Tips on Asking Questions to Get More Accurate Data on GaG

One of the great features of this site is the polls, which is a very nice tool to easily get real statistics.

It's easy to make mistakes in statistics and data analysis, even if you're a mathematician. Many people use this to their advantage to get results they want to prove a point. But even trying your best to get accurate results it's very easy to fail. So I'd like to share a few tips that might help you when getting stats whether it's using polls on this site or doing surverys.

Think about who is answering.

It's always important to know your audience, but knowing who's answering is critical for accurate statistics.

Example:

If you want to know the percentage of people who use facebook, asking on this site won't get you accurate results, because people on this site probably use the internet more than the average person and the majority of this site's users are aged 18-25.

Another example:

Depending on what time of day you ask you'll get more answers from different countries due to different time zones (though on this site questions do usually keep geting answers for at least 24 hours after asking so it's usually not a problem.)

Think of the psychology of people answering.

This is similar to the last one, but it's very important. You'll understand with the example.

Example:

People are more likely to share their opinion if they think it's different. If you ask "Who has been scuba diving?" people who have been scuba diving are more likely to click it. You may get 50% of people voting yes when really hardly anyone has been.

Note: This can be helped by not making the question clear in the title so people have to click on the question to see what it's asking. You will however get a lot less answers by doing this, but you'll get more accurate stats. Also you can ask multiple questions at once, like "which of the following have you done?" with scuba diving being one of the choices, though then you can't really use the poll.

Try to get lots of answers.

Of course the bigger the sample size the more accurate the results, so try to make a title that will make people want to click it, but without having a clickbait title that will only draw in a certain type of people. Also don't write a too big description, half the people will just leave and the other half will answer just based on the title. Make sure to allow anonymous answers. And make sure people know how to answer.

Example:

If you ask "Does this guy look like Billy Connolly?", including a picture of Billy Connolly will get you more answers cause some people might not know him.

Make an option for everyone.

If people can't find an option that's right for them they'll likely just vote anything to see the results. Example:

If you ask "Do you prefer mars bars or snickers?" with 2 options, one for each, people who haven't tried them will vote randomly ruining the results.

Make more choices if possible.

More options = more data and it can help you understand what's actually going on. It's pretty similar to the last one really.

Example:

If you ask "Do you like chocolate" you can make the options "Yes" and "No" and there's an option for everyone, but some people might not like it simply cause they're allergic to it, though they like the taste. You may not care about those so you can make more options to filter the results and get the data you're after.

Don't ask something different in the description.

I've seen this so many times on this site. It can make it unclear which question the poll is for.

Example:

Title:"Do you hate spiders?"

Descrition: "Many people hate them but I like them. Wbu? Do you like spiders?"

Options: "Yes", "No"

You don't know which to vote, half the people will vote based off the title and the other half based on the descrition cause it's what they saw last.

Use words that everyone understands.

If you use words that only people who would vote yes understand that can of course ruin your results.

Example:

If you ask "Have you ever taken Es?" you'll get a reduced amount of votes from people who haven't, cause they're less likely to know what it is.

To conclude

I think that's the top and bottom of it, I won't go into any advanced stuff to save you from the maths. I know it's pretty obvious stuff but maybe some stuff you hadn't considered when asking questions. Hope you found it useful or at least interesting. Have a nice day :)


1|1
2|2

Join the discussion

0/2500

Submit

What Girls Said 2

What Guys Said 2

  • 28d

    Good advice, but most users don't care to. They're more into confirmation bias

    1|0
    0|0
  • 28d

    Good advice on posting polls on GaG

    0|0
    0|0
Loading...