You will learn to work with Ibex and Ibex farm in this lab meeting.
To get you started, I prepared skeleton files for you and specified a few tasks, which, if you go through them, should help you understand how to use Ibex for your own experiments. This is not a graded assignment! Feel free to come up with other tasks and to try different features of Ibex.
- Download/clone all the files from this repo
- Go to http://spellout.net/ibexfarm/ and create an account.
- Donwload the manual for Ibex: http://spellout.net/latest_ibex_manual.pdf. This will help you if you get stuck.
- Create an experiment in Ibex farm and upload the following files from the folder Preference task:
- example_intro.html into chunk_includes (delete an old file in the folder, if there is any)
- judgements.js into data_includes (delete an old file in the folder, if there is any)
- Go through the experiment once; when you are done inspect results in results; these should be self-explanatory. After answering each of the following questions, you should always run the experiment and check the results to see that your modifications worked
- Add extra items to the experiment.
- Add extra fillers to the experiment.
- Modify the html form, add a new column to collect information.
- Add two more conditions per item (e.g., another quantifer).
- Download the results in your R working directory and rename them as "results.csv". Then, run the script data_munging.R in the same folder where you stored results.csv. You should be able to load your results into R. The R script specifies three tasks. Carry them out.
- Create a second experiment and upload the following files from the folder Acceptability:
- example_intro.html into chunk_includes (delete an old file in the folder, if there is any)
- global_stylesheet.css into css_includes (not obligatory)
- acceptability.js into data_includes (delete an old file in the folder, if there is any)
- Go through the experiment once and inspect the results.
- We now want to test that each has to satisfy the Differentiation condition (discussed in class 18-02), unlike every. Think of a way to test this in Acceptability task that does not involve inverse quantifier scope.
- Write a few items following your idea. Create at least one full loop for the Latin square (you have to have as many items as conditions).
- Add two fillers.
- Run the experiment one more time, store the results and adapt data_munging.R script that you used before. Using the script, load your acceptability data into R and calculate median Answer per condition.