf, grouping.var = cRank, type="np")

session2.ellrep.tidy %>% filter(pairType == "inference") %>% left_join(s1.im.aggr, by = c("participant"="participant","group"="group")) %>% filter(participant %in% subs.normalrange$participant) %>% group_by(group,participant_u_id,learningCriteria,cRank) %>% summarise(meanPerf = mean(corr)) -> tmp
## `summarise()` has grouped output by 'group', 'participant_u_id',
## 'learningCriteria'. You can override using the `.groups` argument.
ggstatsplot::grouped_ggbetweenstats(data = tmp, x=group,y=meanPerf, grouping.var = cRank, type="r")