A Simple Rule for the Data Quality Discussion

Our friends at Greenfield Online (Keith Price specifically) start to examine the issue of data quality in their latest newsletter. Two key things: The first is that it seems that Keith believes as the issues around cheating respondents has been a bit overstated (I agree). The second is that a bad survey will always deliver bad results. That's the issue that drives us at Invoke - creating a participant experience that delivers great insights.

A simple rule: Never write or test a survey that ends with you asking "who would take this?"

Here's a snippet of the article (click to read the whole thing):

Enter 2007. In the US, online surveys have reached high market penetration, and for most survey respondents, taking an online survey is no longer a "new" activity. As an industry we've gone from questioning under-participation of respondents, to questioning over-participation. We continue to look into respondents that attempt to cheat or game the system to participate in surveys. The "eager" respondent has become a concern. Data quality is a high priority for Greenfield Online. In fact, we've executed several studies to better understand the impact made by the "eager" respondent. Interestingly, we've found that even in some cases where we've encouraged "eager" respondents to over-state usage to qualify for surveys by offering sizable incentives, no more than 8% of an un-cleaned sample have "exaggerated" to qualify.

What do you think?


Popular posts from this blog

The difference between social posting and social media marketing

Wait, when did I become the "slow down" guy?