The SAOM models estimated in RSIENA can include different specifications of influence and selection effects. The influence effects can be specified by "average similarity" (avSim), "total similarity" (totSim), or "average alter" (avAlt) [29]. Since they are calculated and interpreted in distinct ways, we conducted three different meta-analyses, separately for each specification of influence effect.

The selection effects can be specified either by similarity/identity effects (simX/sameX), or by the covariate-ego × alter effect (egoXaltX) that have to be analyzed separately. We conducted meta-analysis only for simX specification because the egoXaltX specification has been used very rarely (only in 4 articles).

Our data had 3-level nested structure because some articles provided more than one model (separate models for different schools or age cohorts), and several papers were based on the same project/database. To account for non-independence among effect sizes we used multilevel random-effects (MLRE) models with random effects at the coefficient level, the article level, and the database level. For model estimation we employed the metafor package in R (rma.mv) followed by estimating cluster-robust standard errors and confidence intervals [64].

Coefficients of SAOM models are presented as log odds ratios; hence, the mean effect sizes were calculated as logs odds ratios. To provide an effect size evaluation in metrics most often used in meta-analysis [65], we converted average log odds ratios into Cohen’s d.

To account for heterogeneity in effect sizes we pre-specified several moderators: number of networks, network size, number of possible nominations, between-waves time, country, gender composition and socio-economic status of the sample. We followed the practical recommendation to use at least ten studies for each covariate in the meta-regression [65].

Supplementary analyses were conducted on subsets of models (with one article per project/database).

Datasets, codes, and supplementary analyses results are available at https://osf.io/3avsn.

Note: The content above has been extracted from a research article, so it may not display correctly.



Q&A
Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.



We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.