Data
====

1. earnings data [NO FILE INCLUDED BECAUSE OF NAs--NEED TO USE R TO ACCESS]
  - N      : number of observations
  - charity: income received from charity for entire family
  - educ_4 : respondent's education? 1: no hs, 2: hs, 3: some college, 4: college grad
  - immig  : 0 if US citizen, 1 if not
  - race   : race of respondent? 1: White, 2: Black, 3: Hispanic (non-black), 4: Other
  - rearn  : respondent's earnings
  - sex    : 1: Male, 2: Female
  - ssi    : ssi for entire family
  - tearn  : spouse's earnings
  - welfare: public assistance for entire family
  - workhrs: primary earner's hours/week worked last year
  - workmos: primary earner's months worked last year

Models
======

1. Single level model with multiple predictors
  earnings.stan    : lm(earnings ~ male + over65 + white + immig + educ_r + workmos 
				   + workhrs_top + any_ssi + any_welfare + any_charity)
  earnings2.stan   : lm(earnings ~ interest + male + over65 + white + immig + educ_r 
				   + workmos + workhrs_top + any_ssi + any_welfare 
				   + any_charity)
  earnings_pt1.stan: glm(earnings ~ male + over65 + white + immig + educ_r + any_ssi 
			            + any_welfare + any_charity, 
		         family=binomial(link="logit"))
  earnings_pt2.stan: lm(earnings ~ male + over65 + white + immig + educ_r + any_ssi 
			           + any_welfare + any_charity)