Plots various graphs for panel data ---------------------------------------
xtgraph <variable>, t(varname) i(varname) nfull(#) miss subgroup(varname) subtype(string)
All the plots are longitudinal plots. The time variable and the individual vari > ables are REQUIRED in order to obtain any plot. These are specified in the t() and i( > ) options as used by many of the xt commands. Given this all that needs to specif > ied is the outcome variable of interest.
As these plots are often very cluttered, a shadow plot may be required. A shadow plot is exactly the same as the longitudinal plot except that all the lines are plotted in a light colour EXCEPT a few that are plotted in a dark col > our. This highlights a few lines and at the same time displays the complete data variance. The colours may need to be manipulated to obtain clearer highlighted lines since colours are machine dependent.
The miss option highlights the last data point (in time) per individual if th > e the individual has missing data.
For investigation of subgroups the subgroups, as specified by the subgroup() op > tion ( similar to a by option), can be plotted in different colours or using the val > ues of the variable specified by subgroup. The subgroup cannot be strings.
. xtgraph pef, t(day0) i(patient)
. xtgraph pef, t(day0) i(patient) nfull(3) subgroup(drugsort)
. xtgraph pef, t(day0) i(patient) nfull(3) subgroup(drugsort) subtype(s) . xtgraph pef, t(day0) i(patient) nfull(1) miss . xtgraph pef, t(day0) i(patient) nfull(0.4)
Also see --------
Adrian Mander MRC-Biostatistics Unit Cambridge, UK email: adrian.manderatmrc-bsu.cam.ac.uk