use https://www.stata-press.com/data/r18/nhanes2f, clear
svyset psuid [pweight=finalwgt], strata(stratid) singleunit(centered)
* slide 82
* 1. Get the means for the variables "iron" and "hgb"
svy: mean iron hbg
* 2. Make a weighted scatterplot with the variables "iron" and "hgb".
twoway (scatter iron hbg [pw = finalwgt]) (lfit iron hbg [pw = finalwgt])
* 3. Get a frequency table for the variable "female"; include both the weighted (population) and unweighted size.
svy: female, count obs ci cellwidth(12) format(%12.2g)
* 4. Is there a relationship between the variables "female" and "agegrp"?
svy: tab female agegrp // no; p-value = 0.2960
* slide 124
* 1. Run a t-test with the variable "heartatk" as the predictor and the variable "vitaminc" as the outcome
svy: regress vitaminc i.heartatk
* 2. Is the result statistically significant?
* no; the p-value is 0.1813
* 3. Run a regression model with "vitaminc" as the outcome and one continuous and one categorical predictor of your choice. Use two post-estimation commands of your choice.
svy: regress vitaminc i.agegrp bpsystol
contrast agegrp
margins agegrp, vce(unconitional)
* 4. How are the degrees of freedom for the model calculated?
* number of PSUs - number of strata
* 5. Rerun your regression using data from females only
svy, subpop(female): regress vitaminc i.agegrp bpsystol
contrast agegrp
margins agegrp, vce(unconditional) subpop(female)
* slide 143
* 1. Choose or create a binary variable from the example dataset
* I choose diabetes
* 2. Run a logistic regression with this new variable as the outcome variable and three predictors of your choice
svy: logit diabetes i.agegrp i.region weight
* 3. Include an interaction term in the model
svy: logit diabetes i.agegrp##c.weight i.region
contrast agegrp#c.weight
* 4. Graph the interaction
margins agegrp, at(weight=(35(10)175)) vce(unconditional)
marginsplot
* 5. Rerun your model using a subpopulation
svy, subpop(female): logit diabetes i.agegrp##c.weight i.region
contrast agegrp#c.weight
* 6. Get the predicted probabilities
margins agegrp, at(weight=(35(10)175)) vce(unconditional) subpop(female)
marginsplot