Assignment:
Categorical Perception Experiment
Categorical Perception Experiment Assignment
With thanks to Professor Toby Mintz and to Dr. Ed Holsinger.
First, you will carry out two experiments, in sequence, collecting the data online. These experiments explore the phenomenon of Categorical Perception in speech.
Then for your written homework assignment to turn in, you will next write a 3-5 page research report (plus one page having your two results graphs). More on this below.
Turn in your report as a pdf file to the class assignment dropbox. You have the link to upload this in an email from Professor Byrd. Please make sure the filename starts with your last name, e.g. SmithSally_CategPercep.pdf.
Instructions for Collecting the Data via the Online Experiment Portal
This Experiment Portal was created by Ed Holsinger and Kristin Peterson. Thanks!! The stimuli are from Stephens and Holt.
For this lab assignment you will carry out two experiments. You will be the experimental subject. The experiments are demonstrations of the perceptual phenomenon of Categorical Perception.
You should use headphones so that you can clearly hear the stimuli and so that you can block out other noise in the room as much as possible. (The quality [i.e. quiet] of your listening situation and attention will improve the interpretability of your results.)
After you carry out the experiment, you will need to save the Results page—the page with two graphs—to a pdf. So know ahead of time how to create (print to) a pdf from your browser. This page MUST be attached to your submitted assignment. You will receive no credit if this is not attached.
After you have carried out the experiments and have saved the results graphs, you will write a brief report as described below.
First go to the following website to participate in the experiment:
Make sure you have time; the experiment will take approximately 20 minutes to complete. You should finish it in one sitting. (Turn off ad-blockers, as some have been reported to disrupt the display of your results.)
Click the "Experiment #1 - BADAGA" green "START" button.
Answer the briefing questions at left—for the Class ID type "Core103_Spring 2021_YourLastName_YourFirstName"—and click "Submit". You will see some instructions; you will press "Start" to begin the experiment.
2a. Then for Experiment One—the Categorization (or Labeling) portion of the study,—you will choose "ba" "da" or "ga" for each item you hear. Note that you can use the keyboard to log your responses: 1=ba, 2=da, 3=ga, space=Play Next. Each syllable will not be a 'perfect' ba, da, or ga; pick the response that the sound best resembles.
2b. Then for Experiment Two—the Discrimination portion of the study—you will choose which ordered pair of stimuli you hear: "ba-da" or "da-ba" or "da-ga" or "ga-da". Again, you can use the keyboard to log your responses: 1 (left pair), 2 (right pair), space=Play Next, as indicated. Note the pair may not sound exactly like the choices you see--simply choose the best choice you can. If you're not sure, just guess.
3. After you finish responding, you will get a Results Page that looks something like (not exactly like!) the figure below. The top graph is your Categorization or Labeling Results. A "1" on the vertical access indicates that the response occurred 100% of the time; a 0 indicates that the response never occurred. The bottom graph is your Discrimination Results. A "1" on the vertical access indicates that the two stimuli in the pair could be 100% reliably distinguished, i.e. that their order was perceivable on each and every presentation of the pair. Values less than "1" indicate that the items in the pair could not be reliably distinguished, with .5 being a chance response, meaning that you could only randomly guess what order the pair occurred in.
Details: The pair-numbers across the bottom compare stimuli that are three steps apart. That is, pair 1 is stimulus 1 versus stimulus 4; pair 17 is stimulus 17 versus stimulus 20; etc.4. Print this Results browser window page to a pdf. You MUST turn this in as the final page of your assignment so make sure to save it as a pdf.
Written report of experiment to turn in
For the written portion homework assignment to turn in, you will now write a 3-5 page research report (plus one page having your two results graphs [see above]). The structure of this report will be loosely based on the structure of a typical scientific research paper. Although you did not design these experiments, when writing this assignment, try to write it as though you had a specific hypothesis you were experimentally testing regarding the perception of speech along an acoustic continuum. The report will contain four sections, outlined below (plus your previously saved experimental results graphs). Give each section the appropriate section heading.
Turn in your report as a pdf file to the class assignment dropbox. You have the link to upload this in an email from Professor Byrd. Please make sure the filename starts with your last name, e.g. SmithSally_CategPercep.pdf.
Section One: Introduction
Identify the phenomenon that the two experiments are designed to explore. This section should explain why data from the two experiments (as opposed to just one) are important.
Section Two: Method
Describe the method you used to investigate the issues you discuss in the Introduction. This section should provide the reader with the following information:
• What were the two experiments? • What did the subject (you) do in each experiment? • What equipment did you use? •What were the linguistic stimuli and how were they presented [See my note below on this]. • What was the response measure for each experiment? • What is the predicted outcome? Why?
A method section in a formal experiment report would also report on the important stimulus properties and on the stimulus presentation. If you are interested in how these stimuli were synthetically created, you can have look at this paper (e.g., Figure 2 [Ca]). For our purposes, stimulus properties can be characterized as follows: Each stimulus was one step in a continuously and gradually varying acoustic space that morphed three natural, American English speech productions of [ba], [da], and [ga] from one to another to create a 20-step continuum. A formal method section would also note details of the stimulus presentation in the Categorization/Labeling Experiment and in the Discrimination Experiment. For our purposes, the Labeling task presents each of the 20 stimuli twice, in random order. In the Discrimination task , each pair is three steps apart on the 20-step continuum. Each pair is presented twice in each order. So the result for, say, pair 1 [comparing labeling stimulus #1 to stimulus #4)] is the percent of directionally correct responses for those four trials. Trials are randomized in all cases, and the randomization is handled by the engine at the start of the experiment (so everyone will get a different order). You may copy/edit the information in this paragraph and include it in your paper.
Section Three: Results
Present the data you collected. Describe what is displayed in each graph (that is, what the graphs represent and what they mean). Make generalizations and describe patterns when you can: For example, you might say, “My responses to stimuli 1-3 were 100% ‘ba” (rather than “My responses to stimulus 1 was 0% ‘ga’, 0% ‘da’, and 100% ‘ba’. My responses to stimulus 2 were …") Your summary of results should provide the basis for connecting (in the Discussion; next) the pattern of results found in each of the two experiments.
Note, a pdf created from your browser of your Results webpage MUST be submitted as a final page of the turned-in assignment. Assignments without this will receive a zero.
Section Four: Discussion
This is where you remind the reader of the purpose of the experiments, and you explain your findings. What did the two experiments demonstrate? How do they address the questions asked in the Introduction? Explicitly connect the pattern of results in graph one with the pattern of results in graph two. To what extent were the predictions supported? In addition, discuss any limitations or drawbacks you see in the experiments (e.g., sources of measurement error) and how these might affect your results and conclusions.
Section Five--Appendix: Attached page that is the pdf of the Results browser window
This is required. After you've completed your written report and saved it as a pdf file, insert your experiment results page (which you previously saved as a pdf) as the last page of the report pdf and save again as to form the completed pdf to turn in. Your complete file to turn in to the assignment drop box should have your report followed by your results graphs. Both must be included for credit on the assignment.
My sample results. Yours may likely look a bit different.
Turn in your report as a pdf file to the class assignment dropbox. You have the link to upload this in an email from Professor Byrd. Please make sure the filename starts with your last name, e.g. SmithSally_CategPercep.pdf.