Due 30 November 2012

In this assignment you will:

1. develop a hypothesis that you can test,
2. write a script to collect data,
3. collect some data,
4. write a script to analyze the data,
5. write a short informal presentation of your results.

Remember our mantra that scientists use computers to gather, analyze, and visualize data. In this assignment you'll do all three.

First you should develop a hypothesis that you can test. A simple experiment would have the subject (you and possibly your friends) type in response to a prompt from the computer. The measured variables could be accuracy and time. For this exercise you are free to develop your own hypothesis though I encourage you to keep it simple because the number of trials required (how much typing you have to do) grows rapidly as the number of degrees of freedom in the experiment rise.

I am open to experiments involving data from your discipline; just run your proposal by me via email.

Example hypotheses could include:

• Typing home keys is faster than others.
• Typing with the dominant hand is faster.
• Locating misspelled words for which the first and last letters are correct and the inner letters are permuted is harder than locating random letter substitutions.
• The time required to type a response of length N fits an equation of the form A*N + B.
• Typing common digrams (pairs of letters) or trigrams (triplets of letters) is faster than uncommon ones.
• Typing prompts that are slight misspellings of known words is slower than typing correctly spelled words.
• When typing prompts in reverse (i.e. the computer prints dna and you have to type and), the response will be (faster/slower) when the prompt or reversed letters make a common word.

You can use one of these or come up with one on your own. I'll be pleased to give you feedback on candidate hypotheses before you try them. You'll want to randomize your trials to try to minimize the effects of fatigue and training on the results.

After you come up with a hypothesis you'll need to write a script to collect the data. Your script will print some instructions for the participant and then run a series of trials in which the computer prints a prompt and the subject types a response. You should give the subject some feedback on performance and a short rest every 20 or so trials. Your script must record the data to a file for use in a separate analysis script. Dividing things like this will prevent you from having to rerun the data collection when you are debugging your analysis. Your script should automate all aspects of running the experiment and writing the data to a file.

Now that you've got some experimental data, you'll write a script to analyze it to see if you can make any claims about the correctness of your hypothesis. Your analysis script will read the data from the file where the experiment script recorded it.

Finally you'll write a short informal description of your hypothesis, your experimental procedure, and your results. Your description should include an appropriate visualization and statistics on the results. If you know, from other courses, about the statistical analysis of experiments (power analysis and such) include that in your discussion but it is not required.

The description as a PDF file, your scripts, and your data files should put into a zip file and uploaded via Sakai before midnight on the due date.

There is no template for this assignment.