Overview of Internet Rx:
INTERNET RX Improves the validity and reliability of Internet research results by removing respondents who are harmful to the analysis due to their careless and cursory responding.
There has been mounting concern over problems experienced with Internet research. Issues with reliability (failure of online studies to replicate) and a variety of respondent abuses have been widely reported:
Tiny proportion of population accounting for large percentage of total online interviews
Respondents being members of multiple panels
No limits on # of interviews conducted per week, month, etc.
Extremely short interview lengths
The Challenge:
Internet research is by far the fastest growing form of consumer and B2B research.
The problems we are now experiencing will only worsen over time as:
· Demand grows for more and more respondents, and…
· More consumers learn about the “easy way” of earning $$
Industry-wide systems need to be adopted to help control the situation... for example: respondent validation and controls on past participation. But this will take lots of time and money and cooperation between research companies and their clients.
Until new safeguards can be implemented that reduce the problems to acceptable levels, we feel it is imperative that you adopt a method for “cleaning” out the harmful behaviors that are infecting your data.
The Major Offender:
THE “PROFESSIONAL” RESPONDENT
· Objective: Maximize his $ income from online interviews
· Engaged in one or more of the following harmful behaviors:
- “SPEEDING” through the interview in a careless manner
- “STRAIGHT-LINING” - answering with the same response most of the time
(e.g. repeated “top-box” response)
- “SWERVING” -- erratic, thoughtless, random responding
- “SANDBAGGING” - deliberate false or deceptive responding
RESULT OF THESE HARMFUL BEHAVIORS:
· MINIMALLY: “Noise” or unclear results making the analy sis more difficult and less discriminating
· WORST CASE: Invalid results leading to bad decision-making
The INTERNET RX Remedy:
INTERNET Rx removes respondents guilty of the harmful behaviors.
Through use of its proprietary algorithms, INTERNET Rx identifies and removes respondents who:
· Complete the interview in “impossibly” short lengths of time
· Exhibit patterned “straight-lining” responses (either on the evaluative measures within a questionnaire
or on the Screener)
· Engage in erratic/senseless/random responding
INTERNET Rx identifies between 15% - 25% of the Total Sample involved in one or more of these behaviors. When we remove these respondents and analyze the remaining data, the results are more “clear cut.”
“Noise” and respondents harmful to the analysis are minimized.
How INTERNET RX Works:
Conduct fieldwork with a sample size 25% larger than desired.
For example:
If Want… Conduct Study With…
N = 200 N = 250
N = 500 N = 625
N = 1,200 N = 1,500
When the study is completed, we identify the harmful behaviors and remove those respondents from the sample.
The final sample size is +/- 5% within the desired levels.
|
