Data Sampling
CivicScience’s data reporting interface provides real-time results with tools for Census weighting by age and gender, cross-tabulation, margin of error calculations, time-trend analysis, and data export. Results can also be filtered by geography and time. Data mining algorithms identify groups of respondents who have answered specific combinations of questions, and cross-tabulations are tested using Chi-Squared and Tschuprow’s T to measure correlation strength.
To ensure reliable figures, data collection targets 3,000 responses per question per month for this report. The sample size is robust enough to allow for reliable extrapolation of trends across the broader workforce. Samples are weighted according to Bureau of Labor Statistics workforce demographics, considering factors such as age, gender, income, race, and education. This ensures the sample is representative and findings are generalizable. To ensure large enough sample sizes to achieve reliable cross-tabulation figures, CivicScience recommends a minimum single-question sample of 1,000 respondents for analysis.
CivicScience's large sample sizes require minimal reweighting, allowing for robust trend analysis and predictive modeling with high reliability.
220,000+
Responses in Quarterly Edition