Skip to main content
. 2017 Jan 16;19(1):e17. doi: 10.2196/jmir.5816

Table 5.

Variable coefficients (and their standard errors) found by three linear models accounting for different factors associated with percentage weight loss.

Variable Model 1 Model 2 Model 3

Coefficient (SE) P value Coefficient (SE) P value Coefficient (SE) P value
Real world






BMI 0.517 (0.090) <.001 0.415 (0.090) <.001 0.371 (0.044) <.001

Gender –1.357 (0.140) <.001 –1.979 (1.570) <.001 –1.471 (0.781) <.001

Age 0.235 (1.630) .88 0.126 (0.140) .36 0.118 (0.069) .09
Online behavior






Activity
–0.008 (0.005) .11 –0.006 (0.001) .01

Comments received
0.120 (0.007) <.001 0.005 (0.003) .12

Upvotes received
0.002 (0.001) .04 0.002 (0.000) <.001

Number of weights
0.270 (0.120) .02 0.288 (0.058) <.001

Weeks active
0.008 (0.070) .90 0.089 (0.033) <.001

Off days
–0.004 (0.002) .04 –0.004 (0.001) <.001
Post topics






Weight control apps

–57.26 (26.68) .03

Self-esteem (1)a

118.50 (24.49) <.001

Self-esteem (2)a

48.03 (27.47) .08

Workout

37.22 (20.05) .06

Friendship

83.72 (28.31) .003

Health information

44.84 (31.49) .15

Weight goals

77.90 (26.99) .004

Feelings

–125.00 (37.61) <.001

Asking for advice

68.97 (44.70) .12

Motivation

68.51 (52.50) .009

Weight check-ins

42.02 (26.25) .10

Family

–47.16 (30.94) .12
Comment topics






Body transformations

112.60 (59.71) .06

Asking for help

235.70 (96.420) .01

Body measures

–290.90 (108.20) .007

Changing lifestyle

–219.20 (141.40) .12

Counting calories

187.40 (71.55) <.001

Asking for support

–134.20 (63.51) .03

Friendship

223.40 (105.20) .03

Workout

295.30 (72.43) <.001

Motivation (1)a

304.80 (149.60) .04

Weight goals

243.00 (66.50) <.001

Motivation (2)a

251.80 (100.50) .01

Healthy foods

173.70 (91.41) .05

Gratitude

160.20 (72.06) .02

Starting again

–231.40 (106.10) .02

Lifestyle change

221.00 (99.84) .03

a Topics generated by latent Dirichlet allocation may overlap. The posts and comments for both self-esteem and motivation appear in two different topics, here indicated by (1) and (2) after their names.