What we have seen is that once a poll is created, it’s usually put in a prominent place on our customer’s website. The reasoning is quite simple – the intention is to get as many pollers to participate. Even if this leads to higher polling participation, the results might be distorted.
There are many ways ‘Power Imbalances’ can result in skewed polling results :
1. Minimize Poll Creator Bias. If you were biased towards a certain polling option, you would put the polling option as the first option within the poll and usually seed the data accordingly. When the poll creator seeds a lot, power imbalances in your results will be more marked.
2. Do not make discriminatory references. Polling with questions directly or indirectly making references to special privileges, race, class, gender, sexual orientation, education, religion, health status and employment will encourage polls trending towards a power imbalance.
3. Provide the right context. Not providing the right context to your poll can lead in power imbalances. If your pollers came from a search engine with certain keywords, they are usually on your web page looking for something. If they happened to land on your poll and there was not enough context to the question you were asking, the search keywords could be the context they are using to complete your poll. So search optimization on the wrong keywords or not providing the right context to your poll could cause different polling results.
4. Monitor your network. Marketing savvy pollers can cause power imbalances. We have a feature where we let a poller share their poll with other people. If the people they share the poll with, also happens to be ‘thinking’ or ‘influenced’ by the person who shared the polls, then again a power imbalance has occured. This also applies to viral surveys.
About the author: Anup Surendran is the product manager for Micropoll. He has consulted for various technology startups in his career servicing Fortune 50 companies. He works out of Toronto, Canada and blogs frequently on his personal blog.