Skip to main content

Table 6 Unstandardized path coefficients for mediation model of affect towards e-cigarettes, risk perceptions, and e-cigarettes use (n = 5389)

From: Affect, risk perception, and the use of cigarettes and e-cigarettes: a population study of U.S. adults

Direct paths

Path coefficienta

95% CI

p

Affect → E-cigarette use status (ref = Current e-cigarette user)

 Never user (vs. Current e-cigarette user)

− 0.24

− 0.27 – − 0.21

< 0.001

 Former user (vs. Current e-cigarette user)

− 0.17

− 0.20 – − 0.14

< 0.001

Affect → Risk Perceptions

− 1.05

− 1.28 – − 0.83

< 0.001

Risk Perceptions → E-cigarette use status (ref = Current e-cigarette user)

 Never user (vs. Current e-cigarette user)

0.07

0.06 – 0.08

< 0.001

 Former user (vs. Current e-cigarette user)

0.03

0.02 – 0.04

< 0.001

Covariates → Affect

 Age (ref = 18–29 years old)

  30–44 years old

− 0.04

− 0.39 – 0.31

0.818

  45–59 years old

− 0.27

− 0.62 – 0.07

0.119

  60+ years old

0.14

− 0.19 – 0.48

0.401

 Race/Ethnicity (ref = White, non-Hispanic)

  Black, non-Hispanic

0.00

− 0.45 – 0.45

0.994

  Hispanic, any race

− 0.39

− 0.75 – − 0.02

0.036

  Other

− 0.03

− 0.49 – 0.43

0.899

 Education (ref = <High School)

  High School

− 0.18

− 0.62 – 0.25

0.408

  Some College

− 0.57

− 1.02 – − 0.13

0.012

  Bachelor’s degree or higher

− 0.84

− 1.28 – − 0.40

< 0.001

 Female

− 0.26

− 0.50 – − 0.02

0.035

Covariates → Risk Perceptions

 Age (ref = 18–29 years old)

  30–44 years old

1.98

0.71 – 3.25

0.002

  45–59 years old

3.39

2.09 – 4.70

< 0.001

  60+ years old

5.76

4.36 – 7.16

< 0.001

 Race/Ethnicity (ref = White, non-Hispanic)

  Black, non-Hispanic

0.86

− 0.91 – 2.63

0.339

  Hispanic, any race

− 0.24

− 1.62 – 1.14

0.736

  Other

1.13

− 0.57 – 2.82

0.193

 Education (ref = <High School)

  High School

− 0.03

− 1.66 – 1.60

0.969

  Some College

0.28

− 1.29 – 1.86

0.723

  Bachelor’s degree or higher

0.58

− 0.91 – 2.06

0.446

 Female

1.96

1.02 – 2.91

< 0.001

Covariates → Never user (ref = Current e-cigarette user)

 Age (ref = 18–29 years old)

  30–44 years old

0.33

0.07 – 0.58

0.011

  45–59 years old

0.90

0.68 – 1.12

< 0.001

  60+ years old

1.86

1.61 – 2.11

< 0.001

 Race/Ethnicity (ref = White, non-Hispanic)

  Black, non-Hispanic

0.12

− 0.21 – 0.44

0.476

  Hispanic, any race

− 0.48

− 0.71 – − 0.24

< 0.001

  Other

− 0.29

− 0.66 – 0.08

0.124

 Education (ref = <High School)

  High School

0.31

0.04 – 0.57

0.023

  Some College

0.48

0.21 – 0.75

< 0.001

  Bachelor’s degree or higher

1.15

0.84 – 1.47

< 0.001

 Female

− 0.05

− 0.23 – 0.13

0.589

Covariates Former user (ref = Current e-cigarette user)

 Age (ref = 18–29 years old)

  30–44 years old

0.14

− 0.14 – 0.42

0.338

  45–59 years old

0.20

− 0.04 – 0.45

0.107

  60+ years old

0.45

0.16 – 0.74

0.002

 Race/Ethnicity (ref = White, non-Hispanic)

  Black, non-Hispanic

0.26

− 0.07 – 0.59

0.119

  Hispanic, any race

− 0.18

− 0.46 – 0.10

0.200

  Other

− 0.29

− 0.72 – 0.14

0.182

 Education (ref = <High School)

  High School

− 0.01

− 0.30 – 0.28

0.966

  Some College

0.31

0.01 – 0.62

0.045

  Bachelor’s degree or higher

0.39

0.05 – 0.74

0.027

 Female

0.08

− 0.12 – 0.28

0.432

Factor Loadings

 Risk Perceptions

  Lung Cancer

0.99

0.99 – 0.99

< 0.001

  Lung Disease (e.g., COPD)

0.99

0.98 – 0.99

< 0.001

  Heart Disease

0.97

0.97 – 0.98

< 0.001

  Early/Premature Death

0.96

0.96 – 0.97

< 0.001

 Affect

  First Image Rating

0.87

0.85 – 0.88

< 0.001

  Second Image Rating

1.00

1.00 – 1.00

< 0.001

  1. Ref = referent group or category. denotes an estimated model path (for example, “Affect E-cigarette use status” signifies the path for the regression of e-cigarette use status on affect). aCoeffients are either multinomial logistic coefficients (unstandardized; akin to logistic regression coefficients) if e-cigarette use status is the outcome variable (i.e., “E-cigarette use status”, “Former user”, or “ Never user”), standardized factor loadings or residual correlations for the measurement model paths, or are linear coefficients (akin to regression coefficients) otherwise (i.e., “Affect” or “ Risk Perceptions”)