Reasons for Exclusion

source("0_helpers.R")
library(tidylog)
load("data/cleaned.rdata")

library(knitr)
opts_chunk$set(fig.width = 9, fig.height = 7, cache = T, warning = T, message = T, cache = F, error = FALSE)

comma_separated_to_columns <- function(df, col) {
  colname <- deparse(substitute(col))
  df$splitcol <- df %>% pull(colname)
  separate_rows(df, splitcol, convert = TRUE, sep = ", ") %>% 
    mutate(splitcol = if_else(is.na(splitcol), "no", 
                        if_else(splitcol == "" | 
                                  splitcol %in% c(), "included", as.character(splitcol)))) %>% 
    mutate(#splitcol = stringr::str_c(colname, "_", splitcol), 
           value = 1) %>% 
    spread(splitcol, value, fill = 0) %>% 
    select(-colname)
}

all_survey_length <- nrow(all_surveys)
diary_length <- nrow(diary)
diary_social_length <- nrow(diary_social)

diary$reasons_for_exclusion_diary <- ""

Lab participants that did not do the online diary (not merged)

diary <- diary %>% 
  mutate(reasons_for_exclusion_diary = str_c(reasons_for_exclusion_diary,
                                            if_else(is.na(session), "lab_only, ", "", "")
                                       )
  )

## mutate: no changes

Disclosed that they responded dishonestly on that day.

diary <- diary %>% 
  mutate(reasons_for_exclusion_diary = str_c(reasons_for_exclusion_diary,
                                            if_else(dishonest_discard == 1, "dishonest_answer, ", "", "")
                                       )
  )

## mutate: changed 150 values (<1%) of 'reasons_for_exclusion_diary' (0 new NA)

Did not finish diary entry.

diary <- diary %>% 
  mutate(reasons_for_exclusion_diary = str_c(reasons_for_exclusion_diary,
                                            if_else(is.na(ended_diary) & !is.na(modified_diary), "did_not_finish_entry, ", "", "")
                                       )
  )

## mutate: changed 856 values (1%) of 'reasons_for_exclusion_diary' (0 new NA)

Cycle shorter than 20 days.

diary <- diary %>% 
  mutate(reasons_for_exclusion_diary = str_c(reasons_for_exclusion_diary,
                                            if_else(coalesce(minimum_cycle_length_diary, as.numeric(menstruation_length)) < 20, "cycle_shorter_than_20, ", "", "")
                                       )
  )

## mutate: changed 4,736 values (6%) of 'reasons_for_exclusion_diary' (0 new NA)

Cycle longer than 40 days.

diary <- diary %>% 
  mutate(reasons_for_exclusion_diary = str_c(reasons_for_exclusion_diary,
                                            if_else(coalesce(minimum_cycle_length_diary, as.numeric(menstruation_length)) > 40, "cycle_longer_than_40, ", "", "")
                                       )
  )

## mutate: changed 7,461 values (10%) of 'reasons_for_exclusion_diary' (0 new NA)

Next menstrual onset not observed

diary <- diary %>% 
  mutate(reasons_for_exclusion_diary = str_c(reasons_for_exclusion_diary,
                                            if_else(
                                              menstruation_regular == 1 &
                                              coalesce(minimum_cycle_length_diary, as.numeric(menstruation_length)) <= 40 &
                                              coalesce(minimum_cycle_length_diary, as.numeric(menstruation_length)) >= 20 &
                                              is.na(fertile_fab), "next_menstrual_onset_unobserved, ", "", "")
                                       )
  )

## mutate: changed 4,100 values (5%) of 'reasons_for_exclusion_diary' (0 new NA)

Skipped this diary day (days after dropping out not included)

diary <- diary %>% 
  mutate(reasons_for_exclusion_diary = str_c(reasons_for_exclusion_diary,
                                            if_else(is.na(ended_diary) & is.na(modified_diary), "skipped_diary_entry, ", "")
                                       )
  )

## mutate: changed 15,321 values (20%) of 'reasons_for_exclusion_diary' (0 new NA)

usable_diary_days <- diary %>% group_by(session) %>% 
  summarise(usable_diary_days = any(reasons_for_exclusion_diary == ""))

## group_by: one grouping variable (session)

## summarise: now 1,373 rows and 2 columns, ungrouped

Who completed what?

We create a character variable reasons_for_exclusion. We will concatenate (abbreviated) reasons for exclusion in this variable.

all_surveys$reasons_for_exclusion <- ""

Did not finish demographics survey

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(is.na(ended_demo), "didnt_finish_demographics, ", "")
                                       )
  )

## mutate: changed 53 values (3%) of 'reasons_for_exclusion' (0 new NA)

Reported no regular menstruation

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(menstruation_regular == 0, "no_regular_menstruation, ", "", "")
                                       )
  )

## mutate: changed 302 values (18%) of 'reasons_for_exclusion' (0 new NA)

Menopausal or in climacteric period

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(menopause_yes == 1 | menopause_yes == 2, "menopausal, ", "", "")
                                       )
  )

## mutate: changed 41 values (2%) of 'reasons_for_exclusion' (0 new NA)

Older than 50

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(age >= 50, "older_than_50, ", "", "")
                                       )
  )

## mutate: changed 35 values (2%) of 'reasons_for_exclusion' (0 new NA)

Pregnant

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(pregnant == 1, "pregnant, ", "", "")
                                       )
  )

## mutate: changed 23 values (1%) of 'reasons_for_exclusion' (0 new NA)

Breast-feeding

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(breast_feeding == 1, "breast_feeding, ", "", "")
                                       )
  )

## mutate: changed 28 values (2%) of 'reasons_for_exclusion' (0 new NA)

Not primarily heterosexual. This excludes women who reported being equally interested in men and women, women who reported being asexual or aromantic, and participants who did not identify as female gender.

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(sex_orientation >= 4 | gender != 1, "not_heterosexual_female, ", "", "")
                                       )
  )

## mutate: changed 26 values (2%) of 'reasons_for_exclusion' (0 new NA)

Did not finish personality survey

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(!is.na(ended_demo) & is.na(ended_initial), "didnt_finish_personality, ", "")
                                       )
  )

## mutate: changed 196 values (12%) of 'reasons_for_exclusion' (0 new NA)

Changed contraception

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(change_to_nonhormonal != 0 | change_to_hormonal_contraception != 0, "switched_contraception, ", "", "")
                                       )
  )

## mutate: changed 18 values (1%) of 'reasons_for_exclusion' (0 new NA)

Taking sex hormones (other than the pill)

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(medication_name %contains% "Cycloprognova" |
                                                      medication_name %contains% "Cyproderm" |
                                                       medication_name %contains% "DHEA" |
                                                       medication_name %contains% "Hormone" |
                                                       medication_name %contains% "Cyclo-Progynova" |
                                                       medication_name %contains% "Femoston" |
                                                       medication_name %contains% "Gynokadin", "sex_hormones, ", "", "")
                                       )
  )

## mutate: changed 7 values (<1%) of 'reasons_for_exclusion' (0 new NA)

No diary days

all_surveys <- all_surveys %>% 
  left_join(diary %>% group_by(session) %>% summarise(diary_days = n_nonmissing(ended_diary)), by = 'session') %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(!is.na(ended_initial) & diary_days == 0, "didnt_do_diary, ", "", "didnt_do_diary, ")
                                       )
  )

## group_by: one grouping variable (session)

## summarise: now 1,373 rows and 2 columns, ungrouped

## left_join: added one column (diary_days)

##            > rows only in x     287

##            > rows only in y  (    0)

##            > matched rows     1,373

##            >                 =======

##            > rows total       1,660

## mutate: changed 66 values (4%) of 'reasons_for_exclusion' (0 new NA)

Fertility never estimable

all_surveys <- all_surveys %>% 
  left_join(diary %>% group_by(session) %>% summarise(fertility_days = n_nonmissing(fertile_fab)), by = 'session') %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(menstruation_regular != 0  & diary_days > 0 & fertility_days == 0, 
                                                    "fertility_never_estimable, ", "", "")
                                       )
  )

## group_by: one grouping variable (session)

## summarise: now 1,373 rows and 2 columns, ungrouped

## left_join: added one column (fertility_days)

##            > rows only in x     287

##            > rows only in y  (    0)

##            > matched rows     1,373

##            >                 =======

##            > rows total       1,660

## mutate: changed 143 values (9%) of 'reasons_for_exclusion' (0 new NA)

all_surveys <- all_surveys %>% left_join(usable_diary_days, by = "session") %>% 
    mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
            if_else(menstruation_regular != 0  & diary_days > 0 & fertility_days > 0 & !usable_diary_days, "diary_days_not_usable, ", "", "")
                                       )
  )

## left_join: added one column (usable_diary_days)

##            > rows only in x     287

##            > rows only in y  (    0)

##            > matched rows     1,373

##            >                 =======

##            > rows total       1,660

## mutate: changed 45 values (3%) of 'reasons_for_exclusion' (0 new NA)

Women who are in monogamous heterosexual relationships

all_surveys <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion,
                                            if_else(hetero_relationship == 0 & relationship_status != 1, "non_heterosexual_relationship, ", "", "")
                                       )
  )

## mutate: changed 9 values (1%) of 'reasons_for_exclusion' (0 new NA)

testthat::expect_equal(all_surveys %>% group_by(session) %>% filter(n() > 1) %>% nrow(),
                       0)

## group_by: one grouping variable (session)

## filter (grouped): removed all rows (100%)

table(all_surveys$reasons_for_exclusion == "", exclude = NULL)

## 
## FALSE  TRUE 
##   788   872

diary <- diary %>% left_join(all_surveys %>% select(session, reasons_for_exclusion), by = 'session') %>% 
  mutate(reasons_for_exclusion = str_c(reasons_for_exclusion, reasons_for_exclusion_diary))

## select: dropped 368 variables (created_demo, modified_demo, ended_demo, expired_demo, info_study, …)

## left_join: added one column (reasons_for_exclusion)

##            > rows only in x        0

##            > rows only in y  (   287)

##            > matched rows     77,364

##            >                 ========

##            > rows total       77,364

## mutate: changed 27,531 values (36%) of 'reasons_for_exclusion' (0 new NA)

library(UpSetR)
exclusion_reasons <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_sub(reasons_for_exclusion, 1, -3)) %>% 
  select(session, reasons_for_exclusion) %>% 
  comma_separated_to_columns(reasons_for_exclusion) %>% 
  select(-session)

## mutate: changed 788 values (47%) of 'reasons_for_exclusion' (0 new NA)

## select: dropped 368 variables (created_demo, modified_demo, ended_demo, expired_demo, info_study, …)

## mutate: changed 872 values (47%) of 'splitcol' (0 new NA)

## mutate: new variable 'value' with one unique value and 0% NA

## spread: reorganized (splitcol, value) into (breast_feeding, diary_days_not_usable, didnt_do_diary, didnt_finish_demographics, didnt_finish_personality, …) [was 1864x4, now 1660x17]

## select: dropped one variable (reasons_for_exclusion)

## select: dropped one variable (session)

exclusion_reasons_table <- exclusion_reasons %>% 
  summarise_all(sum) %>% sort() %>% 
  gather(reason, n) %>% 
  left_join(all_surveys %>% mutate(reason = str_sub(reasons_for_exclusion, 1, -3)) %>% group_by(reason) %>% summarise(unique = n())) %>% 
  mutate(unique = if_else(is.na(unique), 0L, unique))

## summarise_all: now one row and 15 columns, ungrouped

## gather: reorganized (sex_hormones, non_heterosexual_relationship, switched_contraception, pregnant, not_heterosexual_female, …) into (reason, n) [was 1x15, now 15x2]

## mutate: new variable 'reason' with 57 unique values and 0% NA

## group_by: one grouping variable (reason)

## summarise: now 57 rows and 2 columns, ungrouped

## Joining, by = "reason"

## left_join: added one column (unique)

##            > rows only in x    3

##            > rows only in y  (45)

##            > matched rows     12

##            >                 ====

##            > rows total       15

## mutate: changed 3 values (20%) of 'unique' (3 fewer NA)

exclusion_reasons_hc <- all_surveys %>% 
  mutate(reasons_for_exclusion = str_sub(reasons_for_exclusion, 1, -3)) %>% 
  select(session, hormonal_contraception, reasons_for_exclusion) %>% 
  mutate(hormonal_contraception = if_else(hormonal_contraception, 1, 0)) %>% 
  comma_separated_to_columns(reasons_for_exclusion) %>% 
  # filter(included == 0) %>%
  select(-session)

## mutate: changed 788 values (47%) of 'reasons_for_exclusion' (0 new NA)

## select: dropped 367 variables (created_demo, modified_demo, ended_demo, expired_demo, info_study, …)

## mutate: converted 'hormonal_contraception' from logical to double (0 new NA)

## mutate: changed 872 values (47%) of 'splitcol' (0 new NA)

## mutate: new variable 'value' with one unique value and 0% NA

## spread: reorganized (splitcol, value) into (breast_feeding, diary_days_not_usable, didnt_do_diary, didnt_finish_demographics, didnt_finish_personality, …) [was 1864x5, now 1660x18]

## select: dropped one variable (reasons_for_exclusion)

## select: dropped one variable (session)

exclusion_reasons_hc_table <- exclusion_reasons_hc %>% group_by(hormonal_contraception) %>% 
  summarise_all(sum) %>% gather(reason, n, -hormonal_contraception) %>% 
  spread(hormonal_contraception, n) %>% 
  arrange(`0`)

## group_by: one grouping variable (hormonal_contraception)

## summarise_all: now 2 rows and 16 columns, ungrouped

## gather: reorganized (breast_feeding, diary_days_not_usable, didnt_do_diary, didnt_finish_demographics, didnt_finish_personality, …) into (reason, n) [was 2x16, now 30x3]

## spread: reorganized (hormonal_contraception, n) into (0, 1) [was 30x3, now 15x3]

exclusion_reasons_diary <- diary %>% 
  mutate(reasons_for_exclusion = str_sub(reasons_for_exclusion, 1, -3)) %>% 
  select(session, created_date, reasons_for_exclusion) %>% 
  drop_na(session, created_date) %>% 
  comma_separated_to_columns(reasons_for_exclusion) %>% 
  select( -created_date)

## mutate: changed 39,110 values (51%) of 'reasons_for_exclusion' (0 new NA)

## select: dropped 788 variables (person, short, created_diary, modified_diary, ended_diary, …)

## drop_na: no rows removed

## mutate: changed 38,254 values (39%) of 'splitcol' (0 new NA)

## mutate: new variable 'value' with one unique value and 0% NA

## spread: reorganized (splitcol, value) into (breast_feeding, cycle_longer_than_40, cycle_shorter_than_20, diary_days_not_usable, did_not_finish_entry, …) [was 97954x5, now 77364x22]

## select: dropped one variable (reasons_for_exclusion)

## select: dropped one variable (created_date)

exclusion_reasons_diary_table <- exclusion_reasons_diary %>% 
  select(-session) %>% 
  summarise_all(sum) %>% 
  sort() %>% 
  gather(reason, n) %>% 
  left_join(diary %>% mutate(reason = str_sub(reasons_for_exclusion, 1, -3)) %>% group_by(reason) %>% summarise(unique = n())) %>% 
  mutate(unique = if_else(is.na(unique), 0L, unique)) %>% 
  left_join(exclusion_reasons_diary %>% 
  gather(reason, n, -session) %>% 
  filter(n > 0) %>% 
  distinct(session, reason, n) %>% 
  group_by(reason) %>%
  summarise(n_women = sum(n)))

## select: dropped one variable (session)

## summarise_all: now one row and 19 columns, ungrouped

## gather: reorganized (dishonest_answer, didnt_do_diary, non_heterosexual_relationship, sex_hormones, did_not_finish_entry, …) into (reason, n) [was 1x19, now 19x2]

## mutate: new variable 'reason' with 181 unique values and 0% NA

## group_by: one grouping variable (reason)

## summarise: now 181 rows and 2 columns, ungrouped

## Joining, by = "reason"

## left_join: added one column (unique)

##            > rows only in x     6

##            > rows only in y  (168)

##            > matched rows      13

##            >                 =====

##            > rows total        19

## mutate: changed 6 values (32%) of 'unique' (6 fewer NA)

## gather: reorganized (breast_feeding, cycle_longer_than_40, cycle_shorter_than_20, diary_days_not_usable, did_not_finish_entry, …) into (reason, n) [was 77364x20, now 1469916x3]

## filter: removed 1,371,962 rows (93%), 97,954 rows remaining

## distinct: removed 93,686 rows (96%), 4,268 rows remaining

## group_by: one grouping variable (reason)

## summarise: now 19 rows and 2 columns, ungrouped

## Joining, by = "reason"

## left_join: added one column (n_women)

##            > rows only in x    0

##            > rows only in y  ( 0)

##            > matched rows     19

##            >                 ====

##            > rows total       19

Reasons for exclusion

exclusion_reasons %>% 
  filter(included == 0) %>% 
  select(-included) %>% 
  as.data.frame() %>% 
  {
  upset(., ncol(.), 20, show.numbers = TRUE, order.by = "freq",
      main.bar.color = "#6E8691",
      matrix.color = "#6E8691",
      sets.bar.color = "#53AC9B")
  }

## filter: removed 872 rows (53%), 788 rows remaining

## select: dropped one variable (included)

How to read this plot: The horizontal green bars show for how many women this reason for exclusion applies. The blue bars show how many women are excluded for multiple reasons (e.g., they’re menopausal and not heterosexual). If reasons for exclusion necessarily depend on another (i.e. participants had to finish the first survey to get to the second), we counted only those who had not yet been excluded earlier. If reasons for exclusion depended on each other only stochastically (e.g., age and menopause), we did not do this.

exclusion_reasons_table %>% 
  knitr::kable(caption = "Reasons for exclusion. _n_ shows the number of affected women, _unique_ those who for whom this was the only reason to be excluded.")

Reasons for exclusion. n shows the number of affected women, unique those who for whom this was the only reason to be excluded.

reason	n	unique
sex_hormones	7	4
non_heterosexual_relationship	9	4
switched_contraception	18	15
pregnant	23	2
not_heterosexual_female	26	10
breast_feeding	28	6
older_than_50	35	0
menopausal	41	0
diary_days_not_usable	45	43
didnt_finish_demographics	53	45
didnt_do_diary	66	51
fertility_never_estimable	143	134
didnt_finish_personality	196	141
no_regular_menstruation	302	178
included	872	0

By contraception

exclusion_reasons_hc %>% 
  as.data.frame() %>% 
  {
  upset(., ncol(.), 20, show.numbers = TRUE, order.by = "freq",
      main.bar.color = "#6E8691",
      matrix.color = "#6E8691",
      sets.bar.color = "#53AC9B")
  }

exclusion_reasons_hc_table %>% 
  knitr::kable(caption = "Reasons for exclusion. _n_ shows the number of affected women, _unique_ those who for whom this was the only reason to be excluded.")

Reasons for exclusion. n shows the number of affected women, unique those who for whom this was the only reason to be excluded.

reason	0	1
non_heterosexual_relationship	5	4
sex_hormones	6	1
switched_contraception	8	10
not_heterosexual_female	19	7
diary_days_not_usable	20	25
pregnant	23	0
breast_feeding	27	1
older_than_50	34	1
didnt_do_diary	38	28
menopausal	40	1
didnt_finish_demographics	47	6
didnt_finish_personality	66	130
fertility_never_estimable	67	76
no_regular_menstruation	162	140
included	580	292

In the diary

exclusion_reasons_diary %>% 
  filter(included == 0) %>%
  select(-session, -included, -didnt_do_diary) %>% 
  as.data.frame() %>% 
  {
  upset(., ncol(.), 20, show.numbers = TRUE, order.by = "freq",
      main.bar.color = "#6E8691",
      matrix.color = "#6E8691",
      sets.bar.color = "#53AC9B")
  }

## filter: removed 38,254 rows (49%), 39,110 rows remaining

## select: dropped 3 variables (session, didnt_do_diary, included)

exclusion_reasons_diary_table %>% 
  knitr::kable(caption = "Reasons for exclusion. _n_ shows the number of affected women, _unique_ those who for whom this was the only reason to be excluded.")

Reasons for exclusion. n shows the number of affected women, unique those who for whom this was the only reason to be excluded.

reason	n	unique	n_women
dishonest_answer	150	120	86
didnt_do_diary	187	0	28
non_heterosexual_relationship	338	135	8
sex_hormones	442	181	7
did_not_finish_entry	856	392	542
not_heterosexual_female	896	362	16
pregnant	1047	89	19
breast_feeding	1137	267	22
switched_contraception	1178	540	18
older_than_50	1560	0	28
menopausal	2166	0	39
diary_days_not_usable	2361	0	45
fertility_never_estimable	2527	0	143
next_menstrual_onset_unobserved	4100	2065	207
cycle_shorter_than_20	4736	3162	476
cycle_longer_than_40	7461	2704	222
no_regular_menstruation	13237	6893	245
skipped_diary_entry	15321	6823	1245
included	38254	0	872

included <- diary %>% filter(reasons_for_exclusion == "") %>% distinct(session)

## filter: removed 39,110 rows (51%), 38,254 rows remaining

## distinct: removed 37,382 rows (98%), 872 rows remaining

testthat::expect_equal(
  all_surveys %>% filter(reasons_for_exclusion == "") %>% nrow(), 
  nrow(included)
)

## filter: removed 788 rows (47%), 872 rows remaining

Sanity Checks

library(testthat)
expect_equal(nrow(diary), diary_length)
expect_equal(nrow(all_surveys), all_survey_length)
expect_equal(nrow(diary_social), diary_social_length)
expect_false(any(names(diary) %contains% ".x"))
expect_false(any(names(diary) %contains% ".y"))
expect_false(any(names(all_surveys) %contains% ".y"))
expect_equal(groups(s3_daily), list())
expect_equal(groups(diary), list())
expect_equal(groups(all_surveys), list())
expect_equal(sum(duplicated(all_surveys$session)), 0)
expect_equal(sum(duplicated(s1_demo$session)), 0)
expect_equal(diary %>% drop_na(session, day_number) %>% 
               group_by(short, day_number) %>% filter(n() > 1) %>% nrow(), 0)

## drop_na: no rows removed

## group_by: 2 grouping variables (short, day_number)

## filter (grouped): removed all rows (100%)

expect_equal(diary %>% drop_na(session, created_diary) %>%  
            group_by(session, created_diary) %>% filter(n()>1) %>% nrow(), 0)

## drop_na: removed 14,698 rows (19%), 62,666 rows remaining

## group_by: 2 grouping variables (session, created_diary)

## filter (grouped): removed all rows (100%)

expect_equal(s3_daily %>% drop_na(session, created_date) %>%  
            group_by(session, created_date) %>% filter(n()>1) %>% nrow(), 0)

## drop_na: no rows removed

## group_by: 2 grouping variables (session, created_date)

## filter (grouped): removed all rows (100%)

expect_equal(diary %>% drop_na(session, created_date) %>%  
            group_by(session, created_date) %>% filter(n()>1) %>% nrow(), 0)

## drop_na: no rows removed

## group_by: 2 grouping variables (session, created_date)

## filter (grouped): removed all rows (100%)

expect_equal(diary_social %>% drop_na(session, created_diary, person) %>%  
            group_by(session, created_diary, person) %>% filter(n() > 1) %>% nrow(), 0)

## drop_na: removed 47,466 rows (48%), 50,666 rows remaining

## group_by: 3 grouping variables (session, created_diary, person)

## filter (grouped): removed all rows (100%)

expect_equal(network %>% drop_na(session, person) %>%  
            group_by(session, person) %>% filter(n()>1) %>% nrow(), 0)

## drop_na: no rows removed

## group_by: 2 grouping variables (session, person)

## filter (grouped): removed all rows (100%)

Save

save(diary_social, diary, sex_long, network_nominations, network, s1_demo, s1_filter, s2_initial, s3_daily, s4_followup, s4_timespent, withfollowup, all_surveys, file = "data/cleaned_selected.rdata")