Finally, within each age group, we completed a series of regression analyses for each outcome variable (see Section 2.2.) using the combined Colorado and NMMS data and including the following independent variables: state (Colorado vs. NMMS), year (2009-11 vs. 2006-8), exact age, gender, race-ethnicity, high school diploma (for adults only) and a state-by-year interaction. The interaction term was included to identify if time trends in Colorado and NMMS differed.
