distout and svydistout: help file to accompany Stata programmes for undertaking distributional analysis of continuous outcome variables

The distcout and svydistout programmes estimate and automatically save in the current directory two excel files which store a range of distributional statistics for any continuous variable (e.g. earnings, equivalised household disposable income) overall and for each of the groups specified by the variable include in the variable list (typical variables that can be included in the varlist include gender, age group, socio-economic status). Specifying the bygroup (groupvar) option performs the distributional analysis separately for each subgroup in groupvar. The distributional statistics produced by distout and svydistout are the mean, P10, P20, P30, P50, P70 and P90 percentiles of the continuous outcome variable. Both cross-sectional and over time change analysis of the distribution is undertaken. The cross-sectional analysis produces statistics for the mean and various percentiles of the distribution of the continuous variable varname overall and for each of the groups specified by the variables included in the varlist for each year in the dataset (or for the years specified by the user), as well as differences in the each distributional statistics between the different subgroups defined by each variable included in the varlist (along with the corresponding standard errors and p-values). The over-time analysis produces statistics for the change in the mean and the various percentiles of the distribution of the continuous variable overall and for each of the groups defined by the variables included in the varlist for each year in the dataset (or for the years specified by the user) relative to the previous year, as well as statistics for the differences in the change of each distributional statistics of subgroups relative to the reference group of each variable in the varlist. The programme also allows analysis to be undertaken by subpopulation groups allowing intersectional research. Both a non-survey design and survey design version of the programme are available (distout and svdistout respectively). The survey design version of the programme (svydistout) is undertaking survey correction only for the mean and the change in the mean estimates.