-
Survey Design
Minimize total survey error with adept scientific survey design -
Panel Sampling
Secure samples from probability and non probability panels -
Weighting & Calibration
Construct representative results with our expert data scientists -
Analysis of
Survey DataSeasoned analytical insights and post-field data enhancements
Survey Sampling is Evolving
Time for a reality check: the old-school survey research paradigm was developed based on a number of assumptions that are becoming exceedingly difficult to fulfill. How do we adapt to the changed landscape without compromising the principles of representative sampling, diligent survey administration, effective data analysis, and coherent interpretation of results to produce actionable intelligence?
The new realities of the 21st Century poses a set of growing challenges that require thinking outside of the traditional survey sampling toolbox. While the fundamental principle stays the same – minimizing the so-called Total Survey Error – researchers are now in need of pragmatic alternatives that do not compromise quality and rigor.
All traditional methods are subject to change and evolution, a reality from which sample survey procedures are not exempt. While our existing inferential machinery has served us for decades when producing measurable inferences, the new realities induced by the digital evolution are forcing researchers to look for more pragmatic sampling options. It is from this perspective that traditional sampling methods that rely on single frames, such as RDD or ABS, are bowing out to alternatives that include sampling from multiple frames to improve coverage and reduce cost.
Hybrid sampling, which is an extension of the composite estimation methodology whereby data from different surveys are combined to produced more robust statistics, is rapidly becoming a method of choice for market and survey research applications. In particular, with the growing interest in online modality, it is becoming a common practice to supplement expensive survey data from traditional samples with those secured from online panels – an alternative that can be both cost-effective and improve coverage.
Carried out imprudently, however, naïve mixing of probability-based samples with those from online panels can lead to added cost and reduced inferential integrity of the resulting data. Samples from online panels need to be selected judiciously and the various pools of data need to be combined in an optimal fashion. Simply piling on whatever data panel providers can muster on top of your precious survey data can be akin to mixing fresh spring water with what is readily available from polluted sources just to increase volume. This is where our decades of experience with sampling and hands-on knowledge with sample blending can save your day.
As the leading provider of traditional sampling solutions that have served the industry for decades, Marketing Systems Group is now your one-stop shop for modern sampling alternatives that employ the latest online techniques.