LIWC Extra Credit

Goal: Explore patterns of language using LIWC

Note: This exercise is worth 1 extra-credit point.

Step 1: Planning

Plan for two sets of texts you wish to compare. For example, you may be interested in the language of friends vs. family, males vs. females, authority figures vs. subordinates, different presidential speeches, etc. Think of some psychological contrast that interests you (if you need some help, check out the slides and reading from the LIWC day).

Step 2: Creating your own "corpus" of texts

Once you have come up with a single contrast that interests you (e.g., two different US Presidents), then seek out texts. You might have N = 6 total text samples, 3 per person. Save these texts either in a Word document or in separate text files (e.g., "Clinton1.txt", "Clinton2.txt", etc.).

Step 3: Data collection and results

When you have all your files together, open them one at a time. Copy and paste their text contents one at a time in LIWC's free online interface:

http://www.liwc.net/tryonline.php

When you submit each, LIWC will give you about 7 semantic features that you can analyze, and the two columns to the right will tell you how that text compares to formal / informal texts.

Because you are copying/pasting these texts one at a time, make sure you get 6 LIWC outputs. Take a look at the results overall. Calculate averages that interest you (you don't have to use all features, just a handful is fine). What did you find? Is there anything surprising or interesting? Of course, with so small a set, your inferences are limited. However, I invite you to speculate about psychological processes!

Write at least 200 words answering the questions above, and share it with your TA.