Include an Indicator for the 23 States That Passed MML by 2013;

***Incorporating medical marijuana law (MML) information*********;

data mml;

setmml;

***Include an indicator for the 23 states that passed MML by 2013;

ever_pass = 0;

if abbrev in ("AK""AZ""CA""CO""CT"

"DE""HI""IL""ME""MD"

"MA""MI""MN""MT""NV""NH""NJ"

"NM""NY""OR""RI""VT""WA") thenever_pass = 1;

*****law_status1 is a 0-1 indicator of whether the MML has been passed

yet in the particular state in the particular year;

law_status1 = ever_pass;

if abbrev = "MD" and year < 2003then law_status1 = 0;

if abbrev = "MT" and year < 2004then law_status1 = 0;

if abbrev = "VT" and year < 2004then law_status1 = 0;

if abbrev = "RI" and year < 2006then law_status1 = 0;

if abbrev = "NM" and year < 2007then law_status1 = 0;

if abbrev = "MI" and year < 2008then law_status1 = 0;

if abbrev = "NJ" and year < 2010then law_status1 = 0;

if abbrev = "AZ" and year < 2010then law_status1 = 0;

if abbrev = "DE" and year < 2011then law_status1 = 0;

if abbrev = "CT" and year < 2012then law_status1 = 0;

if abbrev = "MA" and year < 2012then law_status1 = 0;

if abbrev = "IL" and year < 2013then law_status1 = 0;

if abbrev = "NH" and year < 2013then law_status1 = 0;

if abbrev = "MN" and year < 2014then law_status1 = 0;

if abbrev = "NY" and year < 2014then law_status1 = 0;

******earlypass is an indicator of whether the state had already passed MML

before the first year of NSDUH observation 2004;

earlypass = 0;

if abbrev in ("CA","OR","WA","AK","ME","CO","NV","HI", "MD")

thenearlypass = 1;

****Here we create the MMLpass variables using the ever_pass, law_status1, and earlypass indicators;

MMLpass = "after";

ifever_pass = 1 and law_status1 = 0thenMMLpass = "before";

ifearlypass = 1thenMMLpass = "after";

ifever_pass = 0 thenMMLpass = "never";

run;

/** Spline for time trend ***/

data mml;

set mml;

yearcont = year - 2003;*calculate year count;

yearsp = max(yearcont-5, 0);*define the spline knot as 2008;

run;

/** Model: Impact of MMLs on Past month marijuana use by age and gender**/

procmixeddata = mml;

class abbrev age MMLpass gender;

model use = age gender MMLpass age*gender age*MMLpass gender*MMLpass

age*gender*MMLpass

/* historical time trend**/

yearcontyearcont*yearcontyearcont*yearcont*yearcontyearsp*yearsp*yearspage*yearcont age*yearcont*yearcont age*yearcont*yearcont*yearcont age*yearsp*yearsp*yearsp

/* state level covariates **/

Unemp_rateMedian_incoMALE_PCT WHITE_PCT PCT10_24 PCT_NoHS

/solutiondfm = bw;

random abbrev / group=age;

lsmeans age*gender*MMLpass/cl;

/* Code for causal contrasts **/

estimate"AGE 12-17, Men: After to Before"

mmlpass1 -10

age*mmlpass1 -10000000

gender*MMLpass10 -1000

age*gender*MMLpass10 -1000000000000000;

estimate"AGE 12-17, Women: After to Before"

mmlpass1 -10

age*mmlpass1 -10000000

gender*MMLpass010 -100

age*gender*MMLpass010 -1 00000000000000;

estimate"AGE 18-25, Men: After to Before"

mmlpass1 -10

age*mmlpass0001 -10000

gender*MMLpass10 -1000

age*gender*MMLpass000000 10 -1000000000;

estimate"AGE 18-25, Women: After to Before"

mmlpass1 -10

age*mmlpass0001 -10000

gender*MMLpass010 -100

age*gender*MMLpass000000010 -1 00000000;

estimate"AGE 26+, Men: After to Before"

mmlpass1 -10

age*mmlpass0000001 -10

gender*MMLpass10 -1000

age*gender*MMLpass000000000000 10 -1000 ;

estimate"AGE 26+, Women: After to Before"

mmlpass1 -10

age*mmlpass0000001 -10

gender*MMLpass010 -100

age*gender*MMLpass000000000000010 -1 00;

run;