Get Rid of FAKE RESPONDENTS Once and For All!!!


CATCH them and SAVE your survey!!

Do you really know who EXACTLY is filing up your online survey? The question probably haunts every researcher with no surety of the elusive solution. The very EXISTENCE of access panel industry depends upon how AUTHENTIC the respondents are in the panel base. One of the biggest reasons for traditional market researchers to be wary of online surveys is the cloak behind which the identity of the respondent is hidden. This article tries to find a practical look at the options available and the challenges associated with it.

Solution 1: LINK THE ID.

Participating in an online survey is a SERIOUS business as your opinion shapes up a brand therefore a certain level of commitment is required. One of the best solutions to identify a respondent is linking their IDs (for e.g. PAN CARD or PASSPORT NUMBER) to their profiles. Although it is very difficult to gather such information from respondents but it is a sure shot way to ensure that the respondent is one who he claims to be.

In India UID (Unique Identification Number) is being implemented and hopefully having such information will help correctly map the authenticity of the respondent.

The biggest challenge is the general psyche of the population not to share across personal identifiable information to someone which can however be countered by building a TRUST value for the panel base.

There are certain panel companies which are asking such information but most of the times these options are left blank. One should encourage respondents to fill ID information and can be assured for the security of such info.

Solution 2: CREATE TRAP SURVEYS.

Many panel companies claim to have trap questions in their recruitment surveys as well as encourage their clients to have some in the live questionnaire.

However there are some limitations to it – most of the trap questions are very basic in nature and any smart respondent can easily fool the system. Also there is a certain time limitation in terms of putting many trap questions in a live survey.

Therefore it is advisable to run an entirely DUMMY TRAP SURVEY. DTS (or dummy trap surveys) can be formulated specifically with the purpose of identifying and purging BAD RESPONDENTS.

The DTS can be run on a monthly/quarterly basis and can have lots of interlinked logic which should be difficult for a fake respondent to trick. Many panel companies might find it time consuming but in the longer run it is beneficial to their own business.

Solution 3: GIFT DELIVERY @ DOORSTEP.

One of the easiest and probably a very effective methodology which I had implemented in one of my previous companies is to deliver the gift at the physical location of the respondent and carry out a small feedback survey then and there itself.

Although it does have a limitation specifically for a panel company as most of them don’t have a field staff but still it is a good way to cross-check the authenticity of the respondent.

Solution 4: CREATE A FORUM.

Although a few panel companies have started it in the name of community building for the clients but hardly anyone is taking steps in terms of creating a forum/community specifically for the panel members.

It is a double edged sword as it may be counterproductive if there are too many disgruntled panel members – resulting in dreaded attrition.

These panel member specific forums will not only create a feeling of BELONGING to the panel company but also to gauze the commitment level as well as the sign of dubious claims.

It can also become a lateral source of revenue in the longer terms.

Solution 5: GENERIC BLACKLIST.

Although this idea will require lot of collaboration amongst different panel companies but nevertheless a good one – a generic blacklist of bad respondents can be shared across to cross-reference with one’s own panel and then purge them out of the system.

Although not a very effective method considering one can create infinite bogus email IDs but if technologically such an IP address or computer system can identified will help a lot in creating an effective blacklist.

Enhanced by Zemanta
Advertisements

About akshay kanyal
Understanding Survey Research

3 Responses to Get Rid of FAKE RESPONDENTS Once and For All!!!

  1. Some or all of this may help decide that the panelist exists and is who they say they are, Given the various checks that are put in, it is also likely that the respondent will answer questions honestly. However, even if the panelist is genuine and they answer honestly it doesn’t alter the problem that panelist are attitudinally unrepresentative and possibly effected by other research surveys they have done. On average, a panelist is a member of 3 or 4 panels and completes over 2 queationnaires a week for each.

    But lets forget even all of this and just step back and ask yourself what sort of people spend this sort of time for virtually no reward – who are they representative of.

    Access panels can be OK for difficult to find samples and provide relatively sound comparative data, but it is too unreliable for most research that we do.

    • akshay kanyal says:

      Thanks Mike for the insightful comments! You have raised a valid point that “panelist are attitudinally unrepresentative and possibly effected by other research surveys they have done” – survey research can still be considered at a very nascent stage of development. In coming years I can foresee a major focus towards EDUCATING PEOPLE about the constructive contribution they can make for a brand by their survey participation. It probably will help create more representation in the online medium.

  2. Page Schorer says:

    I agree with Mike that people who, for what ever reason, want to do a lot of survey’s are different from the “normal” people who do not want to do a lot of surveys.

    In the early days of on-line research, I was one of those people. I was very interested in what others were doing on-line, joined a couple of panels and did about two surveys a week. After a few months, i had a good idea of what people were doing (and also realized how crappy many of our surveys are) and stopped. Most panelists did not have my motivation but I think all of them have some motivation and it is not “eagerness to help brands”.

    On the other hand, every method I have used during a long career in survey research, has built in selection bias. The people who will stop for a mall interceptor are different from those who will not. The people who will do a door-to-door are different from those that refuse or are not at home. The people who will hang on the phone and answer questions are different from those who say, “No thanks” and hang up. Maybe, the census long form, which used government coercion, has no selection bias but even here I expect it does have at least some bias.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: