Statistical Grounding

Basically, it is a goal of this work to facilitate an ability for social network analysts to include a confidence level when they use social network measures calculated from collected data (observed network) and infered to the population (true network). The data we collect for analysis is propably missing data (for some reason) so what confidence can we have in the measures that are calculated from that data?

Test Statistic

Using the CASOS Lab tools that have been developed (some of which are free-available) we have the capability to study dozens of measures applicable to social networks.

Confidence Intervals

Resampling

Sampling Scheme:

Bootstrapping

Bootstrapping is a resampling procedure (Chernick, 1999, p. xi), so named because it resamples from the original data with replacement.
Boostrapping is a tool, like the jackknife and the delta method, that is used to estimate standard errors.

Variants of the bootstrap include: smoothed bootstrap, the double bookstrap, the Bayesian bootstrap and the parametric bootstrap (Chernick, 1999, p. 5).

Jackknife

REFERENCES