This application proposes to develop three key components in social network methodology: (a) the assessment of the sensitivity of network statistics and model parameters to various types of missing data under different sampling schemes, (b) develops the first notions of effect sizes and power analysis for social network methodology, and (c) cost-effective sampling schemes that maximize the "return on investment" when network researchers are faced with a finite resource pool. These developments are accomplished through a series of modern simulations while being coupled with advanced combinatorial data analytic approaches. All advances are made freely available in a user-friendly software package.
This application develops a foundation for assessing the quality of a sampling scheme for social network analysis to determine if study proposals will have adequate power (and/or sample size) to detect varying effects. Additional consideration is given to augment standard sampling schemes in such a manner as to identify the most informative areas of a network, insuring that funding mechanisms are more likely to exhibit a return on investment.
|Winograd, Rachel P; Steinley, Douglas L; Sher, Kenneth J (2014) Drunk personality: reports from drinkers and knowledgeable informants. Exp Clin Psychopharmacol 22:187-97|
|McCutcheon, Vivia V; Lessov-Schlaggar, Christina N; Steinley, Douglas et al. (2014) Social network drinking and family history contribute equally to first-onset alcohol dependence in high risk adults. Drug Alcohol Depend 141:145-8|