The first sample experiment
Object
To design a Psyscope project that will play five sound files, use button box input to record responses and to cue the next stimulus, and record the responses in a file.
Basics.
The first part of any Psyscope experiment is to define your inputs and outputs. In this sample, the the only user input is the button box clicks to indicate that the sound file has played. The sounds are outputs, as are the data files. The sound files must be prepared, and named as described above. A little tip for experiment construction. You can construct rough sound files on the small Macintosh in the Microcomputer room (B10) using SoundEdit. These files do not give good enough quality, or precise enough timing for use in the final script, but it is an easy way to prepare sound files for preliminary work, if you can face the embarassment of sitting in the lab talking into a microphone, with everyone sitting round laughing. Anyway, I did it, so you can. If you need further details on this, ask Eddie.
When you begin, you start as above, selecting from the design menu on the tool bar, "new experiment". Psyscope will then prompt you for a name for the experiment, so choose something fairly descriptive - not "My Experiment" or "New Experiment". The design window will then open up looking as shown below:-
The first thing you should do, is to define the input and output devices for the experiment. This is done by clicking on the experiment icon at the top of the design window. This will open up the experiment information window, which should look like this:-
Attributes selection
Experimental Attributes
For this experiment, the attributes needed to be set are the Data file to give your output, the Data Info, which will specify how the data file is written, and what is written, the Input Devices - the button box must be specified, the timer must be set to button box. Other devices in the list may be set as required, but read section 5.8.3 of the manual (only about 10 pages:-).
After you have done this, you are ready to start arranging your experiment.
First, you need to have a template. A template is a handy method for arranging your experiment into small units. Instead of having everything on the design window, you can divide your experiment into modules, and each module is independent. This means that you can design each step of the experiment, and test it separately from the whole experiment. To do this, you drag a template object from the object menu, and drop it on the design window. A dialogue box will open, asking for the name of the template - again it is best if you choose a descriptive name for the template, rather than "Template 1". The arrow pointer will change shape into the link pointer, so that the template can be linked with the experiment. The screen will now look like this:-
You now have a template linked to the experiment. Now you will have to link events to the template, which will be the course of the experiment. A useful tip - it is harder than it should be to insert events into an experiment which will occur before events already included. In other words, if you are going to have an instruction sheet (a paragraph event), it is best to place that in the experiment now. Since this is a very simple experiment which requires no real user interaction, this will be ignored. To start placing events on the template, double click on the template icon, which will open up a new window:- This window is where you develop the timeline for the experiment, and place events in the experiment. This is simply done by clicking on the event type in the palette on the lefthand side of the window and dragging the event into the area on the left. This should leave you with the following on your screen:-
you will see that a line has appeared (normally pink) which is the timeline Again, give the event a fairly descriptive name, but a point to watch for is that Psyscope has a fairly long list of words that are reserved for use by the program (how or why, for example). The timeline allows you to see precisely how long the stimulus will occur for, and also to see events in proportion to each other and the whole template. You now need to set the attributes for this event, because currently Psyscope only knows that it is a sound stimulus. There are two types of attribute for each event, the stimulus attribute, which controls the 'appearance' of the stimulus, and the event attribute which controls the way in which Psycope handles the event. To access the screens which control the attributes, you must double click on the event. This will open up the attributes window as depicted on the right. This is showing the stimulus attributes. The attributes shown in the list vary according to the event type, the ones shown are for a sound event, for a full listing of the types, refer to pages 180seq in the Psycope manual. For this experiment, it is only necessary to specify the sound file to use, by clicking on the file pane, and selecting 'Set to' and entering the filename. The other button for Event Attributes has a more significant role, so you will need to click on the Event Attribs button which will open the event attributes window.
This picture below actually shows the window after some work has been done on it, but no matter. First, the duration. If you are using a stimulus that has no fixed time, e.g. a text stimulus, this show how long the event will remain. There are several options under the button, refer to the main manual for details, in general, however, you will need to use 'set to' and set it to 'sound' to allow the sound to play fully. You now need to specify the actions. It is advisable to consider the experiment as a whole - you are wanting to record responses, so moving onto the next stimulus immediately is not required - this is where you set up the use of the button box. First click on where it indicates Set To:, which activates a menu. From this you need to select Set To: which will activate a new window, which will give the conditions for duration. Select button box, so that the square at the end of the line has an X in, as in the picture below.
This means that the timeline will not continue, until it has received a button box stimulus. If you double click on the words 'button box' another dialogue box will open to give you choices as to which button box button is necessary to terminate the stimulus. In this experiment, because of it's simplicity, this has little effect, but in future experiments the choices will have to be limited. In this experiment, the release box will have to be unchecked, otherwise two button box events are recorded, and the next stimulus disappears very rapidly. At present, however, Psyscope does not know that you want to record the button box events, so you must specify this.
On the event attribute dialogue screen, (shown above) the second option is actions. This determines what actions Psyscope takes on receiving an input, so this needs to be set. Click on the bar with an arrow on it, and select Set To:. A new dialogue opens, shown on the right This screen already has the conditions set, but when first seen, the box is empty. you will need to click on new, to add a Condition. This will bring up the label at the bottom marked Default, you will need to click on this and set it to Set To: , which will open up a familiar looking dialogue box (identical to the one above), from which you should select Button box. This in it's turn will activate the Actions pull down menu, which is a large dialogue, but at the moment the only one we are interested in is the RT action. This will record the action in the data file. All RT events are recorded in the data file.
The procedure for this stimulus needs to be repeated for each individual stimulus (but later we will look at a way that you can specify on stimulus/condition/attribute sequence for a group of events). The picture below shows the finished experiment:-
When this experiment is run now, it will record the events as designated by the RT actions in the log file e.g.:-
******************************************************************
PsyScope 1.1b7 started: 1/17/97 12:00:13
Script file: Training1 Script
Run on: Power Macintosh
RunNumber: 1
Input devices active: BBox
Trial Condition Time state
1 -1400 1
1 -1257 1
1 -1126 1
1 -989 1
1 -844 1
***********************************************************
but this is not too useful, because Psyscope doesn't line the columns up with the headers, so you tend to end up squinting to work out what is happening. A far better idea is to import the data file into an analysis program (e.g. Excel, Anova, SPSS, Systat) by using the text import feature - in Excel, just open the file and say yes to the prompts, and you will develop a spreadsheet that looks like this:-
**************************************************************
PsyScope 1.1b7 started:
2/12/97 10:03:38
Script file: Training1 Script
Run on: Power Macintosh
RunNumber: 1
Input devices active: Key
Trial Conditio Time state key
n
N/A
N/A
N/A
N/A
N/A
**************************************************************
PsyScope 1.1b7 started:
2/12/97 10:05:40
Script file: Training1 Script
Run on: Power Macintosh
RunNumber: 1
Input devices active: BBox Key
Trial Conditio Time state key
n
1 550 1 N/A
1 749 2 N/A
1 681 4 N/A
1 733 1 N/A
1 765 4 N/A
**************************************************************
PsyScope 1.1b7 started:
2/12/97 10:05:52
Script file: Training1 Script
Run on: Power Macintosh
RunNumber: 1
Input devices active: BBox Key
Trial Conditio Time state key
n
1 657 2 N/A
1 570 1 N/A
1 599 4 N/A
1 638 2 N/A
1 491 1 N/A
******************************************************************
PsyScope 1.1b7 started:
2/12/97 10:06:04
Run on: Power Macintosh
RunNumber: 1
Input devices active: BBox Key
Trial Conditio Time state
n
1 545 1
1 547 4
1 447 2
1 287 1
1 299 4
The points to note are as follows:-
1) The header for each section contains information referring to version of the program, time and date of the experiment, which script file is being used, which machine, the run number, and the active input devices. It is worth noting here, the elementary mistake that was made in constructing the data file - the Run Number is not being calculated for each run of the experiment. This should be set to be calculated at the start of each run, from Subject Information in the tool bar. The header information can be excluded from the data file by bringing up the data file attributes from the experiment attributes, selecting Data Info, and on that screen selecting "Don't include header information". This will make subsequent analysis of the data easier.
2) The data is arranged in such a fashion as to facilitate data extraction. If you look at the state column, you will see the numbers in the column - these refer to which button box button was pressed, so it is a simple matter of inserting an extra column with the "correct" answer. In this case there were no "correct" responses, so this is pointless. However if I had specified push the red button (1) for words of less than five letters, push the green button (4) for five letter of greater, the results should be 1, 1, 4, 1, 1. Any deviation from this shows an error. Also, if key is the response device, the key column would show which key was pressed. You should also note as well that the data file that is the output from the experiment is in tab-delimited format. This is important to know because every program you use for data analysis will ask you whether it is a comma-delimited, space-delimited or tab-delimited. That is about all there is for the first Psyscope experiment. I know that it seems unnecessarily verbose for what is in point of fact a very simple experiment that could more easily be accomplished with other utilities, but this simple experiment introduced all of the basic concepts that occur in Psyscope.
All material Copyright Eddie Dubourg/Department of Linguistics, University of Edinburgh