Survey Documentation and Analysis (SDA) Tools We Use: Survey Documentation and Analysis (SDA) Jon Stiles, UC DATA May 3, 2012
Part I. The Basics of SDA What is SDA? What data is available in SDA? What can you do with SDA? The parts of the SDA interface Menu Variable List Active variables Analysis Specification
Part II. Working with SDA 1. Finding data/variables/subjects - search - documentation 2. Analysis - Components - rows, columns, selection, controls Procedures - crosstabs, means, correlations Aids in Analysis Recoding Saving new variables Downloading
Part I. The Basics of SDA What is SDA? SDA (Survey Documentation and Analysis) is a set of programs for the documentation and Web-based analysis of survey data. It was developed and is maintained by the Computer-assisted Survey Methods Program (CSM) at UC Berkeley. It was developed as a companion program with CASES (Computer Assisted Survey Execution Program), a package for collecting survey data based on structured questionnaires, using a variety of modes of data collection. It operates on a transposed file structure, which makes analysis of datasets, especially large datasets, extremely fast.
Part I. The Basics of SDA What data is available in SDA? LOTS! Many popular social science datasets (e.g. the GSS, the ANES, the PUMS from the Decennial Census, the ACS, the CPS Annual Demographic Files,…… can be found in SDA format. Many archives (ICPSR, IPUMS, CPANDA, Roper, SDA, UCDATA, ISSR) provide access to at least some of their holdings via SDA. ICPSR: 778 studies out of 8283, 108 out of 457 studies added this year IPUMS: Decennial/ACS, CPS, NAPP
575 Studies at ICPSR (http://www.icpsr.umich.edu/access/sda.html)
A brief aside: search filtering ICPSR
http://www.icpsr.umich.edu/icpsrweb/NACDA/
http://www.icpsr.umich.edu/CPES/
http://www.icpsr.umich.edu/icpsrweb/IFSS/
http://www.icpsr.umich.edu/SAMHDA/
http://www.icpsr.umich.edu/RCMD/
Multiple Census Samples at IPUMS (http://usa.ipums.org/usa/sda/)
And CPS (March files) data, as well (http://cps.ipums.org/cps/sda/)
Orange County Polls (UCI) (http://data.lib.uci.edu/)
http://sda.berkeley.edu/archive.htm
Part I. The Basics of SDA What can you do with SDA? SDA can be used to: learn about a dataset (metadata, paradata) search for variables of interest investigate sample sizes and variable distributions perform statistical analyses transform, manipulate and create variables for each unit extract and download subsets or full datasets
Part I. The Basics of SDA The parts of the SDA interface Menu Variable List Active variables Analysis Specification
Basic Menu Analysis – choose Getting Started – note that you can resize /rearrange
Collapsed Variable Tree
Active Variables
Analysis Specification
Part II. Working with SDA 1. Finding data/variables/subjects - search - documentation The Codebook Menu
Part II. Working with SDA 1. Finding data/variables/subjects - search - documentation The Search Menu
Part II. Working with SDA Analysis - Components - rows, columns, selection, controls Procedures - crosstabs, means, correlations Screens will vary depending upon what procedure you are using. Start with exploratory – frequencies, cross-tabulations
Part II. Working with SDA Aids in Analysis Recoding Saving new variables Downloading
Recoding variables – on the fly age (5-18) age (r: 5-18) age (d: 5-18) age (c:13,5) Can be used in row, column, control (Crosstabs) Recoding variables – Web interface
Creating new variables – Web interface
Part II. Working with SDA Aids in Analysis Recoding Saving new variables Downloading
SDA Projects Under Development What Next? SDA Projects Under Development Complex Standard Errors The calculation of standard errors for percentages and means in complex samples has been revised to incorporate new developments for subpopulations and for differences. Complex standard errors for regression coefficients will also be implemented. Charts Charts were added to the MEANS program in version 3.4. Charts will also be added to the regression program in a future version. Data Documentation Initiative (DDI) Many organizations are working together to arrive at a common standard for documenting data files using XML. CSM is participating in this ongoing effort and has been involved in developing tools to read and write DDI metadata files.