Survey Design and Implementation

Our staff is well versed in multilevel sampling procedures (for area-based, RDD, or list-based sampling frames), creating complex sampling weights, and computing variances to address all levels of survey complexity.

Insight researchers select and apply the most appropriate data analytic methods, ranging from simple cross-tabulations to other more complex inferential techniques, including:

• Multivariate and univariate regression analysis
• Analysis of variance
• Cluster/factor analysis
• Hazard/survival analysis
• Optimization and classification
• Regression trees

We are experts in developing targeted data collection instruments (including psychometric and cognitive testing of questionnaires). We also have the expertise and experience to integrate quantitative and qualitative research—both primary and secondary—into higher-level, “triangulated” analysis and interpretations.

Key program evaluation services we offer include:

• Survey design plan. Insight's process for conducting surveys begins with developing a thorough sampling plan (details include defining the target population, compiling the sampling frame, calculating sample sizes, developing stratification plans, and designing methods for sampling, weighting, and variance estimation). This step also involves development of a survey methodology plan, which includes specifying the initial survey mode and follow-up strategies for achieving desired response rates, particularly for reaching hard-to-reach or underserved populations.
• Questionnaire design and pretesting. We use cognitive interviews and psychometric testing to validate every draft survey before finalizing the instrument. Additionally, we pretest the instrument in its intended mode for quality assurance purposes. For example, web surveys are subject to extensive usability tests, while computer-assisted telephone interview (CATI) surveys are tested to detect any programming problems.
• Frame development.
• Sample selection, weighting, and variance estimation.
• Data collection. Insight works closely with our clients to design a data collection approach that is optimally suited to each survey's target population, since different populations require different contact strategies and response modes.
• Editing and imputation.
• Multivariate modeling of complex survey data.
• Tabulation and report development and presentation.
• Public Use File (PUF) development.
• Nonresponse bias analysis.
• Incentive analysis.

^ Back to Top