fbpx

DATA INTEGRITY, WEIGHTING, AND ANALYSIS

Data integrity and proper application of statistical methods are essential to gaining a true understanding of your survey audience.

Data integrity and proper application of statistical methods are essential to gaining a true understanding of your survey audience. There are specific methods for cleaning, randomizing and matching that must be adhered to in order to ensure statistically significant results. Triton employs enterprise grade software tools, including Microsoft SQL Enterprise Server 2012 and IBM SPSS, along with rigorous data handling procedures.

Upon completion of calling, the raw survey data is weighted using industry standard statistical procedures to ensure the sample reflects the overall population, typically in terms of age, gender, ethnicity, political party affiliation, geography, etc. This processing step is essential because different segments of the population answer the phone in different ways. For example, women answer the phone more than men, older citizens are home more and participate more often than younger people, and rural residents typically answer the phone more frequently than urban residents. Without a proper weighting model, in most cases survey samples are heavily skewed one direction or another and are not representative of the target population.

While reporting can vary depending on customer requirements and budget, standard service usually includes full statistical analysis is provided in comprehensive crosstabs and graphical summary report. Turnaround time is generally five days or less. Multiple reports with different weightings or crosstab arrangements are available post survey for little or no additional cost. Triton will continue to work with you to help you understand the results and consider your next course of action.

Surveys that deliver timely, highly valuable results!

CALL TOLL-FREE (877) 284-7990

REQUEST PROPOSAL












    COMPANY PROFILE

    Explore the benefits of choosing Triton for your polling, survey, and/or research needs.