####Overview and Initial Diagnostics
This document will examine the final HIV in the Workplace dataset (with the exception of Lebanon). We will first take a look at how the data breaks along regional, income and country lines; as well as look at the impact of using projection weighting when analyzing the data across countries.
The total number of individuals represented by the ILO survey, across 49 countries is
print(total_proj_n)
## [1] 4025604983
We now consider the break-out of respondents by region and country income
print(region_list)
## # A tibble: 7 x 4
## region_unaids countries unweighted_n projection_n
## <fct> <int> <int> <dbl>
## 1 Eastern and southern Africa 9 9036 211216841.
## 2 Western and central Africa 11 11117 205427575.
## 3 Asia and the Pacific 9 13618 2719724281.
## 4 Eastern Europe and central Asia 7 8012 198313770.
## 5 Latin America and the Caribbean 6 6012 212840828.
## 6 Middle East and North Africa 6 6088 214500208.
## 7 Western/central Europe and Northern Ameri~ 1 1008 263581480.
print(wbi_list)
## # A tibble: 4 x 4
## wbi4 countries unweighted_n projection_n
## <fct> <int> <int> <dbl>
## 1 Low income 5 5035 115204180.
## 2 Lower-middle income 23 25236 1748906068.
## 3 Upper-middle income 16 19576 1848590809.
## 4 High income 5 5044 312903926.
One strategy for cross-country analysis is to apply projection weights. Given the countries in this set, we see that this approach will give greater representation to respondents from China and India, as seen in the below chart and table
rank_graph + theme_economist()+
theme(legend.position="none")
As the above graph shows, the countries of India represent over a majority of all respondents when apply projection weights on the pooled cross-country data.
####HIV 3 INDEX OVERALL SCORE.
This analysis considers how many items a respondent answered CORRECTLY in the HIV3 series items. They include:
HIV3. To the best of your knowledge, can HIV be transmirred in any of the following ways?
HIV3A Through unprotected sex (Yes/No) HIV3B Through hugging or shaking hands (Yes/No) HIV3C By sharing needles (Yes/No) HIV3D Through kissing (Yes/No) HIV3E By sharing a bathroom (Yes/No)
Index is calculated by scoring the number of items a person answered correctly and then is divided by the number of total questions answered. The resulting figure is a percentage between 0 and 100, with 100% indicating the person answered all of the questions correctly.
Note there are no results available for the UAE, as Gallup was not allowed to ask these questions in that country.
##Overall country level summary statistis
summary(HIV3_index_data$index_score)
## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
## 42.89 68.52 78.07 74.52 82.47 89.83 1
###Graph of country results
hiv3_graph<-ggplot(data=HIV3_index_data,aes(y=index_score, x=index_rank))
hiv3_scatter<-hiv3_graph + geom_text(aes(label=country_abv),
hjust=0, nudge_x=0.05, check_overlap=TRUE)+
labs(x="Index score rank", y="Index score", title="HIV 3 Index: Country Results")
hiv3_scatter + theme_economist()
## Warning: Removed 1 rows containing missing values (geom_text).
The median country resultstands at 78.07 – meaning about half of the countries in the study had an overall HIV3 Index score ABOVE this figure, while hald of the countries had a score below it
We see several Sub-Saharan African countries report the best overall index score, including South Africa (ZAF), Ethiopia and Namibia. The US finishes fourth, with a score of 86.6; Zimbabwe and Gabon boast scores over 85.
On the other end of the spectrum, Egypt has the lowest score at 42.89. Morrocco is the only other country with a score below 50, at 49.54.
The below table allows you to search for the country results.
library(DT)
## Warning: package 'DT' was built under R version 4.0.3
###Table of country results
datatable(HIV3_index_data, options=list(pageLength=50))
##wp21335/HIV1: Should people with HIV be allowed or not allowed TO work in a job with direct contact with those who do not have HIV?