Mktg, Inc., a market research company specializing in measuring sample source quality, released the results of a consistency audit it performed on DMS Research's River™ Sample.
The audit encompassed data collected in three waves from December of 2007 to May 2009.
According to Mktg, Inc. President Steven Gittelman PhD.
''DMS passed fourteen of fourteen measurable categories with a fifthteenth category left out due to a small sample size.'' According to Gittelman, ''The analysis determined that the DMS River™ product exhibits remarkable continuity over time.''
DMS Vice President of Marketing, Melanie Courtright, announced that
'DMS is committed to sampling excellence, which means sample that is representative and replicable. Variability in survey results should not be introduced by changes in the sample. We are committed to providing consistent sample every time.''
DMS announced today that they will be continuing with a regular audit as part of Mktg's Consistent Track™ program.
According to Courtright
'The Consistent Track™ program provides us with a combination of an outside audit and a management report that sets goals for the future. We find it to be a useful quality assessment tool that answers a critical client question: How will I know if any tracking differences are due to data or sampling changes? This process will ensure the sample itself is not causing shifts.''
Summary of DMS-River Consistency Results
The objective of this program is to capture the variability in online respondent sources. All survey data, by its inherent nature, is distributed. That is, in any survey, the responses are expected to vary. What we wish to measure here is the stability of those responses or the consistency of the source.
Analysis was based on the response to questions on three types of information:
Demographics (including age, income, education, and marital status distribution) reflect the traditional classification of respondents. Note that some of these variables were used to quota control the sample. As such, it measures the consistency of the quota process.
Structural Segments (based on buyer behavior, sociographic issues, and media use) reflect the cultural, social, and behavior characteristics of the respondents. While the distribution of these segmentation schemes may vary between countries and regions, they should be consistent within them and between instances of using the samples.
Source Performance (including erroneous, professional and satisficing behavior metrics) reflect the quality of survey results. These include the issues of incorrect responses, speeding through the survey, and participating in frequent surveys.
Two references were used to gauge the consistency of results: (1) overall average response (local mean), and (2) the external ''Grand Mean'' representing the average responses for the standard questionnaire from a number of sources.
The basis for evaluating consistency is the congruity of sample-set results with these references within the expected statistical error. To summarize the results of this study we examine the average deviation across the sample-sets against the overall average response, and the corresponding deviation from the most recent sample set.
Consistency of the DMS-River Data Source
In terms of overall consistency, 93.3% of the metrics (14 out 15) for the average deviation across DMS-River sample-sets were within the error bounds.
When considering only large inconsistencies (greater than a 50% discrepancy over the error bound) this becomes 100.0% with no metrics significantly outside the error bounds. The largest percent variant was Media Segments with a deviation of 25% of the error bound.
Mktg, Inc., through its Sample Source Audit Division, has pioneered new metrics to measure online sample quality.
Consistent Track™ is the latest, an audit system that measures the variability of panel output data over time. Buying behavior segmentations are compared to detect data shifts.
Other online quality control products of Mktg, Inc. include:
The Grand Mean Project™--- The Grand Mean is a new non-probabilistic metric that anchors online data to provide a reference for tracking and comparison purposes. So far over ninety panels have participated globally.
Optimum Blending Solutions™---Using the Grand Mean in each market as a new metric. Panel blending strategies minimize the RMS (Root Mean Square) error from the Grand Mean of the panel sources that are now available in the marketplace.
Consistent Track™---The measurement and certification of panel consistency through a unique audit program.
CPR™ (Consistency, Predictability, Reliability)---A unique tracking program that measures the consistency of sample sources to identify the real changes in data and reduce the background error often experienced in market research data.
QMetrics™---A unique method of identifying the worst online respondents.
Crop Duster™---De-duplication digital fingerprinting software.
East Islip, NY -June 2009