Online Surveys

An online survey asks a sample of people from a larger population a series of questions through a website or app, and a computer program then collates and analyses the answers. Because answers are stored digitally, survey software can collate them straight away.

Carrying out an online survey

You write the questions, publish the survey online, share the link (by email, social media or a QR code on a poster) and collect the responses. Survey software then adds up the answers and draws charts, so a thousand replies can be turned into statistics in seconds:

ratings = [4, 5, 3, 5, 2, 4, 5]  # star ratings from survey replies

responses = len(ratings)
average = sum(ratings) / responses

print("Responses collected:", responses)
print("Average rating:", round(average, 2))

Output:

Responses collected: 7
Average rating: 4.0

Advantages

• Large numbers quickly: a link can reach thousands of people in a day, far faster than posting forms.
• Low cost: nothing is printed or posted, so creating and sending it costs almost nothing.
• Wide, targeted distribution: you can share it in a gaming forum, a school app or a hobby group to reach exactly the audience you want.
• Automatic analysis: answers are counted instantly instead of being typed up by hand.
• Anywhere with internet: people can reply at home, on a bus or at break.

Disadvantages

• Digital divide: it needs a device and internet access, so people without them are missed and the sample may not represent everyone.
• Emails can land in spam or junk folders and never be seen.
• People may lie, skip questions, give joke answers or give up halfway, which spoils the data.
• Fixed tick-box options can stop people giving the answer they actually want.
• Poorly worded questions or unclear instructions produce poor data, however many people reply.
• Data could be lost, and storing personal details creates a cyber security risk.

Surveys are strong for quantitative statistics and concrete results, such as “62% of players would pay for extra levels”. A survey used to produce statistics, though, may not be scientifically high quality: with a small or unrepresentative sample, or leading questions, the figures can look convincing while proving very little.