If you do research that involves collecting data from human participants, and most psychology research does, you need a survey platform. Most people default to Google Forms because it’s free and familiar. Some go to commercial tools like Qualtrics or SurveyMonkey because their institution pays for them. A few build custom solutions in code.
I use LimeSurvey for all of my research, and have for years. It’s free, open source, and self-hosted on my own server at survey.psycholo.ge. In this post, I want to explain what LimeSurvey is, why I chose it, and why I think more researchers, especially in psychology and the social sciences, should be using it.
The Moment I Switched
I first started using LimeSurvey in 2020, during the COVID-19 pandemic. Like a lot of researchers, I suddenly needed to collect data online at scale, and I had a study design that was more complex than what off-the-shelf tools could comfortably handle. I needed conditional logic, filters, branching scenarios. I needed certain questions to appear only for certain respondents based on their earlier answers. I needed to control which response options different participants would see.
Google Forms simply couldn’t do it. The conditional logic was too limited, the question types too basic, and the scalability was uncertain for the number of respondents I expected. I needed something more powerful, and ideally, something I controlled myself.
That’s when I found LimeSurvey, and I haven’t looked back since.
What LimeSurvey Actually Is
LimeSurvey is a free, open-source survey platform with a wide range of built-in question types: single choice, multiple choice, matrix questions, ranking, sliders, file uploads, and many more. It supports complex conditional logic, sophisticated branching, multilingual surveys, and quotas. It can handle anonymous and tracked responses. It exports data in multiple formats including SPSS, R, CSV, and Excel.
The first big difference between LimeSurvey and a commercial tool like SurveyMonkey is that LimeSurvey is genuinely open source. You can install it on your own server, modify the code, build your own question types if the built-in ones don’t fit your needs. That last point is something you simply cannot do with SurveyMonkey or Qualtrics. With them, you get whatever the company decided to give you. With LimeSurvey, you get whatever you can build.
The second big difference is the response limit. Commercial platforms charge based on usage: number of responses, number of surveys, number of questions per survey. With a self-hosted LimeSurvey instance, you can collect as many responses as you want, run as many surveys as you want, ask as many questions as you want. The only limit is your server’s capacity.
Self-Hosted vs. Cloud
LimeSurvey actually offers both a cloud version (LimeSurvey Cloud) and a self-hosted option (LimeSurvey Community Edition, which is free). I chose the self-hosted route, and I’m running it on the same infrastructure that hosts the rest of the Psycholo.ge ecosystem.
The trade-offs are real. The self-hosted version is much cheaper, especially when you’re collecting large volumes of data. I have unlimited responses and full control over the system. I can integrate it with my own workflows, customize it deeply, and keep all the data on infrastructure I trust.
The cloud version has its own advantages, particularly around scalability. My self-hosted instance can comfortably handle around 150-200 simultaneous respondents at peak. For most of my research, that’s more than enough. But if you’re running a major survey with thousands of people responding at the same moment, the cloud version is far more robust at scale.
For my purposes, such as academic research, psychometric validation studies, replication projects, recruitment for my own program, self-hosted is the right choice.
What Conditional Logic Actually Looks Like
Let me give a concrete example, because abstract talk about “conditional logic” doesn’t capture why this matters.
For my Cultural Life Scripts replication study, respondents are asked to write down 7 important life events in their own words. Then, in a later section, they need to evaluate each of those 7 events on several dimensions – emotional valence, age expected, importance, and so on.
In a basic survey tool, you’d typically present this as a matrix where the rows are labeled “Event 1,” “Event 2,” “Event 3,” and so on. The respondent has to remember what they wrote for each numbered event and mentally map their evaluations back to it. This is cognitively taxing and a major source of measurement error.
In LimeSurvey, I can pull the actual text the respondent typed earlier and use it as the row label. Instead of “Event 3,” they see the actual event they described, say, “graduating from university.” The matrix shows the events by name, in the respondent’s own words. The cognitive load drops dramatically, and the data quality improves.
This kind of dynamic content insertion is exactly the sort of thing that simpler tools can’t do. And it’s just one of many examples. I can also enforce logic checks (minimum age, response validation, range restrictions on numeric inputs), randomize question order to control for order effects, and route different respondents through entirely different paths through the survey based on their earlier answers.
Customization
Beyond what comes built in, I’ve done significant customization on my LimeSurvey installation. I have a custom theme that matches the Psycholo.ge Solarized aesthetic. I’ve modified some default question behaviors to better fit Georgian-language conventions. And in one project, I built a custom question type from scratch, a question that displays a specific text passage to respondents and records data about their reading behavior, all saved automatically to the server.
This level of customization is simply impossible with closed-source platforms. With LimeSurvey, the codebase is yours. If something doesn’t work the way you need it to, you can change it.
The Learning Curve
Building basic surveys in LimeSurvey is genuinely easy. The interface is straightforward, and once you understand the basic structure (survey → question groups → questions), you can build sophisticated instruments quickly.
The deeper customization, server configuration, theme development, custom question types, does require more technical comfort. But you don’t need any of that to get started. And when you do hit something tricky, the LimeSurvey community is one of the most helpful technical communities I’ve encountered. The forums are active, the responses are thoughtful, and even niche questions get serious engagement.
If you have a researcher question, someone has probably asked something similar. If you have a technical question, someone has probably solved it.
Should You Switch?
Here’s my honest answer: yes.
I genuinely believe more researchers should be using LimeSurvey. The freedom to design exactly the survey you need, the ability to collect unlimited responses without watching a meter, the option to build custom features when standard ones don’t fit, the export formats designed for proper data analysis, all of this adds up to a research tool that respects how research actually works.
It is not overkill, even for simple studies. The basic survey-building experience in LimeSurvey is comparable in difficulty to Google Forms, but the ceiling is enormously higher. You start with a simple survey today, and tomorrow when you need conditional logic or a custom question type, you don’t have to migrate platforms.
For Georgian researchers in particular, this is also a question of independence. We don’t need to depend on commercial platforms hosted elsewhere. We can build our own infrastructure, on our own servers, in our own languages, fully under our control. That’s exactly what survey.psycholo.ge represents – a Georgian-hosted research platform that any of my collaborators or students can use for their own work.
The tools we use shape what research is possible. Better tools mean better research. And LimeSurvey, in my experience, is the best research survey tool out there.
—
Giorgi Tchumburidze
May 2026