Online surveys

…there are some pitfalls to observe – depending on how you do it. If you haven’t read the chapter about target groups yet, please do so before you continue reading.

Two different approaches fall under the category online survey. Yet, they have entirely different consequences for your results.

• You have a list of Email addresses (or WhatsApp/Facebook etc. contacts) and you ask those relevant for your survey, but only those, for participation. Either you send them the survey as an attachment or you secure your online questionnaire on a platform with a password they get.
• You have an open to all online survey where the participants will self-select. You ask people in relevant community networks to participate and advertise in media or on your and your friends blogs.

For both approaches you need to reflect:

• Does your whole target group have good access to the internet and is used to the technology – even if someone has an email address he might not know how to answer questions on the net. Which parts of the group will be excluded through an online survey?

• Think about data security. Just as when you ask people face to face they will have to believe you that you will keep their information anonymous. But when communicating through the net, others may access the data. Is this a realistic threat? If yes, use an online survey platform which guarantees security.

• The anonymity of the net may prompt more honest results than face to face questioning. But you have no spontaneous feedback and no direct communication online! Therefore, your questions have to be very clear, comprehensible for everyone and to the point.

Using Email or specific Social Media contacts

For an online survey where you know beforehand who you are reaching the same considerations apply as we discussed for random and convenient samples. It is just a different communication tool, not a different method.

If you have a complete (or almost complete) list of email addresses of your target group, you can do a random sample and get really sound results. Maybe you can get the list of all students at your college? Or the religious community you want to survey has a list of their members – and would give it to you? Or you work in a hospital and your target group is medical personnel?

If you don’t have such a list, but use the email and social media contacts you have or any kind of lists you can find, then this would be a convenient sample and you have to observe the same considerations as described in the chapter about convenient samples.

Using email and not face to face questioning has one shortcoming: Not everybody will answer. The discrepancy between your target group and the number of people really participating is called a coverage error. You must mention the coverage error (by percentage) when you publish your study. If half of those contacted fill out the questionnaire your results are still significant. If less than half answer, you should contact them again.

Self-select Option

This is an interesting option, when it comes to intimate questions – and usually questions concerning FGM are considered intimate by many women. Because the participants are sure that no one, not even you, knows who they are, they may answer more honestly.

The central problem is: This is definitely not representative. This is even the case if you find in the end that the participants’ background represents more of less what would be expected in your community/society. People tend to answer questions voluntarily when they are interested in the topic. The “silent majority” might have completely different attitudes. Those who participate in your survey most likely have already heard about FGM, probably they are already aware of the problems surrounding FGM, and you will attract a large number of people who have an outspoken opinion about the issue – both: pro and con.

Most likely the percentage of people who answer that FGM exists in their family/surrounding will be much higher than it actually is in your target group, because people who have never heard of it will simply not participate in your survey. Probably the number of parents who have stopped committing FGM on their daughters and are opposed to it will also be higher, because people who think it is normal and don’t question it might not be as interested in your survey.

But the opposite could happen as well. Maybe some FGM defenders will try to blast your survey and even give fake answers under false flag. Some participants may not belong to your target group at all, e.g. you are targeting women in Kuwait, but women from Iraq or the U.A.E. find your survey and don’t understand that it is not for them. To avoid this you need to verify in your first question that the participant belongs really to your target group. This should also cancel out migrant workers if you don’t want them in your target group.

Now, you may wonder: What use is such a self-select survey if the results can come out so wrong?

• It can tell you if FGM is (still) a relevant problem at all
• It can give you a clue in which subgroups of your target group it is a problem and in which not. For example: If you ask women from Iraq to participate and only Sunni women do then you can assume that Shia women are not (as) affected.
• It can give you useful results on motivations, how it is done, what age etc.

Last, but not least, it can be a campaigning tool. By advertising your survey in different networks, groups and media outlets you are raising awareness for the issue.

There are free tools for creating online surveys like SurveyMonkey or pick one from the Survey Software Comparison on Wikipedia.

Next

Go to How to use the basic questionnaire.