Page 162 - GHES-3-1

Global Health Economics and
Sustainability
Affect heuristics in substance use

Table 3. Results of the probit models estimated separately for the consumption of each substance by adolescents; Brazil 2019
Variables Alcohol Drugs Cigarettes
(1a) (2a) (1b) (2b) (1c) (2c)
Friend_alcohol 0.5997*** 0.5356*** 0.2423*** 0.1661*** 0.3262*** 0.2561***
Friend_drugs 0.4632*** 0.4055*** 1.0111*** 0.9728*** 0.6239*** 0.5950***
Friend_smoke 0.2558*** 0.2555*** 0.4399*** 0.4190*** 0.5421*** 0.5001***
Parent_alcohol 0.4163*** 0.3986*** 0.0789** 0.0597 0.0249 0.0551*
Parent_smoke 0.2484*** 0.2623*** 0.2993*** 0.3121*** 0.4035*** 0.3820***
SLI −0.8922*** −0.5978*** −0.2122**
RBI 1.0566*** 0.9413*** 0.8974***
Work 0.1229 −0.1304 0.0503
Study −0.0692 −0.2890*** −0.1859***
Study_work 0.1681*** −0.1329** 0.0050
Another_plan 0.0464 −0.1001 0.0007
Male −0.1563*** 0.0142 0.0744***
Age 0.1271*** 0.1373*** 0.1272***
School_mom 0.0121** 0.0324*** 0.0094
n_people −0.0469*** −0.0569*** −0.327***
White 0.0799*** 0.1104*** 0.0634*
Priv −0.1545*** −0.1001** −0.1633***
Urb 0.1369** 0.2339*** −0.1444**
Intercept 0.0178 −2.0803*** −1.6495*** −4.0252*** −1.1938*** −3.2829***
Respondents 58837 58837 58837
Wald Test 2222.36*** 17644.15*** 4472.77*** 11912.14*** 1495.42*** 9071.81***
Pseudo R 2 0.1284 0.1656 0.2102 0.2407 0.1522 0.1783
Source: Author’s elaboration based on NSHS 2019 data. Note: *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

cigarettes, and illicit drugs). Models (1a), (1b), and (1c) among adolescents remained when socioeconomic and
in Table 3 present only the variables of substance use by demographic variables and risk behaviors were controlled.
peers as explanatory variables. Models (2a), (2b), and (2c) Among models with the consumption of illicit
add to the other previously described control variables. All drugs among adolescents as their dependent variable,
models displayed in this table are statistically significant at all variables in the model (1b) that indicated the
the 1% level. The Wald test was utilized to determine the
global significance of the model. consumption of substances by peers exerted a positive
effect on the probability that an adolescent had already
Of the models with the consumption of alcoholic consumed illicit drugs. However, the consumption
beverages as their dependent variable, model (1a) was of alcoholic beverages by parents ceased to exercise a
estimated by including only the variables that represented statistically significant effect when other control variables
the consumption of substances by peers. In this model, were included in the model (2b). The coefficients of illicit
having a friend who consumes alcohol, having a parent who drug, alcohol, and cigarette consumption by friends and
consumes alcohol, having a parent who smokes, having cigarette consumption by a parent remained positive and
a friend who smokes, or having a friend who consumes statistically significant after the inclusion of the control
illicit drugs increased the probability that an adolescent had variables.
already consumed an alcoholic beverage. In model (2a), the
values of the coefficients of these variables did not show many Of models with cigarette consumption by adolescents
changes when the other control variables were included, as the dependent variable, the coefficient of the alcohol
and the statistical significance remained the same. Thus, the consumption by a parent was not statistically significant in
positive effect of peer consumption on alcohol consumption model (1c) but variables that indicated substance use by

Volume 3 Issue 1 (2025) 154 https://doi.org/10.36922/ghes.3829