Skip to main content

Table 3 Characteristics of initial and final cluster analysis solutions, presented in order from the first model (all variables included) to the final model (only two variables included), LSAC data a

From: ‘Social screens’ and ‘the mainstream’: longitudinal competitors of non-organized physical activity in the transition from childhood to adolescence

Model characteristics

Model 1

Model 2

Model 3

Final model

Number of clusters/segments

2

3

2

2

Silhouette measure of cohesion and separation

0.1 (poor)

0.0 (poor)

0.5 (good)

0.8 (good)

Predictor importance of variables

 Texting/emailing/social media

1.00

1.00

1.00

1.00

 Other internet use

0.50

0.49

0.46

0.39

 Other physical activity

0.03

0.03

0.01

 Shopping

0.15

0.01

 Reading for leisure

0.13

0.01

 TV/movies

0.04

0.01

 Music for leisure

0.04

 Electronic gaming

0.04

 Homework/study

0.01

 School lessons

0.01

 Sleeping/napping

0.00

 Daily living activities

0.00

 Verbal communication

0.00

  1. a.Models were conducted iteratively until segments differed significantly in all input variables. Non-significant variables were excluded from subsequent models.