The present lab is a useful exercise for several reasons: 1) it uses linear regression to evaluate the relationship between an independent and dependent variable 2) it examines the speed-accuracy trade-off and 3) it tests the "internal images" hypothesis.

You are investigating whether people can form mental images that closely resemble the actual perception in order to make decisions. Cooper and Shepard (1978) tested this hypothesis by presenting their subjects with letter stimuli (such as a capital 'F') in a variety of orientations from the vertical (e.g., 60, 90, 120 degrees). Subjects were required to decide whether or not the stimulus was a normal 'F' or a mirror image and the time it took to make this decision was measured. Cooper and Shepard (1978) found that reaction time (RT) varied as a function of the angle of rotation. Specifically, this relation was a positive linear one so that the larger the angle of rotation (i.e., increased orientation from the vertical), the longer it took for subjects to make the decision. They concluded that subjects were mentally rotating the image of the stimuli in their mind so that they could more easily make the comparison.

The study we will be doing is similar to that of Cooper and Shepard's (1978) except geometric stimuli will be used. The experiment is a within-subjects design where all subjects will be tested on all levels of the independent variable, the angle of rotation (with five levels, 0, 45, 90, 135 and 180 degrees). There are two dependent variables in this study. The main measure is reaction time in milliseconds (msecs). The second is an accuracy measure, specifically the number of errors made for each angle.

As previously mentioned, this lab is also concerned with the speed-accuracy tradeoff. This is an ever-present problem in RT studies. Basically this tradeoff defines an inverse relationship between speed and accuracy such that concentrating on speed reduces accuracy and vice versa. If we want to make inferences about changing RT's, we need to ensure that the level of accuracy is not changing in the opposite direction or our assumptions may be incorrect (i.e., may be a result of changing accuracy levels than changes in the independent variable).

We will use simple regression analysis to see if RT is a linear function of the angle of rotation. This function is described by the equation:

Y' = bx + a

where Y' is the predicted value of Y;

b is the slope of the regression line (or the change in the dependent variable associated with one unit change in the independent variable); and,

a is the Y-intercept (the value of the dependent variable when the independent variable is 0).

So, in a study like that of Cooper and Shepard's (1978), b would tell us how much time it would take to respond when the rotation of the stimuli is increased by one degree and a tells us the response time when the angle is 0 degrees. Therefore, we can infer that b indicates how long it takes to mentally rotate an image one degree and a indicates how long it takes to compare images and respond once they have been rotated.

We can also predict how long it takes to decide if the images are the same or not depending on the rotated angle of the images. For example, if the regression equation was found to be Y' = 10x + 1200, we can estimate that it would take 1650 milliseconds to respond when the angle was 45 degrees from the vertical orientation (i.e., 1650=10(45) + 1200). The same concept applies to using simple regression to analyze the number of errors.

Students from the Research Design and Analysis 2  classes will be used as participants
Materials

• Experimental Psychology: A Computerized Laboratory Course
• standard stimulus: a vertically-orientated geometric shapes presented on the left side of the computer screen in one of its two mirror image forms.
• test stimulus: the same figure as the standard will be presented simultaneously on the right side of the screen, rotated at one of 8 possible angles (0, 45, 90, 135, 180, 225, 270 and 315). The data collected on stimuli rotated at 225, 270 and 315, however, will not be analyzed in the present lab.

Double click on the experimental software icon
Once the main menu is displayed, select Mental Imagery (#7) and then Experiment 2, Mental Rotation.

You will then be provided with instructions. Throughout the experiment, you will have your left forefinger on the "S" key (same) and your right forefinger on the "D" key (different) and your thumb on the spacebar. At various times during the trials, you will be required in indicate your readiness by hitting the spacebar. In each trial, an * will appear; you should focus on this point. Shortly, the * will disappear and the two shapes will appear. The two shapes will either be the same or mirror reversals of each other. Your task will be to decide whether or not the two stimuli are the same or are mirror-reversals (and therefore different). If they are the same, press "S" and if they are different, press "D". So, if the shapes are the same version of the figure but oriented differently, then press "S" but, if they are mirror reversals, regardless of their orientation, press "D".

There will be 8 practice trials and 64 test trials. Respond as quickly and as accurately as possible. You will receive feedback on your judgements. If you take too long to respond, you will hear "beep-beep" and if you are incorrect, you will hear "burr".

You may note that in some of the test stimuli, the shape is slightly blurry compared to the test stimuli. This is a result of poor graphics and is not a difference we will be concerned with. The only difference relevant in this study is the one that indicates mirror reversal.

After the 64 test trials, you will be shown your data. RT is the median RT in msec. RECORD THIS INFORMATION IMMEDIATELY IN THE TABLE PROVIDED!! Do not record the data for the 14 computer-generated subjects. We will only be interested in the data from the students in this class.

Pass your data in to Jill by Friday March 12 (section A) and Monday Mar 15 (section B).  Penalty is loss of 5 marks off your lab report mark.
 
 
 
 

 Recall, that our main prediction is that reaction time will positively vary as a function of the angle the test stimuli is rotated. For our results we will do two regression analyses; one for each dependent variable. Also, report the means and sd RT's for each angle and provide two figures.

From the results you should be able to infer psychological processes, i.e., what would increased reaction time with larger angles suggest? Remember to also look at the errors as a function of angles. Does the data suggest a speed-accuracy tradeoff? In other words, do the number of errors vary inversely with reaction time. Expand using articles discussed in the introduction.

Lawrence Erlbaum & Associates Ltd.

     Cooper, L.A., & Shepard, R.N. (1978). Transformations on representations of objects in space. In E.C. Caterette

and M.P. Friedman (Eds.), Handbook of Perception (Vol.8): Perceptual Coding. New York: Academic.
 
     Kosslyn, S.M., Ball, T.M., & Reiser, B.J. (1978). Visual images preserve metric spatial information: Evidence

from studies of image scanning. Journal of Experimental Psychology: Human Perception and Performance, 4, 47-60.
    
     Shepard, R.N., & Metzler, A. (1971). Mental rotation of three-dimensional objects. Science, 171, 701-703.

The following articles are on reserve.  You need to find at least one article on your own,  not listed in references above or on reserve.  You should use a minimum of 3 relevant empirical articles in your introduction

Cooper, L.A., & Shepard, R.N. (1978). Transformations on representations of objects in space. In E.C. Caterette and M.P. Friedman (Eds.), Handbook of Perception (Vol.8): Perceptual Coding. New York: Academic.

Kosslyn, S.M. (1980).  Image and Mind.  Cambridge: Harvard University Press. (Both the book itself and ch.8 from the book are on reserve)

Pylyshyn, Z.W. (1979). The rate of "mental rotation" of images: A test of a holistic analogue hypothesis.  Memory & Cognition, 7 (1), 19-28.
 
 
 
 
 
 

• separate columns for  gender, age, angle of rotation, rt, and errors.  There will be a number of empty cells in the gender and age columns because of the type of design we are using.

• Open Statistics menu, select Regression and then Linear. This will open the Linear Regression dialog box.
• Highlight RT in the source list and click on the arrow pointing to the Dependent box so that RT appears there.
• Then, highlight angle and click on the arrow pointing to the Independent box so that angle appear there. Click OK or press <enter>

Do the same for errors except select Errors for the dependent variable.

You should save your output at this point:

To obtain means, sd's  choose Compare Means,  Means under the Statistics menu.  Move RT and error into the Dependent variable box and angle into the independent variable box.  Click on the option button and add minimum and maximum to the stats in the list (mean, standard deviation should already be present in the box on the right)

Finally, create two scatterplots:

• Under the Graphs menu, select Scatter and then select Simple from the bar graph definitions. Now click the Define button. Another dialog box will open.
• Highlight RT and click the arrow button pointing to the Y-axis box.
• Then, highlight angle and click the arrow pointing to the X-Axis box.
• Click OK or press<enter>

You will want a regression line in your figure. To do this:

• Hit the Edit Button. Then open the Chart menu and select Options.
• Under Fit Line, click on the small box besides Total, then click on the Fit Options button, select Linear Regression and click on the Continue button.
• Click OK or press<enter>. The regression line will appear.

Save and print your output to attach at the end of your lab report.
 

Name_____________________________

Section:   A  (M/W)         B (T/TH)

Age: _______

Sex:     M      F