Hi everyone,
I have the following issue. I'm currently working on the following project 8. Measuring the non-monetary cost of unemployment – Doing Economics. The thing is, I have to create a table with countries (3 of them) as the rows and the differences in the means of certain variables (namely, life satisfaction for three conditions of employment - full time, unemployed, retired), the standard errors of these differences and the confidence interval width for these differences. The code for doing this is given. It is the following
<
List chosen countries
country_list <- c("Turkey", "Spain", "Great Britain")
df.employment.se <- lifesat_data %>%
Select Wave 4
subset(S002EVS == "2008-2010") %>%
Select the employment types we are interested in
subset(X028 %in% employment_list) %>%
Select countries
subset(S003 %in% country_list) %>%
Group by country and employment type
group_by(S003, X028) %>%
Calculate the standard error of each group mean
summarize(se = sd(A170) / sqrt(n())) %>%
spread(X028, se) %>%
Calculate the SE of difference
mutate(D1_SE = sqrt(Full time
^2 + Unemployed^2),
D2_SE = sqrt(Full time
^2 + Retired^2))
df.employment <- df.employment %>%
Select chosen countries
subset(S003 %in% country_list) %>%
We only need the differences.
select(-Full time
, -Retired, -Unemployed) %>%
Join the means with the respective SEs
inner_join(., df.employment.se, by = "S003") %>%
select(-Full time
, -Retired, -Unemployed) %>%
Compute confidence interval width for both differences
mutate(CI_1 = 1.96 * D1_SE, CI_2 = 1.96 * D2_SE) %>%
print()
However, when running the second chunk of code I get the following error "Error: object 'D1_SE' not found". Now, what's weird is that the original code works. However, the error occurs when I attempt to change the 3 countries mentioned in the 1st chunk above. The code chunks related to the variable D1 are the following:
<
1)
Set the employment types that we want to report
employment_list = c("Full time", "Retired", "Unemployed")
df.employment <- lifesat_data %>%
Select Wave 4
subset(S002EVS == "2008-2010") %>%
Select only observations with these employment types
subset(X028 %in% employment_list) %>%
Group by country and then employment type
group_by(S003, X028) %>%
Calculate the mean by country/employment type group
summarize(mean = mean(A170)) %>%
Reshape to one row per country
spread(X028, mean) %>%
Create the difference in means
mutate(D1 = Full time
- Unemployed,
D2 = Full time
- Retired)
df.employment %>%
Combine with the average work ethic data
inner_join(., df.work_ethic, by = "S003") %>%
ggplot(., aes(y = D1, x = mean_work)) +
geom_point(stat = "identity") +
xlab("Work ethic") + ylab("Difference") +
ggtitle("Difference in wellbeing
between the full-time employed and the unemployed") +
theme_bw() +
Rotate the country names
theme(axis.text.x = element_text(
angle = 90, hjust = 1),
plot.title = element_text(hjust = 0.5))
Full-time vs unemployed
cor(df.employment$D1, df.work_ethic$mean_work)
and the two given above
Please, help me out with this. Thank you for your time.