From 4fa82a429d2bd2501e3cf6714bb39b72d8e76c0f Mon Sep 17 00:00:00 2001 From: Ben Anderson <dataknut@icloud.com> Date: Fri, 3 Jun 2016 19:18:36 +0100 Subject: [PATCH] Amended pooled 1983/87 test dataset name to avoid confusion later. --- MTUS-W6-adult-survey-data-processing.Rmd | 37 +++++++++++++++--------- 1 file changed, 23 insertions(+), 14 deletions(-) diff --git a/MTUS-W6-adult-survey-data-processing.Rmd b/MTUS-W6-adult-survey-data-processing.Rmd index 88f3126..8d30aa7 100644 --- a/MTUS-W6-adult-survey-data-processing.Rmd +++ b/MTUS-W6-adult-survey-data-processing.Rmd @@ -213,17 +213,17 @@ Before we do this we test for significant differences on core time-use dimension ```{r tableDifferences1983_1985} # Check distributions for 1st diary day -t_dt <- MTUSW6UKsurvey_DT[(survey == 1983 | survey == 1987) & diary == "1st diary day"] +subset1983_1987_DT <- MTUSW6UKsurvey_DT[(survey == 1983 | survey == 1987) & diary == "1st diary day"] kable(caption = "Months data collected", - table(t_dt$mtus_month, t_dt$survey) + table(subset1983_1987_DT$mtus_month, subset1983_1987_DT$survey) ) # essentially 1983 = autumn/winter, 1987 = spring/summer kable(caption = "Days data collected (day may be incorrect)", - table(t_dt$mtus_day, t_dt$survey) + table(subset1983_1987_DT$mtus_day, subset1983_1987_DT$survey) ) kable(caption = "Mean minutes per day by 1983/87 survey - 1st diary day", - t_dt[, + subset1983_1987_DT[, .( sleep = mean(main2), wash_dress = mean(main4), @@ -405,25 +405,34 @@ kable(caption = "Survey year & pooled survey year (ba_survey)", There do not appear to be large differences but we will test whether the survey year significantly predicts minutes per day in these activities given other characteristics (which may themselves have varied between the two samples). +The following analyses use a subset of the main data which only contains 1983 & 1987 data. + ```{r testDifferences1983_1985} -t_dt[, survey:= as.factor(survey)] +kable(caption="Cases for 1983/87 pooling testing", + table(subset1983_1987_DT$survey, + subset1983_1987_DT$ba_survey, + useNA = "always" + ) + ) + +subset1983_1987_DT[, survey:= as.factor(survey)] # Transformations based on spreadlevel plot of original un-transformed data -sleep <- lm((main2*main2) ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = t_dt) +sleep <- lm((main2*main2) ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = subset1983_1987_DT) -wash_dress <- lm(sqrt(main4) ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = t_dt) +wash_dress <- lm(sqrt(main4) ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = subset1983_1987_DT) -t_dt$eat <- t_dt$main5 + t_dt$main6 -eat <- lm(eat ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = t_dt) +subset1983_1987_DT$eat <- subset1983_1987_DT$main5 + subset1983_1987_DT$main6 +eat <- lm(eat ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = subset1983_1987_DT) -t_dt$paid_work <- t_dt$main7 + t_dt$main8 -paid_work <- lm(paid_work ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = t_dt) +subset1983_1987_DT$paid_work <- subset1983_1987_DT$main7 + subset1983_1987_DT$main8 +paid_work <- lm(paid_work ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = subset1983_1987_DT) -cook <- lm(main18 ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = t_dt) +cook <- lm(main18 ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = subset1983_1987_DT) -laundry <- lm(main21 ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = t_dt) +laundry <- lm(main21 ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = subset1983_1987_DT) -pub_etc <- lm(main39 ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = t_dt) +pub_etc <- lm(main39 ~ survey + mtus_month + ba_age_r + ba_nchild + hhtype, data = subset1983_1987_DT) ``` -- GitLab