Skip to main content
. Author manuscript; available in PMC: 2024 Apr 1.
Published in final edited form as: Addiction. 2022 Dec 5;118(4):727–738. doi: 10.1111/add.16095

Table 5.

Effects of sociodemographic predictors on transitions across W1-W4.

Nonusers as reference a Class membership at time t+1
Class membership at time t ECIG SLT CIG
Beta SE Beta SE Beta SE
Age Nonusers −0.11 b 0.05 0.02 0.10 0.09 0.06
ECIG 0.39 c 0.17 0.07 0.28 0.17 0.17
SLT −0.42 0.31 −0.16 0.25 0.45 0.39
CIG −0.04 0.20 −0.62 0.23 0.08 0.18
Gender (ref=male) Nonusers −0.02 0.08 −1.22 0.26 0.37 0.11
ECIG 0.34 0.34 −10.60 0.43 −0.20 0.39
SLT 10.25 0.84 0.73 0.49 −0.06 0.84
CIG −0.24 0.50 0.08 1.28 −0.26 0.34
Parent education Nonusers 0.08 0.05 −0.09 0.10 −0.27 0.06
ECIG 0.13 0.19 0.42 0.36 −0.01 0.30
SLT 1.25 0.33 −0.06 0.33 0.34 0.54
CIG 0.29 0.26 0.81 0.21 0.23 0.23
Black (ref=White) Nonusers −0.25 0.16 −0.51 0.40 −0.15 0.19
ECIG −0.97 0.61 −10.04 0.58 −1.32 0.82
SLT 2.00 1.65 −1.42 0.80 −9.64 0.77
CIG −0.84 1.11 2.17 1.53 −0.05 0.54
Hispanic (ref=White) Nonusers 0.05 0.09 −0.63 0.28 0.11 0.12
ECIG −0.31 0.40 −1.00 1.05 0.75 0.51
SLT 2.13 1.68 0.49 0.76 1.01 0.96
CIG −1.28 0.68 −8.91 0.73 −0.78 0.36
Other (ref=White) Nonusers 0.00 0.11 −0.71 0.35 −0.08 0.15
ECIG −0.61 0.51 −0.46 1.09 −0.50 0.57
SLT −7.60 1.66 1.44 1.10 1.64 1.32
CIG 0.03 0.69 2.17 1.50 0.15 0.43
ECIG users as reference d Class membership at time t+1
Class membership at time t NON SLT CIG
B SE B SE B SE
Age ECIG −0.35 0.16 −0.14 0.32 0.01 0.19
Nonusers 0.44 0.17 0.28 0.34 0.20 0.21
SLT 0.88 0.35 0.54 0.41 1.07 0.43
CIG 0.50 0.25 −0.31 0.39 0.31 0.27
Gender (ref=male) ECIG −0.41 0.34 −12.11 0.40 −0.27 0.44
Nonusers 0.40 0.36 10.91 0.48 0.67 0.46
SLT −10.07 0.93 1.15 0.89 −9.90 1.09
CIG 0.67 0.59 11.23 1.36 0.64 0.68
Parent education ECIG −0.24 0.20 −0.20 0.40 −0.45 0.34
Nonusers 0.18 0.21 0.02 0.41 0.10 0.35
SLT −1.09 0.38 −1.28 0.45 −0.81 0.61
CIG −0.14 0.32 0.54 0.47 0.03 0.42
Black (ref=White) ECIG 1.03 0.61 −8.11 0.60 −0.17 0.90
Nonusers −0.79 0.63 7.85 0.70 0.24 0.94
SLT −2.78 1.81 4.44 1.71 −11.70 1.71
CIG 0.05 1.25 10.85 1.85 1.06 1.40
Hispanic (ref=White) ECIG 0.29 0.39 −1.19 1.05 −0.32 0.54
Nonusers −0.35 0.41 0.51 1.09 0.37 0.56
SLT −2.47 1.75 −1.13 1.93 −0.75 1.82
CIG 0.94 0.79 −8.28 1.39 0.87 0.86
Other (ref=White) ECIG 0.67 0.50 −0.42 1.11 0.08 0.66
Nonusers −0.68 0.52 −0.29 1.17 −0.15 0.68
SLT 6.14 1.74 7.96 1.83 8.29 1.69
CIG −0.70 0.86 1.85 1.88 −0.04 0.92

Betas (SEs) shaded in gray met the transition probability cutoff (>.10 across all waves), and thus are discussed in the text.

a

Betas for Nonusers at time t represent the effects of the predictor on transitioning to ECIG, SLT, or CIG at time t+1 (compared to remaining Nonusers) among those who were Nonusers at time t. Betas for ECIG, SLT, and CIG at time t represent the differences in the effects of the predictor on transitioning to ECIG, SLT, or CIG at time t+1 (compared to transitioning to Nonusers) among those who were ECIG, SLT, or CIG at time t.

b

As an example, this entry is interpreted as follows: older Nonusers (compared to younger Nonusers) were less likely to transition to ECIG use than to stay a nonuser; that is, for every one-unit increase in age, the log-odds of transitioning from Nonusers to ECIG (compared to remaining Nonusers) changed by −0.11.

c

As an example, this entry is interpreted as follows: compared to Nonusers, older ECIG users (compared to younger ECIG users) were more likely to continue ECIG use than to transition to nonuse; that is, for every one-unit increase in age, the difference in the log-odds of remaining ECIG (compared to transitioning from ECIG to Nonusers) changed by 0.39. This means that, for every one-unit increase in age, the log-odds of remaining ECIG (compared to transitioning from ECIG to Nonusers) changed by −0.11 + 0.39 = 0.28.

d

Betas for ECIG at time t represent the effects of the predictor on transitioning to Nonusers, SLT, or CIG at time t+1 (compared to remaining ECIG) among those who were ECIG at time t. Betas for Nonusers, SLT, and CIG at time t represent the differences in the effects of the predictor on transitioning to Nonusers, SLT, or CIG at time t+1 (compared to remaining ECIG) among those who were Nonusers, SLT, or CIG at time t.

Bolded values significant at p < .05.

Polytobacco use classes excluded from these analyses due to variable interpretations across waves and notably small sample sizes (<4% at most waves).