I got a dataset with 2 group of ppl (group1 is male, group 2 is female), and I got their odds ratio, and 95% CI. Represented as “ll, ul” in my dataset.
Now I tried to graph a two-way scatter graph, with their risk_score as x-axis and OR as y-axis, with 95%Ci. And I only need 2 legends in my graph, which are two dots of OR, but do not need legend of 95% CI (ll/ul)
i tried two ways, by constructing two datasets,
For dataset 1, I pooled all ppl and divided into 2 groups, and here is my code:
Code:
twoway ///
(rcap ll ul risk_score, vertical) ///
(scatter or risk_score  if group == 1, mcolor(red)) ///
(scatter or risk_score  if group == 2, mcolor(blue)) ///
xtitle ("Risk score") ytitle("ORs") legend(label1(2 "OR with male HH"） label(3"OR with female HH”)
The results did not return a graph, plus, it said "parentheses do not balance”.
So I tried another way by using dataset2. My code is:
Code:
twoway scatter or risk_score
twoway (scatter male_ors risk_score, yaxis(1)) || rcap male_ll male_ul risk_score /// 
       (scatter female_ors risk_score, yaxis(2)) || rcap female_ll female_ul risk_score ///
   xtitle ("Risk score") ytitle("ORs") legend(label1(2 "OR with male HH"） label(3"OR with female HH")
And Stata showed similar results as above.
May I know how to deal with this? Thank you so much!