Linguist Forum
General Linguistics => Feedback, Help and Forum Policy => Postings => Topic started by: HVandeWiele on December 20, 2013, 08:43:36 AM
-
Dear all,
as part of my master thesis, I will run a psycholinguistic experiment on the processing of questions in English. I first need to check the naturalness of the stimuli I am going to use. The questionnaires available on this page aim at collecting such judgement:
https://sites.google.com/site/linguisticquestionnaires/home
If you are a native speaker of English and want to spend 15 minutes helping out, that would be wonderful!
Many thanks in advance to the participants, and merry Christmas to all!
-
Ok, done. I hope that helps!
Note: there are some strange typos "desesperately", "My grandmother is nighty-two years old.".
Many of these sentences (questions and answers) do not sound natural in form (honestly they sound non-native). An example: "Which store is cheaper finally?" / "Is it true that he beats his wife finally?" -- "Finally" in those sentences is just bizarre. But I don't think it affected my choice of response.
I focused on meaning. I don't know if that is a problem for your research. You should have native speakers individually rate everything on a 0-5 scale if you want to know that everything is natural, then you can check which one fits which context best.
I would strongly suggest asking a native speaker to proofread this before you actually try to run it as an experiment-- that's very important for reliability!! Honestly (and I hope this is helpful, not insulting) if I were a reviewer for a journal, or your advisor on the thesis, and I saw that these sentences were so unnatural, I would not accept the results of your experiment.
Also, I think my answers were based primarily on closest association with the answer, often with overlapping words. If it said "Don't eat chicken" I chose the response involving the word "chicken". The actual linguistic form (grammar, etc.) didn't seem to be a factor. I didn't always choose just based on overlapping words, but that seemed to be the major pattern.
Finally, I'd suggest using Amazon's Mechanical Turk if you want a lot of data quickly.