Datawrangling

Data and Functions

source("0_helpers.R")
knitr::opts_chunk$set(warning = FALSE, message = FALSE)
pander::panderOptions("table.split.table", Inf)
pander::panderOptions('round',2)
pander::panderOptions('digits',2)
pander::panderOptions('keep.trailing.zeros',TRUE)

load("data/cleaned_selected.rdata")

Select Variables

Initial Survey

data = all_surveys %>%
  select(session,
         age,
         education_years,
         net_income,
         starts_with("bfi_agree"),
         starts_with("bfi_consc"),
         starts_with("bfi_open"),
         starts_with("bfi_extra"),
         starts_with("bfi_neuro"),
         religiosity,
         duration_relationship_years,
         duration_relationship_month,
         contraception_at_all,
         contraception_method,
         contraception_approach,
         contraception_method_broad,
         contraception_meeting_partner,
         partner_attractiveness_face,
         partner_attractiveness_body,
         relationship_satisfaction_overall,
         relationship_satisfaction_2,
         relationship_satisfaction_3,
         relationship_problems_R,
         relationship_conflict_R,
         satisfaction_sexual_intercourse,
         reasons_for_exclusion)

Diary

data_diary = diary %>%
  select(session,
         reasons_for_exclusion,
         high_libido, sexual_desire_for_whom,
         sex_active, sex_activity_anal_sex, sex_activity_bdsm_dom, sex_activity_bdsm_sub,
         sex_activity_cuddling, sex_activity_cunnilingus, sex_activity_cybersex,
         sex_activity_dirty_talk,
         sex_activity_fellatio, sex_activity_kissing,  sex_activity_masturbated_by_partner,
         sex_activity_masturbated_partner, sex_activity_masturbation, sex_activity_other,
         sex_activity_phone_skype_sex, sex_activity_pornography, sex_activity_sex, 
         sex_activity_touch_other, 
         sex_activity_toys, sex_activity_unclear,  sex_other, sex_solo, sex_unclear,
         days_done)

Exclusion

Initial Survey

n_excluded = data %>% filter(reasons_for_exclusion != "") %>% nrow()
data = data %>% filter(reasons_for_exclusion == "")

481 people were excluded.

Diary

Exclusion criteria based on initial

session_codes = data$session

data_diary = data_diary %>%
  filter(session %in% session_codes)

length(data_diary$session)

## [1] 53332

The 1179 participants filled out 53332.

Skipped diary entry

n_skipped_diary = data_diary %>%
  filter(reasons_for_exclusion %contains% "skipped_diary_entry") %>%
  count()

data_diary = data_diary %>%
  filter(!(reasons_for_exclusion %contains% "skipped_diary_entry"))

745 days were skipped.

Dishonest answers

n_dishonest = data_diary %>%
  filter(reasons_for_exclusion %contains% "dishonest") %>%
  count()

data_diary = data_diary %>%
  filter(!(reasons_for_exclusion %contains% "dishonest"))

142 days contained dishonest answers.

Filled out less than 14 days

number_of_days = data_diary %>%
  group_by(session) %>%
  summarise(n = n()) %>%
  rename(number_of_days = n)
  
data_diary = left_join(data_diary, number_of_days, by = "session")

data_diary_include = data_diary %>%
  filter(as.numeric(number_of_days) >= 14)

data_diary_exclude = data_diary %>%
  filter(as.numeric(number_of_days) < 14)

data_diary = data_diary_include

170 participants were excluded because they filled out less than 14 days - resulting in 1014 excluded days.

Summary

In total 968 participants with 51431 diary days were included for the analyses in which outcomes were based on diary information.

Wrangle data

Income

Set factor level correctly

data = data %>%
  mutate(net_income = factor(net_income,
                             levels = c("euro_lt_500", "euro_500_1000",
                                        "euro_1000_2000", "euro_2000_3000",
                                        "euro_gt_3000", "dont_tell")))

Relationship status (single vs partnered)

data = data %>% mutate(
  relationship_status = ifelse(is.na(duration_relationship_month), "Single", "Partnered"))

qplot(data$relationship_status) + coord_flip()

Relationship duration

data = data %>%
  mutate(relationship_duration = duration_relationship_years * 12 +
           duration_relationship_month)

qplot(data$relationship_duration)

data = data %>%
   mutate(relationship_duration_factor = 
            factor(ifelse(relationship_status == "Single",
                   "Single",
                   ifelse(relationship_duration <= 12,
                          "Partnered_upto12months",
                          ifelse(relationship_duration <= 28, 
                                 "Partnered_upto28months",
                                 ifelse(relationship_duration <= 52,
                                        "Partnered_upto52months",
                                        ifelse(relationship_duration > 52,
                                        "Partnered_morethan52months",
                                        NA))))),
                   levels = c("Single", "Partnered_upto12months",
                              "Partnered_upto28months", "Partnered_upto52months",
                              "Partnered_morethan52months")))

qplot(data$relationship_duration_factor)

ggplot(data, aes(relationship_duration)) +
  geom_histogram(aes(fill = relationship_duration_factor), bins = 100)

Contraception approach

data = data %>%
  mutate(# Participants, who indicated having no penetrative sex:
         contraception_method = if_else(contraception_at_all == 5,
                                   "barrier_no_penetrative_sex",
                                   contraception_method),
         # Participants, who indicated having no penetrative sex
         # use no contraceptives
         contraception_approach = factor(ifelse(
           contraception_method =="barrier_no_penetrative_sex",
           "nothing",
           as.character(contraception_approach))),
         # Fixed contraception_approach for
         # "hormonal_pill, barrier_coitus_interruptus"
         contraception_approach = factor(ifelse(
           contraception_method == "hormonal_pill, barrier_coitus_interruptus",
           "hormonal_pill_only",
           as.character(contraception_approach))),
         # Fixed contraception_approach for
         # "hormonal_pill, barrier_no_penetrative_sex"
         contraception_approach = factor(ifelse(
           contraception_method == "hormonal_pill, barrier_no_penetrative_sex",
           "hormonal_pill_only",
           as.character(contraception_approach))),
         # Fixed contraception_approach for
         # "hormonal_other, barrier_condoms"
         contraception_approach = factor(ifelse(
           contraception_method == "hormonal_other, barrier_condoms",
           "hormonal_other+condoms",
           as.character(contraception_approach))),
         # Fixed contraception_approach for
         # "hormonal_other, barrier_condoms, barrier_coitus_interruptus"
         contraception_approach = factor(ifelse(
           contraception_method ==
             "hormonal_other, barrier_condoms, barrier_coitus_interruptus",
           "hormonal_other+condoms",
           as.character(contraception_approach))),
         contraception_approach = factor(ifelse(
           contraception_approach == "hormonal+barrier",
           "hormonal_pill+condoms",
           as.character(contraception_approach))),
         contraception_approach = factor(ifelse(
           contraception_method %contains% "awareness" &
             contraception_method %contains% "condoms" &
             !(contraception_method %contains% "hormonal"),
           "awareness+condoms",
           as.character(contraception_approach))))


data = data %>%
  mutate(contraception_approach = factor(contraception_approach,
                                         levels = c("hormonal_pill_only",
                                                    "hormonal_other_only",
                                                    "hormonal_pill+condoms",
                                                    "hormonal_other+condoms",
                                                    "barrier_pessar",
                                                    "awareness",
                                                    "awareness+condoms",
                                                    "condoms",
                                                    "other",
                                                    "nothing")))

qplot(factor(data$contraception_approach, levels = rev(levels(data$contraception_approach)))) +
  coord_flip()

table(data$contraception_approach)

## 
##     hormonal_pill_only    hormonal_other_only  hormonal_pill+condoms hormonal_other+condoms 
##                    254                     61                    159                     17 
##         barrier_pessar              awareness      awareness+condoms                condoms 
##                     85                     20                    100                    380 
##                  other                nothing 
##                     14                     89

Current contraceptive status (hormonal vs. non hormonal)

data = data %>% mutate(
  contraception_hormonal = factor(ifelse(contraception_approach %contains% "hormonal",
                                         "yes",
                                         "no")))
qplot(data$contraception_hormonal) + coord_flip()

table(data$contraception_hormonal)

## 
##  no yes 
## 688 491

Changed contraception since meeting their partner

There will be NAs because we asked about contracetion when meeting partner only if participants were currently in a relationship

crosstabs(~contraception_hormonal + contraception_meeting_partner, data = data)

##                       contraception_meeting_partner
## contraception_hormonal   0   1 <NA>
##                    no  251 150  287
##                    yes 133 240  118

data = data %>% mutate(
  contraception_change_since_meeting_partner =
    ifelse(contraception_hormonal == "yes" & contraception_meeting_partner == 1,
           "congruent_hormonal",
           ifelse(contraception_hormonal == "no" & contraception_meeting_partner == 0,
                  "congruent_nonhormonal",
                  ifelse(contraception_hormonal == "yes" & contraception_meeting_partner == 0,
                         "switched_to_hormonal",
                         ifelse(contraception_hormonal == "no" &
                                  contraception_meeting_partner == 1, "switched_to_nonhormonal",
                                NA)))))

qplot(data$contraception_change_since_meeting_partner) + coord_flip()

data = data %>% 
  mutate(congruent_contraception = 
           factor(ifelse(contraception_change_since_meeting_partner %contains% "congruent",
                  1, 0)))

Contraception meeting partner

data = data %>%
  mutate(contraception_meeting_partner = factor(if_else(
    contraception_meeting_partner == 1,
    "yes", "no")))

Attractiveness Partner Scale

data = data %>%
  mutate(attractiveness_partner = as.numeric(
    (partner_attractiveness_face + partner_attractiveness_body)/2))

Relationship Satisfaction Scale

data = data %>%
  mutate(relationship_satisfaction = as.numeric(
           (relationship_satisfaction_overall +
              relationship_satisfaction_2 +
              relationship_satisfaction_3 +
              (6 - relationship_problems_R) +
              (6 - relationship_conflict_R))/5))

Diary information

calculate mean of libido, sex_active_frequency and actual sex frequency based on diary

data_diary_means = data_diary %>%
  group_by(session, number_of_days) %>%
  summarise(diary_libido_mean = mean(high_libido, na.rm = T),
            diary_sex_active_mean = mean(sex_active, na.rm = T),
            diary_sex_active_sex_mean = mean(sex_activity_sex, na.rm = T),
            diary_masturbation_mean = mean(sex_activity_masturbation, na.rm = T))

qplot(data_diary_means$diary_libido_mean)

qplot(data_diary_means$diary_sex_active_mean)

qplot(data_diary_means$diary_sex_active_sex_mean)

data_diary_means = data_diary_means %>%
  mutate(diary_sex_active_sex_sum = as.integer(round(diary_sex_active_sex_mean*number_of_days, 1)),
         diary_sex_active_sum = as.integer(round(diary_sex_active_mean*number_of_days, 1)),
         diary_masturbation_sum = as.integer(round(diary_masturbation_mean*number_of_days, 1)))

data = left_join(data, data_diary_means, by = "session")

Set sex_freq and mas_freq as missing, if participants indicated having no penetrative sexual intercourse and using no other form of contraception

data = data %>%
  mutate(diary_sex_active_sex_mean =
           ifelse(contraception_method == "barrier_no_penetrative_sex",
                  NA,
                  diary_sex_active_sex_mean),
         diary_sex_active_sex_sum =
           ifelse(contraception_method == "barrier_no_penetrative_sex",
                  NA,
                  diary_sex_active_sex_sum),
         diary_sex_active_mean =
           ifelse(contraception_method == "barrier_no_penetrative_sex",
                  NA,
                  diary_sex_active_mean),
         diary_sex_active_sum =
           ifelse(contraception_method == "barrier_no_penetrative_sex",
                  NA,
                  diary_sex_active_sum),
         diary_masturbation_mean =
           ifelse(contraception_method == "barrier_no_penetrative_sex",
                  NA,
                  diary_masturbation_mean),
         diary_masturbation_sum =
           ifelse(contraception_method == "barrier_no_penetrative_sex",
                  NA,
                  diary_masturbation_sum))

Select Data

data = data %>%
  select(-duration_relationship_years, -duration_relationship_month,
         -contraception_at_all, -contraception_approach, -contraception_method_broad)

Save Data

save(data, data, file = "data/cleaned_selected_wrangled.rdata")