This is probably one of the most complex topic in survey development. Good on you that you think about this, many who create survey don't even try.
First of all, it highly depends on your target group and your topic.
If the topic is relevant to your target group, you are already one step ahead. Many people actually want to share their expertise, knowledge or opinion. But if the topic isn't of interest for the participant, they will loose interest soon. So make sure you aks the right people.
Keep your surveys short and simple.
Gamifying is an option, but it is also dangerous. Depending on the target group, it might be the wrong move. In my experience, physicians are not having it. They want things to be straight forward and objective. Try to approach them with projective research methods and they will get easily annoyed. The prefer clean, direct questions, without any "play". Now for a younger audience gamify the questions might be vital, especially for kids.
Try to avoid long, boring item batteries. Long array type questions usually have a high dropout rate. Try to give your respondents the answer options they want to give. Nothing more frustrating than to have answer options and none of them apply to your case. At least for me. Do good planning, do some qualitative tests of your surveys, see if you cover all the answer options possible.
In market research we usually offer some "incentives". Either a little present, some amount of money, a prize draw, etc. to keep people engaged. But be aware that this might attract people to the survey just for the sake of the incentive and this might also increase the possibility of fraud in the survey, so be delicate with that and don't go full in. Make it a token of appreciation for the time your target group spends on your surveys, not something that will attract the wrong crowd.
In your case, if you want them to take part in multiple surveys, offer something at the end if they completed all surveys, for example.
But I think the most important thing is what I have mentioned at the beginning: Make the survey relevant to your audience. If the topic interests them, things are a lot easier.
In practice it is a lot harder to get people to participate on market research for insurance than for cars, etc.
Don't just place question after question, try to lighten things up from time to time with a little feedback. Let them know where they are in the survey and why you ask those questions, "interact" with the survey takers.
Usability: Think about your users when creating the survey. A specific question type might be nice for you as a researcher, but could be a nightmare to respond.
Don't repeat the same questions over and over again. Sometimes you need to ask the same item battery for different stimuli, brands, products. Keep the number of repetition to a minimum. You don't need to ask every one about every iteration. You could give each respondent x out of y iterations and rather include more respondents than to torture your respondents with the 10th iteration of the same item battery. Rather ask only about 5 brands and get another respondent to answer about the other 5 brands. You will need a bigger sample, but you don't scare your respondents away that much.
Joffm, Tpartner, Jelo, Tammo and Denis will surely have something to add to this.
I answer at the LimeSurvey forum in my spare time, I'm not a LimeSurvey GmbH employee.
No support via private message.
