Lead partner : John Forth - National Institute for Economic and Social Research - United Kingdom
This activity will be dedicated to a discussion on the various possible survey methods available, on their advantages and drawback. Survey methods used in National systems are often a reflection of existing institutional constraints ; however, if alternative possibilities are available in a country, the chapter on survey methods in the guidelines should help evaluating them in terms of quality of the data, but also of its potential use in descriptions and analyses.
Detailed issues regarding survey methods include :
- The choice of target population
- Choice of respondents
- Sampling methods
- Matching methods
- Methods for panel constitution
- Collection methodology (face to face, computer assisted telephone interviews, etc)
- Weighting techniques (longitudinal, cross-sectional, employee, employer, and linked “weight”
- Quality assurance with discussion on variance calculations (using bootstrap weights) and possible errors such as sampling and non-sampling errors, coverage error, response and non-response errors, processing errors
- Discussion on optimal sample size, and response rates
- Discussion on issues such as stratum jumpers, edits, imputation and outlier detection
- Discussion on analytical techniques to identify how microdata can be used for analysis (which weights to us when creating estimates, how to calculate the correct standard errors to take in consideration the complex survey design, etc)
- A draft of the chapter on statistical methods will be discussed with members from the consortium and presented to the stakeholders at the time of the third Meadow General Assembly meeting to be held in February 2009.