Tips/tricks for conducting internet research
From PsychWiki - A Collaborative Psychology Wiki
Revision as of 03:25, 21 June 2008
Got a tool which has helped you do your research more effectively. Share your experience here.
Read more about why you should contribute to Psychwiki
Hosted vs. Installed vs. build your own
There are 3 main ways that you can go about doing research on the internet.
- You can use a pre-built product that requires no technical skill beyond the ability to use a web interface with the same complexity as most web mail programs. This is the easiest and most appropriate for basic surveys for people who don't have the desire to learn web technology.
- You can go get server space, find software that has been written by someone else, and follow the instructions to install it. This requires some technical ability, but the advantage is that you can learn what you need to learn to customize things. As well, the fees for server space are often lower than the fees for a fully hosted solution.
- You can go get server space and write your own program from scratch. This is obviously the hardest thing to do, but is a requirement for some studies where pre-existing solutions do not have anything resembling the functionality that you need.
For more information on conducting research on the internet and an online wiki directory of survey hosting services and survey software, visit the Cii Wiki at http://cii.wikispaces.com
► Hosted Solutions
http://www.surveymonkey.com - This is perhaps the most popular hosted solution on the internet. The free version allows 100 responses to 10 questions. The "professional" solution costs $20/month and allows 1000 responses with an unlimited number of pages/questions. There is some advanced functionality such as the ability to skip questions, randomize response order, and require responses to questions.
http://www.qualtrics.com/ - I searched through all the online survey software websites a few months ago looking for the best options to conduct my studies online, and finally decided on Qualtrics. Their software has the most options, and is the easiest to use. They have educational prices, but its still not cheap.
Similar to Survey Monkey, but higher priced ($49/month+) and I don't see any additional functionality.
They have all the basic functionality (question presentation logic and randomization) and tout their service as being more secure which justifies their higher price (they charge per question and per respondent). I do not have any experience with them to be able to validate their claims either way.
http://www.experimentcentral.org/tess/ - This NSF funded project allows one the ability to do a survey on a representative sample of internet users, something which is almost impossible to do on your own. Unfortunately, it looks like funding may run out for this project as of this writing.
► Code you have to install
http://psych.fullerton.edu/mbirnbaum/web/surveyWiz.htm - This tool allows you to create HTML which you can then upload to your web site. A full discussion of it accompanies Birnbaum's book, Behavioral Research on the Internet (http://psych.fullerton.edu/mbirnbaum/web/examples.htm). The data will either be emailed to you or be downloadable from the author's website.
Marketing Survey Tool
This code, from http://www.surveyworld.org/, does some wonderful things once you figure out how to configure it. It can be used to create randomized surveys with rotating blocks of questions. Some questions (ie. informed consent, debriefing) can be placed at fixed spots while others appear in random order. Not only the questions can be randomized, but the responses can be randomized. Questions can take text, numeric, multiple choice, radio button, matrix, or drop down list box responses. Data is easily downloadable in comma delimited form for import into a statistical package. Of all the tools I've seen, this one works the best. The only drawback is that it requires some work to configure without the best documentation in the world.
► Build your own tool
Not for the faint of heart, but you can do some amazing things once you know what you're doing. For example, one can build in Flash or dynamic HTML that allows a user to interface with an experiment in much like they do with an arcade game. One can add multimedia and time responses. One can do longitudinal studies with automatic emails generated to solicit continued response.
Teaching you how to do build an internet research experiment from scratch is not possible on this page, but we can point you in the right direction.
- Start with our FAQ page where you can see answers to common questions and ask your own questions -> http://www.psychwiki.com/wiki/Internet_Research_FAQ
- See also the APA Online Psychology Laboratory which has a section called Developer's Corner where you will find various resources designed to help you design and build web-enabled psychology studies including code for selected studies posted on the OPL.
◄ Back to Research Tools mainpage