Minimizing both the number of bad respondents who take a survey and their impact on the survey results can seem as futile as Sisyphus pushing the rock up the mountain. Bad respondents come in all flavors: professional respondents, speeders, retakers (people who take the same survey multiple times), and outright frauds (people who aren’t who or what they claim to be).

Researchers have tried different approaches to these problems, including increasing sample size, eliminating one or two biggest types of bad respondents, or even ignoring the problem altogether. Unfortunately, the first two approaches can actually cause more damage than doing nothing at all. Let’s look at these three approaches more closely.

**Approach 1: Increase the sample size**

When concerned about accuracy, the common prescription among researchers is to increase the size of their sample. Indeed, this approach reduces sampling error, margin of error, and the impact of multicollinearity, and increases the confidence level in the results. However, larger sample sizes are a double-edged sword. Because a larger sample size reduces the standard error in the data, it also increases the t-value. As a result, a small difference between two or more respondent groups can greatly increase the chance of committing a Type I error (rejecting a true null hypothesis).

Similarly, if a sample has bad respondents, a larger sample size can actually exacerbate their impact on survey results. After all, bad respondents are likely to respond to survey questions differently than legitimate respondents. A larger sample size (even if every additional new respondent is good), will simply reduce the degree of these differences needed for statistical significance, and inflate the chance of drawing an erroneous conclusion from the survey findings.

**Approach 2: Tackle the biggest offender**

When faced with multiple problems, it is human nature to focus on eradicating the one or two worst problems. While that might work in most situations, eliminating only one type of bad respondents can actually cause more problems.

Assume that a survey’s results include responses from both professional respondents and speeders. Assume also that the survey has some ratings questions. What if – compared to legitimate respondents – the former rates an item higher than average, and the latter lower than average?

By having both types of bad respondents in the survey, their overall impact on the mean may be negligible. However, if you take out only one of them, the mean will become biased in favor of the type that was left alone, again exacerbating the impact of bad respondents.

**Approach 3: Do Nothing**

While doing nothing is preferable to the other two approaches, it has its own problems. Return to the example of two types of bad responders. While leaving both of them alone will keep the mean close to what it would be in the absence of both types, it will also inflate the variance of the data, resulting in an estimate of the mean that is untrustworthy. Hence, removing one type of bad respondents causes biased results while doing nothing causes inefficient results, neither of which has pleasant outcomes.

**What to do about bad respondents**

Bad respondents cannot be totally eliminated, but they can be minimized. The best ways to go about this include:

- Ask how your sample vendor screens people wishing to join its panel;
- Find out how your vendor ensures that panelists who are on other panels are precluded from being sent the same survey;
- Determine how your vendor tracks the survey-taking behavior of its panelists, assesses the legitimacy of each, and purges itself of suspected bad respondents; and
- Determine how your vendor prevents a person with multiple e-mail addresses – if you’re doing online surveys – from trying to register each one as a separate panelist.

Tags: bad respondents, hypothesis testing, professional respondents, sample bias, sample size, Sampling, statistical analysis, Statistics, survey research, Surveys, type I error

May 27, 2009 at 8:14 pm |

[…] Insight Central Insights for Maximum Marketing Performance « How Much Damage do Bad Respondents do to Survey Results? […]

December 11, 2009 at 8:22 pm |

I am always looking for brand-new informations in the net about this matter. Thanx!!