Resources

Find out more about the tools used in the lab to conduct our research.

In-house equipment

The lab is home to several tools for online experimentation. We house an anechoic chamber for high-quality recording and presenting auditory stimuli, as well as two SR research eye trackers.

• SR Eyelink 1000 We use this tower-mounted eye tracker, which uses a high speed camera to record eye movements at up to a 2000 Hz (twice per millisecond) to study reading in a natural setting. The camera can also be used with a desktop mount to permit free movement of the head.

  See lab video on the Sign Up page for an overview of a reading experiment.

Offline judgments and corpora

• Ibex Farm. Alex Drummond's web-based tool for collecting judgments online. My tutorial from a talk at USC on how to use Ibex Farm with Amazon's Mechanical Turk, along with a handy Excel script for formatting the materials, and a sample R-file, can be found here.

  http://spellout.net/ibexfarm

• English Lexicon Project. Powerful database for collecting lexical characteristics such as frequency, neighborhood size, bigrams, etc.

  http://elexicon.wustl.edu

• English pond. Another tool for extracting lexical characteristics from a list of words from English, Spanish, French, Dutch, and German. Particularly useful for determining neighborhood size.

  http://clearpond.northwestern.edu/englishpond.html

• UCLA Corpus resources. A summary with links of the corpora accessible on the department's server.

  http://www.linguistics.ucla.edu/general/CorporaSummary/Corporasummary2.htm

Real time data collection

• Mouse tracker. For presenting mouse trajectory experiments, with a very intuitive interface.

  http://www.mousetracker.org/

• Linger. A versatile tool for lexical decision, priming, and self-paced reading. Requires the Tcl/Tk framework (and the occasional hack).

  http://tedlab.mit.edu/~dr/Linger

• UMass eye tracking software. Multiple programs developed by Chuck Clifton and others at UMass Amherst for all aspects of text based experiments. For use with the SR Eyelink II or 1000.

  http://www.psych.umass.edu/eyelab/software

• Experiment builder. From SR Research, a nice suite of presentation and analysis tools for just about any type of eye research. We've used it for a pupillometry experiment. Great software support, too.

  http://www.sr-research.com/eb.html

Lab developments: Comments and suggestions welcome!

The following code was written by Jesse Harris, unless otherwise noted. Please send him bug reports or suggestions for improvements directly.

• CombineResults.rb. Simple to the point of trivial, this Ruby program writes results from Linger's .dat files to a single file with the experiment name automatically appended along with the number of subjects run. Primarily for DOS phobics. If Ruby is installed on Windows, simply place in the same folder as your .dat files, and then double click on the icon to run. Also works with Mac and Linux.


• Robodoc.py. We've collaborated with Adrian Staub and Chuck Clifton on a new version of Robodoc, a program that automatically cleans eye tracking data of blinks and track losses. Learn more about this handy code here.


• NPR web crawler. A few Python programs that crawl NPR and download transcripts into XML format and links to audio files of radio interviews into a directory. Can be tweaked to crawl other news sites. Requires a working knowledge of Python. To be posted with instructions soon!


• Eye movement while reading corpus. This corpus collects eye movements from 12 Pomona College students in 2014, silently reading 25 articles from the Penn Treebank. These texts were taken from the freely accessible sample of the Wall Street Journal included in the Natural Language Tool Kit. Materials were chosen to align with already parsed text. Currently, only raw asc files are available for download, and may be examined using EyeDoctor. The corpus will be updated with cleaned and processed data shortly, formatted along the lines of the Dundee corpus.


• Los Angeles Reading Corpus of Individual Differences. The Los Angeles Reading Corpus of Individual Differences (LARCID) is a corpus of natural reading and individual differences measures. The corpus is currently a feasibility pilot of eye tracking data collected from 15 readers. Five texts from public domain sources were included. In addition to the eye tracking measures, a battery of individual difference measures, along with basic demographic information, was collected in a separate session. Individual difference measures included the Rapid Automatized Naming, Reading Span, N-Back, and Raven's Progressive Matrices tasks. Pilot data, write up, and R-markdown files can be found on this Open Science Framework page. Sample reading sessions shown below: