Take Twitter polls with a pinch of salt

Twitter polls have become a popular way to engage online audiences on a variety of topics and test their interests. They are simple to create, run, and understand. It is not uncommon for political pundits and bloggers to use them to generate online and offline discourse about hot-topic issues.

In contrast to other surveys, Twitter polls are easy to conduct and have a higher chance of success than other methods. With a Twitter poll, you can ask a question and let the respondents choose between two and four responses.

By default, the surveys expire in 24 hours unless the “pollster” changes the duration, but not less than five minutes or longer than seven days.

The creator of polls cannot tell who voted and who didn’t, which gives the impression that anyone can vote as they wish. A voter has a single vote unless they have multiple accounts; in that case, each account can cast a single vote.

 With the upcoming general election rapidly approaching, political pundits — and basically anyone — have been using Twitter polls to gauge the popularity of parties, presidential candidates, and their running mates, generating animated arguments. Are we paying unwarranted attention to those polls? Do they accurately reflect the mood of the voters?

When testing the feelings of a population on an issue, a poll must give results representative of the majority and not just for the people who use Twitter. For example, to find out the fish distribution in a lake, you need to observe the lake in different places —not just in one convenient spot.

Accordingly, the science of opinion polls states that surveys must focus on finding respondents whose views, when combined, provide a good indication of what the whole population of voters thinks at a particular time.

Twitter polls are not scientific; they are not conducted systematically and cannot capture public opinion since respondents may not represent the majority. Ideally, they represent people with Twitter accounts and internet access, people who follow the “pollster” on Twitter, and those online when the poll is active. The survey results are only to be taken with a pinch of salt as they are laced with what research scientists call “sampling bias”.

The political commentators in Kenya, some of whom hold prestigious academic titles, take Twitter poll results for gospel truth, disregarding their weaknesses. Scientific or not, people naturally give more credibility to results that favour their opinion.

Twitter polls, if left unchecked, can mislead voters, leading to disgruntled citizens when the proper election results are announced. The media have a responsibility to make the public aware of the shortcomings of online self-polling.

 However, Twitter polls are not without value – they stimulate public discourse. In brand management, and done the right way, they can help stimulate engagement by gathering useful feedback from an audience.