Monday 17 December 2018

M8: Data Analysis and Population Size Estimations using RDS Analyst (RDS-A)

Description

Respondent-driven sampling (RDS) is being widely used as an effective method to sample populations that are connected through social networks.  Some of the populations most often sampled are migrants, members of political parties, young populations, men who have sex with men, sex workers, people who inject drugs, and others.

RDS is a chain-referral sampling used worldwide to recruit “hidden” populations that are less likely to be recruited from public venues. Since the first RDS study was conducted in 1994 in the United States, this method has become very popular throughout the world. However, many users of RDS struggle with understanding the theoretical parameters involved in analysing network data and selecting the appropriate estimators and software to analyse data collected through RDS methods. Recently a new analysis software, RDS Analyst (RDS-A), has become available. This program has a more user friendly graphical user interface which makes analysis easier and faster than before.  This software allows for the recoding and creation of variables, sample and population estimates, cross-tabs, means and other descriptive statistics as well as stratified analysis and regression. The software also has diagnostics for identifying bottlenecks, convergence, homophily and differential recruitment activity and for producing graphical plots. In addition, RDS-A provides all estimators used to analyse RDS data including RDS I, RDS II (with and without data smoothing) and the Successive Sampling estimator. Finally, the program has a newly developed method to impute network sizes and to estimate the size of hidden populations directly from the network sizes in RDS data.

Objectives

The course provides participants with knowledge and understanding of the theoretical and statistical concepts upon which RDS is based as well as step by step instruction on how to explore, diagnose problems in and analyse RDS data using RDS-A. The course addresses all steps in RDS data management and analysis, including the best practices for more advanced analyses. The training course will address new issues in RDS and parameters that need to be addressed when planning 2nd and any subsequent rounds of RDS. Importantly, the course addresses interpretation of RDS data, particularly the external validity.

Each participant will work on a computer with the following software:

  • RDS Analyst 2.0
  • Excel
  • NETDRAW (freeware for recruitment chains)
  • STATA or SPSS (for logistic regression)

Key topics of the course are: 

  • Latest information about RDS implementation and theoretical assumptions.
  • Implementing and diagnosing problems in RDS samples
  • RDS-A and other software that is currently available, their capacity and limitations
  • The meaning and importance of technical concepts such as homophily, social network sizes, differential activity, coarsened data and convergence
  • The interpretation of diagnostic plots and data from RDS-A
  • Preparation of data files for RDS-A (Excel, STATA, SPSS, Text, SAS, etc.)
  • Analysis of data using RDS -A
    • Adjusted and unadjusted estimates
    • Stratified analyses
    • Exploratory and descriptive statistics
    • Regression analyses
  • Development of plots using the plot builder
  • Graphical presentation of recruitment chains using NetDraw and RDS-A
  • Analysis of trend data and joining of multiple datasets
  • Export of individualized weights to conduct regression analyses
  • Regression analyses using either SPSS, STATA or R
  • Network size imputation
  • Population size estimation using RDS network size data

Teaching Methods

The course consists of lectures and individual work on computers so that participants can get hands-on experience in RDS data management and analysis. Participants are encouraged to bring their own RDS data sets to work on, particularly if they need to write up the results from RDS surveys for surveillance reports or scientific manuscripts.  Those without actual data will be provided with datasets. Participants should be familiar with basic statistics and Excel software. Participants are expected to bring their own laptop computers to the course. They will be given statistical programmes at the course to work with. 

Target Audience

This training invites principal investigators, statisticians, data managers and other researchers who have experience using RDS.

Course lecturers and facilitators

  • Lisa Johnston, PhD, Independent Consultant; University of California, San Francisco, Global Health Sciences; Assistant/adjunct professor, Tulane University, School of International Public Health and Tropical Medicine.  Website: www.lisagjohnston.com.
  • Katherine McLaughlin, PhD Candidate, Statistics Department, UCLA
  • Senad Handanagic, MD, MSc, WHO Collaborating Centre for HIV Surveillance, School of Medicine, University of Zagreb

Course organizer:

Lucija Sikic, University of Zagreb School of Medicine
training@snz.hr                                                                                                                                                                                                                                      

The course is held at:

WHO Collaborating Centre for HIV Surveillance
Andrija Stampar School of Public Health
University of Zagreb School of Medicine
Rockefellerova 4
10000 Zagreb
Phone: +385 1 4590 142 / 4590 100
Fax: +385 1 46 84 212

Download the Programme

Programme - Data Analysis and Population Size Estimations using RDS Analyst (RDS-A).docx

Created by hrvo.je.