Have a personal or library account? Click to login
The ‘GROW Social Network’ datasets Cover

The ‘GROW Social Network’ datasets

Open Access
|Feb 2020

Figures & Tables

Response rateThe most conservative ‘intent-to-treat’ response rates, based on the full sample denominator, regardless of retention or administration issues or face-to-face intervention dose:
Full network dataset
Baseline: n=469/610 (76.9%) (delayed admin)
Month 3: n=505 (82.8%)
Month 12: n=522 (85.6%)
Month 36: n=503 (82.5%)
Discussion and advice network dataset
Week 3: n=216/304 (71.1%)
Week 6: n=213/304 (70.1%)
Month 3: n=198/304 (65.1%)
[Trial retention at 3 years was 550/610 (90.2%) based on primary outcome, child BMI]
Non-respondent biasDifferential SN completion rates:
Full network dataset
Baseline:
Intervention n=251/304 (82.6%)
Control n=218/306 (71.2%)
Month 3:
Intervention n=270/304 (88.8%)
Control n=235/306 (76.8%)
Month 12:
Intervention n=263/304 (86.5%)
Control n=259/306 (84.6%)
Month 36:
Intervention n=258/304 (84.9%)
Control n=245/306 (80.1%)
[Trial differential dropout based on primary outcome (child BMI) was 91.4% (n=278/304) for the intervention group; 88.9% (n=272/306) for the control group]
Theoretical groupingThese data were collected as part of a community-based pediatric obesity prevention intervention. At baseline, 304/610 parents were randomized to the intervention group and assigned to one of 30 small groups of approximately 10 participants. The intervention was delivered in these small groups of the same participants who met weekly for 12 weeks
Publications using this dataDiscussion and advice network dataset
Gesell et al. (2016)
Gesell et al. (2020)
Data contextRandomized controlled trial
RespondentsLow-income parents with children 3 to 5 years of age, predominantly Hispanic
LongitudinalYes
  • 4 timepoints over 3 years

  • 3 timepoints over 3 months

TemporalityIn the Full Network Dataset the network is sparse, with a notable trend for participants in the intervention arm to form ties with other intervention participants (increased from 16.4% at baseline to 28.1 and 37.6% at 12 and 36 months, respectively), whereas participants in the control group formed ties at a notably slower rate and had comparatively fewer ties to participants in the intervention (increased from 10.1% at baseline to 12.0 and 17.1% at 12 and 36 months, respectively). (Paper under review)
In the Discussion and advice network dataset, 34% of participants did not seek advice from anyone, 22% sought advice from one person, and 44% sought advice from two or more people. Seven participants listed the maximum of seven possible advice nominations
Analytical or pedagogical utility
  • Analysis of social network and health behavior data collected at the same time points

  • Analysis of development of new social ties within the context of a group intervention, including comparison of intervention and control group, and comparison among 30 small intervention subgroups

  • Analysis of multiplex ties

Known issues
  • Rolling recruitment and data collection over 1.5 years affected the utility of the Full Network dataset. It limited potential bidirectionality (e.g., earlier recruits could not nominate later recruits at baseline), and the temporal proximity of outcome collection between nominator and nominee is not always guaranteed

DOI: https://doi.org/10.21307/connections-2019.017 | Journal eISSN: 2816-4245 | Journal ISSN: 0226-1766
Language: English
Page range: 123 - 128
Published on: Feb 14, 2020
Published by: International Network for Social Network Analysis (INSNA)
In partnership with: Paradigm Publishing Services
Publication frequency: 1 issue per year

© 2020 Sabina B. Gesell, Evan C. Sommer, Shari l. Barkin, published by International Network for Social Network Analysis (INSNA)
This work is licensed under the Creative Commons Attribution 4.0 License.