Skip to main content

Table 9 Latent variable structure, direct effects, and covariances of the LIWC-based SEM model for relapse

From: Analysis of associations between emotions and activities of drug users and their addiction recovery tendencies from social media posts using structural equation modeling

Relationships between variables

Estimate

Standardized estimate

Standard error

Z value

P value

Latent variables

 Anti-social—> 

Friend΄

1.000

0.597

 Anti-social—> 

We΄

0.569

0.339

0.147

3.8

0.000

 Anti-social—> 

Shehe΄

0.897

0.542

0.152

5.9

0.000

 Anti-social—> 

You΄

1.194

0.719

0.126

9.4

0.000

 Anti-social—> 

Male΄

1.015

0.604

0.129

7.8

0.000

 Anti-social—> 

Female΄

0.785

0.465

0.151

5.1

0.000

Regressions

 Anti-social—> 

Relapse

1.402

0.852

0.235

5.9

0.000

 Religion΄—> 

Relapse

0.151

0.156

0.063

2.3

0.016

 Motion΄—> 

Relapse

0.135

0.139

0.055

2.4

0.014

 Tone΄—> 

Relapse

− 0.044

− 0.045

0.077

− 0.573

− 0.573

Correlations

 Shehe΄

Female΄

0.467

0.652

0.057

8.1

0.000

 Friend΄

You΄

0.230

0.431

0.052

4.4

0.000

 You΄

Female΄

− 0.090

− 0.152

0.047

− 1.9

0.057

 Friend΄

Shehe΄

0.006

0.010

0.037

0.1

0.863

 Shehe΄

You΄

− 0.166

− 0.301

0.035

− 4.7

0.000

 Shehe΄

Male΄

0.395

0.614

0.053

7.4

0.000

 Friend΄

Male΄

0.210

0.338

0.048

4.3

0.000

 Anti-social΄

Tone΄

0.257

0.440

0.060

4.2

0.000

 Anti-social΄

Religion΄

0.193

0.330

0.054

3.5

0.000

 Motion΄

Tone΄

0.321

0.324

0.076

4.2

0.000

 Anti-social

Motion΄

0.043

0.073

0.052

0.8

0.409

 We΄

Tone΄

0.300

0.325

0.069

4.3

0.000

  1. The symbol ‘—> ’ is used to represent a path or direct effect in our SEM model. The negation of a variable is indicated by a prime. “Anti-social”, “motion΄”, and “religion΄” had a positive impact on relapse behavior