Many national data sets used in educational research are not based on simple random sampling schemes, but instead are constructed using complex sampling designs characterized by multi-stage cluster sampling and over-sampling of some groups. Incorrect results are obtained from statistical analysis if adjustments are not made for the sampling design. This study demonstrates how the use of weights and design effects impact the results of contingency tables and chi-square analysis of data from complex sampling designs.
Walker, David A. and Young, Denise Y.
"Example Of The Impact Of Weights And Design Effects On Contingency Tables And Chi-Square Analysis,"
Journal of Modern Applied Statistical Methods:
2, Article 15.
Available at: http://digitalcommons.wayne.edu/jmasm/vol2/iss2/15