This document summarizes some general guidelines for developing simple cognitive experiments using Psyscope.
The first section describes some issues that should be decided before even getting into Psyscope to write your experiment.
The second section provides a general description of the more basic features in Psyscope that you will typically use.
The third section provides a summary of important stages to work through in Psyscope when writing the script for your experiments. It is not intended to be a comprehensive explanation, but a checklist of ordered steps.
§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§
Section 1. What You Should Have Prepared Before Opening Psyscope
Much heartache and frustration can be avoided by carefully working through the following details before attempting to do anything on the computer. Generally, it is not a good idea to develop a study from scratch while staring at the computer screen. The following five points move from broader to more specific concrete issues to work out.
1. Think through the paradigm you have chosen.
For instance, what exactly is the task being asked of subjects? What kind of judgment are they to make? What are they being asked to respond to and how will they respond?
What sort of stimulus materials will be presented?
2. Determine what sort of design you wish to develop.
What are the critical independent and dependent variables of interest? Which of the independent variables are between or within subject manipulations?
What other variables must you consider balancing or controlling for?
3. What are the critical events within each trial a subject will see?
Which events in a trial will be constant and unchanging? Which events will differ from trial to trial? How many different ways can these attributes vary? Are there many different instances or just a few? For example, in the Simple RT experiment of the tutorial the foreperiod could take on 4 different values, while in a lexical decision task in which each word would be seen only once by each subject a different stimulus word would be needed for every trial in the study.
4. Determine how many different trials will be presented in the experiment.
Work out all possible trial types in accordance with the experimental design you established earlier. Each trial should be identifiable with a condition name that reflects which level of the independent variable(s) it represents. The correct response for the trial should also be recorded.
When you have done this it is also a good idea to check that you have the appropriate number of trials in each condition. Usually there are equal numbers of each.
5. Create a text file containing stimulus information.
At this point you should be ready to prepare a text file that will list the relevant values for the different condition types, the stimulus items, and other information associated with individual trials (e.g., position, color, correct and error responses, etc.).
§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§
Section 2. Important Features of Psyscope to be Familiar With
Although there are many features available in Psyscope we have identified a subset that provide a solid foundation for beginning (and advanced) researchers to examine a wide range of empirical questions.
Design window
Perhaps the most important window to understand is the Design window. This is the window that presents a hierarchical tree-like diagram of the elements in an experiment.
The Experiment icon is represented in the Design window as the red psi symbol. When the boxes to ³Show events² and ³Show lists² are checked, the overall organization of the experiment can be seen in the form of a tree diagram with the Experiment icon at the highest node. When the Experiment icon is opened the objects directly linked to it are listed. Here you can specify the order and frequency that these objects should be performed.
Design window Tools, Objects, and Events
Down the lefthand side of the Design window are iconic representations of different operations and elements that can be brought in to the Design window in developing a study.
Tools
The top icons represent Tools, or operations, that manipulate elements in the design window. These tools allow you to select elements in the Design window (for further editing or defining), to link elements to each other, and to cut links between elements.
Objects : Templates and Lists
The icons below Tools are Objects. Objects are organizing structures in the experiment.
Templates. Templates define a series of events comprising a trial. The Template is where the basic temporal sequence of events such as stimulus onset, offset, and response collection are set.
Lists. Lists are used to provide specific values for any attribute that will vary from trial to trial. For instance, in the Simple RT example of the tutorial the foreperiod could be 1500, 2000, 2500, or 3000 msec on any particular trial. These values were defined in a List from which the Psyscope program selected at random on each trial.
Internal Lists. The items in a List can be defined internally within the Psyscope script. When a list is defined in this way it is represented by the icon that looks like a single sheet of paper with the upper and lower edges curved under. Internal Lists are typically used when there are not many different values to be chosen.
External Lists. When the items in a List are prepared in a text file outside of Psyscope the External List object must be used. The icon for this object looks just like the icon for the internal list except that an additional blank page is overlaid in the lower right corner of the scrolled paper. External lists are typically used when there are many different trials in a study.
Events in a Template
Event icons are visible beneath the Object icons along the lefthand side of the Design window. These icons represent the many different types of events that can occur within a single trial of an experiment. These events can, and frequently do, take on different values in the course of an experiment. For instance, the blank delay might vary from 1500 to 3000 msec, or a different word (Single text event) might appear on each trial of a lexical decision task. The events you will frequently encounter are listed below.
Blank time delay (icon : hourglass) : This event presents a blank screen and is typically used between trials.
Text event (icon : pink A) : This event presents any alphanumeric stimulus event. These may be words, nonwords, or numbers -- basically any symbol on the keyboard. Sentences may also be presented as long as they do not extend beyond a single line.
Paragraph (internally defined; icon : blue rectangle) : This event presents a page of text on the screen starting from the upper left and wrapping around down the screen. Typically used for presenting short instructions, but could be used for presenting short stories.
Document (defined in an external text file; icon : grey page of text) : This event serves the same function as the Paragraph above, but the text is prepared in an external text file. Useful when large amounts of text are to be presented or when contents of the text will vary from one trial to another.
Pictures (in PICT form: icon : frame with paintbrush) : Allows pictures prepared in Superpaint and saved in separate PICT files to be presented as stimuli.
Pasteboards (icon : yellowish page of text with red A in lower right) : This event is useful when several elements (pictures, paragraphs, sounds, or words) are to appear at the same time in different places or colors on the screen. The whole pasteboard is presented all at once.
Input (icon : key w/ finger pressing) : This event accepts single key responses from subjects.
Key Sequence (icon : keyboard with hands) : This event allows a subject to type characters into the computer during a trial. Usually this is in response to a paragraph that prompts for some input. The typed characters are displayed on the screen.
Sound Event (icon : speaker horn) : This event present sounds saved in files created in SoundEditTM.
Button Box Output (icon : 3 button box) : This event takes output from the button box.
§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§§
Section 3. General Stages when Developing Your Script
Outlined below are a list of discrete steps to help you in developing the script for your experiment.
*IMPORTANT TIPS
Remember, it is ALWAYS a good idea to save frequently in case the computer crashes. I have also found it extremely useful to use different names as you add elements to your script keeping track of different versions with a number so that the highest number is the most recent version (e.g., LexDec Script 5). This is useful because later changes may introduce an error in the program and it would be easier to return to an earlier version than to remember what all of your most recent changes were.
Starting from scratch
Once inside Psyscope, go to the Experiment menu and select ³New Experiment.² Provide a name for the experiment when prompted. The red Psyscope icon should be present in the Design window.
There are some small, yet nontrivial, steps to take to make your experiment run properly. These have to do with specifying
what sort of timing device you will use
what sort of input subjects will be making and
what will be saved in the data files
Under the Experiment menu select Input devices. Check the boxes next to every type of response you will have subjects make during the experiment.
Next, select Timing Devices under the same Experiment menu. If you will be using the button box to collect responses, use the button box timer. Otherwise, use the Macintosh built-in timer.
Lastly, under the Experiment menu select Data Output. Here, you will need to indicate that a response label will be saved with each RT. Do this by checking the box next to response label. We use the response label to indicate whether the subjects response for a particular trial was correct, an error, or too long (timeout).
Define the global structure of the study
Typically, experiments will be comprised of three templates :
- one to issue task instructions to the subject
- one to define the trials in the experiment
- one to inform the subject of the end of the study
Add these templates to the Design window and link them to the Experiment icon.
Define lists
Next, add and define the fields of any internal or external lists you will need when assigning the attributes of events in a trial.
The Trial Template
Here is the heart of the experiment. While you may add events to the Trial Template from the Design Window, it is best to open the window for the Trial Template and add events from here. This way you can see the temporal relation between events as you create the trial.
Because you have already defined any lists you may need, each time you add an event you can define all necessary stimulus attributes (those that are fixed and those that change from trial to trial).
Scoring Responses, Saving RTs to Data Files and Other Sorts of Condition-Action Pairs
Each of these operations is accomplished in the Condition-Action window. You can get there by going to the stimulus attributes window for an event (usually the event the subject is to respond to) and clicking the radio button so you can view the Event Attributes. The Condition-Action window can be viewed when you change the Actions attribute from ³Default² to ³Set to":
Saving Unique Datafiles for Individual Subjects
You may want to have the program prompt you for certain subject demographics to save in the data file. This can be achieved by asking to see ³Subject Info² from the Design window. There is a small bar above the drawing area of the Design window with the profile of a face. Clicking here will open a window that shows you what information will currently be requested when running the experiment.
In this window is also a small box that says ³Data File.² Clicking in this box will open a window that allows you to instruct the computer to generate a unique datafile name for each subject you test.
Final checks
Are the templates and events within templates presented in the proper sequence?
Are fields in a list linked to the proper attribute?
Are the appropriate number of total trials presented for the study?
Are items in the lists selected in an appropriate sequence (sequential or random)?
Will subjects need a rest during the session? (Done by opening attributes in the Experiment icon window.)
Does the datafile look ok? Is all relevant information there? (Can check this by running a datafile through PsySquash.)
Written by: Penny Yee, Hamilton College
Modified by: Tiffany Mattson, Hamilton College
Last Modified: 6/4/96