Graph showing how the suicide rate changes over time

[Yasmin Pewsey]

Hi everyone,

I'm currently completing my undergraduate dissertation, titled "The Macroeconomic Determinants of Mental Health". I'm using a country-level panel data set spanning 30 years to uncover how/if different elements of the macroeconomy affect the suicide rate and the prevalence of depression (my defendant variables). For my introduction, I would like to code a graph that shows how the suicide rate/prevalance of depression changes over time by income group. I've quickly drawn an example graph to show you the type of thing I am hoping to code:

I know the code for this is probably quite simple, but I've been playing around with the Graphics tab and still no luck

Is anyone able to point me in the right direction, or provide some sample code?

Thanks

[Maarten Buis]

It depends on what your data looks like. Can you give us an example? See the Statalist FAQ (link is in the black bar at the top) on how to do that.

[Yasmin Pewsey]

Hi Maarten, thanks for the response. Below is a sample of my data:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float CountryNum str48 Country int Year str2 EconomicRegion2019 double(DepressionPrev SuicideRate)
1 "Afghanistan"    1990 "L" .0408560809156679 10.3185041279549
1 "Afghanistan"    1991 "L" .0407920823075622 10.3270104500041
1 "Afghanistan"    1992 "L" .0407048655975331 10.2714111890416
1 "Afghanistan"    1993 "L" .0406604799497857 10.3761228683411
1 "Afghanistan"    1994 "L" .0406540039260417 10.5759145441461
1 "Afghanistan"    1995 "L"  .040652433609251 10.6823495644828
1 "Afghanistan"    1996 "L"  .040678033606743  10.720222847466
1 "Afghanistan"    1997 "L" .0407403547944048 10.7918222377962
1 "Afghanistan"    1998 "L" .0408268864157185 10.8624074437158
1 "Afghanistan"    1999 "L" .0409067316795364  10.970915166516
1 "Afghanistan"    2000 "L"   .04094524573839 11.0947432200641
1 "Afghanistan"    2001 "L" .0409156572981134  11.175178260235
1 "Afghanistan"    2002 "L" .0408385569503497 11.0544724130193
1 "Afghanistan"    2003 "L" .0407592711271612 10.9310926684929
1 "Afghanistan"    2004 "L" .0406903464897289 10.8397902521684
1 "Afghanistan"    2005 "L" .0406516571182214 10.6556263827148
1 "Afghanistan"    2006 "L" .0406444092991514 10.5384752685864
1 "Afghanistan"    2007 "L" .0406317409140969 10.3911287480941
1 "Afghanistan"    2008 "L" .0406195887185866 10.2191536282447
1 "Afghanistan"    2009 "L" .0406159595492444  10.036656612642
1 "Afghanistan"    2010 "L" .0406175837588872 9.87539258237426
1 "Afghanistan"    2011 "L" .0406322682291403 9.72186925459132
1 "Afghanistan"    2012 "L" .0406566596950581 9.57065057900413
1 "Afghanistan"    2013 "L" .0406855615178575 9.44766138671321
1 "Afghanistan"    2014 "L" .0407188560107943 9.34673059822699
1 "Afghanistan"    2015 "L" .0407468648129695  9.3191139906738
1 "Afghanistan"    2016 "L" .0407303005981798 9.25159986184833
1 "Afghanistan"    2017 "L" .0407133963201374 9.18856839991993
1 "Afghanistan"    2018 "L" .0407756849206637                .
1 "Afghanistan"    2019 "L" .0409066133095214                .
2 "Albania"        1990 "M" .0135389303240527 3.98516989836589
2 "Albania"        1991 "M" .0135502057140485 4.19900670506314
2 "Albania"        1992 "M" .0135777845116861 4.09420009962095
2 "Albania"        1993 "M" .0136186414709952 3.99379570351613
2 "Albania"        1994 "M" .0136682498678804 3.83212351391277
2 "Albania"        1995 "M" .0137239670371148 3.99658373439064
2 "Albania"        1996 "M" .0138052084962864 4.26962291647126
2 "Albania"        1997 "M" .0138913763134865 4.56311396281924
2 "Albania"        1998 "M" .0139944659595864 4.75940781317434
2 "Albania"        1999 "M" .0141579882997326 4.80816867262347
2 "Albania"        2000 "M" .0142945363954687 4.86701957066737
2 "Albania"        2001 "M" .0143817686998483 5.05108394895781
2 "Albania"        2002 "M" .0144704840713434 5.26610402317284
2 "Albania"        2003 "M" .0145496917720361 5.46462575107308
2 "Albania"        2004 "M"  .014607189344053 5.63102363876003
2 "Albania"        2005 "M" .0146309176458602  5.6163774990296
2 "Albania"        2006 "M" .0146058322370591 5.50065241930541
2 "Albania"        2007 "M" .0145421548308992 5.32997865056312
2 "Albania"        2008 "M" .0144648407622318 5.42147558752574
2 "Albania"        2009 "M"  .014396565240448 5.28660048510971
2 "Albania"        2010 "M" .0143578318442172 5.25775401253507
2 "Albania"        2011 "M" .0143473121041298 5.23957186339348
2 "Albania"        2012 "M" .0143479582908106 5.19961656388038
2 "Albania"        2013 "M" .0143579667396815 5.19681601742163
2 "Albania"        2014 "M" .0143704052657081 5.21657018291087
2 "Albania"        2015 "M" .0143755482159518 5.19108328893683
2 "Albania"        2016 "M" .0143741629542606 5.15557910659458
2 "Albania"        2017 "M" .0143870497149837 5.10878060847002
2 "Albania"        2018 "M" .0144507194229321                .
2 "Albania"        2019 "M" .0145666629082159                .
3 "Algeria"        1990 "M" .0337183948574448 4.83886233359688
3 "Algeria"        1991 "M" .0335498123756564 4.84281793097349
3 "Algeria"        1992 "M" .0333951355035503 4.86731620221792
3 "Algeria"        1993 "M" .0332605278855215  4.9140743366917
3 "Algeria"        1994 "M" .0331522638401438 5.05118197693958
3 "Algeria"        1995 "M" .0330766954683956 5.09269165288559
3 "Algeria"        1996 "M" .0330261782493018 5.14083132957216
3 "Algeria"        1997 "M" .0329815185829691 5.19511371887633
3 "Algeria"        1998 "M" .0329415748411108 5.19407668778325
3 "Algeria"        1999 "M" .0328994100389051 5.26244102486573
3 "Algeria"        2000 "M" .0328549583420267 5.26835867189677
3 "Algeria"        2001 "M" .0327982445593262 5.26595622642517
3 "Algeria"        2002 "M" .0327226528671152 5.24978922465377
3 "Algeria"        2003 "M" .0326365762650973 5.09878634995847
3 "Algeria"        2004 "M"  .032543715033233 5.03952855448661
3 "Algeria"        2005 "M" .0324525875369634 4.96704895260906
3 "Algeria"        2006 "M" .0323197100116501 4.88188892161778
3 "Algeria"        2007 "M" .0321304928701967 4.79079219172001
3 "Algeria"        2008 "M"  .031934481771384 4.69123762263069
3 "Algeria"        2009 "M" .0317789288347278 4.60304134538344
3 "Algeria"        2010 "M" .0317137677401988 4.50749334350865
3 "Algeria"        2011 "M" .0317476197172612 4.46207032704655
3 "Algeria"        2012 "M" .0318349755520022  4.3884416229494
3 "Algeria"        2013 "M" .0319426742870085 4.34927012045409
3 "Algeria"        2014 "M" .0320380707995079 4.32945657727614
3 "Algeria"        2015 "M" .0320894784225569   4.291621910652
3 "Algeria"        2016 "M" .0320177569054523 4.18781470021409
3 "Algeria"        2017 "M" .0319652678732525 4.12430646193392
3 "Algeria"        2018 "M" .0322075321464128                .
3 "Algeria"        2019 "M" .0327175811419446                .
4 "American Samoa" 1990 "M" .0139444095860065 6.84665005335045
4 "American Samoa" 1991 "M" .0138646941594683 6.81114819125232
4 "American Samoa" 1992 "M" .0137899571839374 6.76453751640861
4 "American Samoa" 1993 "M" .0137231295873954 6.74624558385127
4 "American Samoa" 1994 "M" .0136664473648662 6.74661508871285
4 "American Samoa" 1995 "M" .0136219903946288 6.71902462154722
4 "American Samoa" 1996 "M" .0135841042212095 6.68802941335749
4 "American Samoa" 1997 "M" .0135443252803234 6.71325790835888
4 "American Samoa" 1998 "M" .0135059668887098 6.64327331694032
4 "American Samoa" 1999 "M" .0134726649713064 6.53326996735081
end

The second to last column is the prevalence of depression, and the last column is the suicide rate.

I hope this helps

[Rich Goldstein]

what variable in your example is "income group"?

[Ken Chui]

Looking at the data posted in #3, I believe it also needs the country's population from whom the prevalence or rate was computed. It's because we cannot average the rates in all the "M" economic region; the rates have to be weighted by the population size.

[Yasmin Pewsey]

Hi Rich,
The 4th column is the income group. L = low income, M = medium income, H = high income. I don’t currently have the data on population, is there a way I can weight them equally?

Thanks

[Nick Cox]

A plain average is equivalent to equal weighting.

[Yasmin Pewsey]

Hi Nick,

Yes, an average is equivalent to equal weighting. Do you know how to code a graph showing how the average prevalence of depression/suicide rate is changing over time by income group, as mentioned in #1?

[Nick Cox]

I should think everyone who has answered knows this. The uncertainty is just exactly what you want. It appears to be something like

egen mean = mean(SuicideRate) , by(EconomicRegion2019 Year)

separate mean, by(EconomicRegion2019) veryshortlabel

line mean? Year ytitle(something sensible)

I would be queasy about using means here for the usual reasons, but you'll know what graph fits with the rest of what you've done.

[Yasmin Pewsey]

Thanks Nick,

When I used the code:

egen mean = mean(SuicideRate) , by(EconomicRegion2019 Year) 
separate mean, by(EconomicRegion2019) 
line mean Year

It generated this graph:

I know I didn't exactly copy your code. Stata displayed the error "option veryshortlabels not allowed" for the veryshortlabels command, and when I typed ytitle(XXX) it said "variable XXX not allowed". So I just took these two commands out of the code. That being said, this graph doesn't show what I hoped. I am wanting to display 3 lines, one for low-income, one for middle-income, and one for high-income countries, showing how the suicide rate has changed year on year. Do you know how I can code for this?

Thanks

[Nick Cox]

Yes. I just fixed one error in my previous. veryshortlabel is an undocumented option, except for https://www.stata-journal.com/sjpdf....iclenum=gr0023 and mentions here which you aren't expected to know about. I wrote separate before it was folded into official Stata. But veryshortlabels is not alllowed.

This is closer to what you want

separate mean, by(EconomicRegion2019) veryshortlabel  

line mean? Year, sort

What you did has spurious connections between the means for different groups. You should edit the variable labels. Also, they won't sort in the right order as they will sort H L M so a further guess is that what you really want could be more like

line mean1 mean3 mean2 Year, sort 

[Yasmin Pewsey]

Hi Nick,
It works! I used the code:

egen meanS = mean(SuicideRate) , by(EconomicRegion2019 Year)
separate meanS, by(EconomicRegion2019)
line meanS1 meanS2 meanS3 Year, sort

And it produced this wonderful graph:

I played around with the titles etc in the graph editor. Thank you so much, you've helped me a lot!

[Nick Cox]

Direct labelling is a better idea. This self-contained example shows the principle.

clear
set obs 5
gen year = 2000 + _n 
gen high = 12
gen medium = 10
gen low = 8
set scheme s1color 

line high medium low year, lc(orange_red blue black) xsc(r(2001 2005.5)) legend(off) || scatteri 12 2005 "high", mlabcolor(orange_red) ms(none) mlabsize(medsmall)  || scatteri 10 2005 "medium", mlabcolor(blue) ms(none) mlabsize(medsmall) legend(off) || scatteri 8 2005 "low", mlabcolor(black) ms(none) mlabsize(medsmall)

[Ken Chui]

I'd suggest renaming the title to "Mean country-level suicide rate (per 100,000, unweighted) over time" and y axis to "Mean country-level suicide rate (per 100,000, unweighted)". Let's say country A (100 pop) has 90% prevalence (90/100) and country (5000) has 20% prevalence (1000/5000). If we just average their rate it's 55%. But in reality the rate is only 1090/5100 = 21.4%. Calling it "mean rate" is very misleading. To avoid perpetuating any misunderstanding, it'd be better to state that up front.

[Yasmin Pewsey]

Hi Ken,
I will rename the title, thank you for the suggestion