Graph hazard rates values with 95% CI over a dependant variable

[Daniel Richards]

Hello All,

I would like use Stata to make a graph Hazard of rates values with a 95% over different values of a independent variable. But I am really stuck on how I can do this? I have been exploring the sts graph function but it seems to graph over time. I have attached an example of a similar graph that I would like to make. This example is taken from Strahilevitz. M., Odean, T, Barber, B, (2011) "Once Burned, Twice Shy: How Naive Learning, Counterfactuals, and Regret Affect the Repurchase of Stocks Previously Sold" Journal of Marketing Research, 48 pp. S102-S120, doi 10.1509/jmkr.48.SPL.S102,

In this graph, the hazard rate over different values of "Return Earned" are presented. In my data I have a variable "commission" and I want to graph hazard rate values over.
Any help would be greatly appreciated.

Regards,
Daniel Richards

[Clyde Schechter]

So, how have you calculated your hazard ratios? If you used -stcox- or -stcrreg- or one of the -streg- models, you can follow that up with -margins- and -marginsplot-. The 95% CI's won't be presented as curves around the average hazard ratio, but rather as error-bars above and below the points you select to plot. Moreover, you can customize the appearance of the plot you get by using almost any option available in -graph twoway-. Or, if you run -margins- with the -saving()- option, you can store the results in a data file and then use -graph twoway- to create a plot more like what you show in your post if that particular format is important.

[Daniel Richards]

Thanks Clyde,

I have been using the stcox command. What you said about storing the hazard ratios makes sense!! I just needed to come from the problem from a different angle. Much obliged!

Regards,
Daniel Richards