This experiment is a replica of a previous study carried out by Shepard and Cooper (1973). The study aims to look at whether the theory of mental rotation is reliable and valid.

University Degree Historical and Philosophical studies

Mental rotation and the effects of angular orientation on images presented as normal or mirror reversed

Abstract

This experiment is a replica of a previous study carried out by Shepard and Cooper (1973). The study aims to look at whether the theory of mental rotation is reliable and valid. Participants were required to look at a number of images, and asked to decide whether the image was a normal or mirror reversed version of specific characters. They were to answer by pressing a button which related to either left or right facing. The time taken to respond was calculated and was theorised that the response time would be greater proportionately to the degree of orientation away from the vertical, or upright, position of the normal character. It was concluded that the response time was in fact greater the further away from the norm the character was presented.

Introduction

Mental rotation is the ‘imagined turning of a form or shape from one orientation in space to another’ (Colman, 2006. p 453). Mental representations and mental rotation have been part of a psychological debate for a long time. Some psychologist (e.g. Watson, 1913; Pylyshyn, 1973) argue that mental images are not important, and that mental representations are language based and similar to the things they represent. However another view, put forward by Kosslyn (1980), suggests that mental images are a significant part of mental representations, and that we have the ability to ‘generate a surface representation which is active while we carry out a task’ (Collins et al, 1994). There is further evidence to support this theory. For example, when a person is asked to imagine a dog beside a house, they often know that the size difference between the two objects are of a different size; suggesting that humans do have a ‘bank’ of stored mental images. Rock (1973) (as cited in Corbalis et al, 1978) states that ‘an observer doesn’t recognise a shape until he or she has assigned a “top” and “bottom axis”, this further helps to understand the significance of mental rotation.

As shown by Shepard & Metzler (1971), participants are able to mentally rotate an image, so as to identify if two images are the same or different. In the experiment, participants were shown a series of images. These images were of two 3-dimensional images, ‘constructed by placing ten cubical blocks face to face’ (Colman, 2006. p 453) to form a structure, presented at angles to each other (rotated, but not mirror reversed). Participants were asked to state whether the images were of the same object or whether they were different. They concluded that is was necessary to rotate one of the images, in ones mind, so that one image can mentally be placed over the other to identify a match between the two. The results of this experiment also show that identification became harder the further the orientation from the normal orientation. The difference in the response time, suggests that participants were mentally rotating the image to come to their decision.

This is a preview of the whole essay

The Shepard & Cooper experiment (1973) designed an experiment to test the theory of mental rotation. In the experiment participants were asked to simply name which alphanumerical character was presented to them as quickly as possible, with as few errors as possible. The alphanumeric characters were presented at different degrees of orientation, and were either normal or backwards facing. There were two conditions in the study; Information and no information. In the information condition all participants were told which character would appear and what degree it would be presented at, the reason for this was that in order to mentally rotate an image, participants needed to know ‘both the identity and angular orientation of each character, for otherwise they could not have known where the upright was’ (Corbalis et al, 1978). In the no information condition, participants were not told which character was to be expected. The results were much the same the Shepard & Metzler experiment, where subjects took longer to identify a character if it was at a greater angular orientation. This experiment is a direct replica of the ‘no information’ experiment mentioned above.

In series of experiments carried out by Corbalis et al (1978), participants were asked to identify which character was presented while it was at different spatial orientations (different angles from the norm). Participants were expected to state whether the character was facing the correct or mirror revered way. Experiment 2 is the closest in relation to our experiment, as each participant was given a different character to identify, and asked to press one button if ‘their’ character appeared and another if any other character appeared. The character was shown facing both normal and backwards, and results show that participants identified their given character quicker if it was the normal way. However the participant, who was assigned the letter J, was quickest at identifying his character when compared with the other 5 participants. The rest of the participants had a delayed response time when identifying their assigned characters. This suggests that mental rotation was taking place during the response time delay.

All previous studies (Shepard & Metzler, 1971; Shepard & Cooper, 1973; Corbalis et al, 1978) propose that mental rotation takes time to complete. The results from these studies prove that the further the differing orientation from the norm or if a character is mirror reversed, the more time the participants will take to respond. This experiment is aimed to provide further information to mental rotation, and it is proposed that the outcome will be the same as earlier studies, in that the further orientation from the norm, the larger the response time will be.

Method

Design

Repeated measures experiment with a within participants design. The independent variable was the presented orientation of the character. It had six levels, one for each of the orientation degree. For example; condition one would be with the character at an orientation of 0º, and condition two with the character at a 60º angle, and so on. The dependant variable was response time, measured in milliseconds.

Participants

All participants were volunteers from an undergraduate psychology lecture. The group of participants consisted of 34 people; 6 males and 28 females. They were aged between 18 and 24 (average age of 18.85). All 34 participants took part in each stage of the experiment.

Apparatus/materials

The materials used were simply a computer and relevant computer program.

Procedure

Participants were asked to take a seat in front of a computer screens and place one finger over the letter ‘m’ and another over the letter ‘z’ on the keyboard. They were asked to complete a trial experiment first, so that the instructions were clear; this was called the practise trial. In the actual experiment there were 12 blocks of 12 trials, making a total of 144 experimental trials. There were six test characters, 3 letters and 3 numbers (G, J, R, 2, 5, and 7). There were six different degrees of orientation, ranging from 0 to 300 (0º, 60º, 120º, 180º, 240º, and 300º). Each character was presented in a randomised order, and appeared at each orientation twice, once facing forward and once as a mirror reversed image. The characters appeared on the screen and stayed until a response was made, followed by a 500ms gap before participants were presented with the next trial character. Participants were expected to respond by pressing a button which corresponds to which way the character was facing, ‘m’ being is the character was in the facing the normal way, and ‘z’ if it was presented as a mirror image. See appendix A for full instructions given to each participant.

Results

Table1 shows the mean response time calculated from the raw data (appendix B). Also shows how far the reaction time deviates from the norm (std. dev.). In the table, it is clear that participants took the longest (2307 ms) when the character was at 180º (upside down).

Table 1: The mean response time and standard deviation for each different orientation

The graph is a visual representation of the mean response time (RT) for each different angular orientation, as shown in table 1. It is what was expected, as the result is a reflection of previous studies carried out on the same topic. The graph shows a proportionate difference dependent on the orientation.

In some cases the raw data is to be ‘cleaned’, by removing any participant’s results, if the accuracy was < 59% (see appendix B for accuracy rating), however this did not need occur in this experiment as no participant scored below 59%.

Figure 1: The mean time taken to differentiate if the image was a normal or mirror reversed version, depending on orientation.

Discussion

In the experiment carried out here, participants were asked to identify if a character was a normal or mirror image version of itself. All characters used were easily recognisable. For example, the letter ‘R’ would still have the same defining features whether it was normal backwards; a straight vertical line, a semi circle up top and a diagonal line pointing down to the right. This is beneficial to participants, but the response time still seems to suggest that mental rotation is occurring. If a letter is presented upside down or at 180º then it becomes difficult to identify whether it is normal or backwards. The fact that participants took longest to say which way a character was facing at this orientation, provides substantial evidence to support the proposed hypothesis, in that mental rotation does take a considerable amount of time to process. However even the longest response time does not show a considerable delay in identification, this suggests that the characters were effortlessly recognisable (as mentioned previously). The pattern of the mean response times is as would be expected. For 0º the identification was quick at 1257 ms. The response time rises slightly from one orientation to the next, this trend continues as characters as displayed at greater degrees away from the upright position. It is further supported as participants start to respond more rapidly as the character begins to come full circle and proceed to the normal upright position.

The results given here agree with previous findings by psychologists introduced earlier. The results disagree with the theories put forward by Watson and Pylyshyn, as they support the idea of mental rotation. The graph is almost an exact replica of that seen in the Shepard & Cooper study, suggesting that their theory was correct, and that the experimental results can be generalised for a larger population. Evidence can be seen from both sets of results to conclude that both studies are reliable and consistent.

For future studies, I believe it would be useful to use a larger group of participants to expand the sample. Male and female differences should also be looked into and maybe even left/right handed participants to see if this effects results. More recent research for further reading can be found in studies such as Tarr and Pinker (1989), Takano (1989) and Cohen & Kubovy (1993).

References

Collins, A. F., Levy, P., Morris, P. E., & Smyth, M. M. (1994). Cognition in Action. London: Psychology Press/Erlbaum.

Colman, A. M. (2006). A Dictionary of Psychology (2nd ed.) Oxford: Oxford University Press.

Corbalis, M. C., Zbrodoff, N. J., Shetzer, L. F., & Butler, P. B. (1978). Decisions about identity and orientation of rotated letters and digits. Memory and Cognition, 6 (2), 98-107.

Eysenck, M. W., & Keane, M. T. (2000). Cognitive Psychology: A Student’s Handbook. (4th ed.) East Sussex: Psychology Press Ltd.

Kosslyn, S. M., & Thompson, W. L. (2003). When is early visual cortex activated during visual mental imagery? Psychology Bulletin 129 (5), pp. 732-746

Shepard, R. N., & Metzler, J. (1971). Mental Rotation of Three-Dimensional Objects. Science 171, pp. 701-703.

Appendices

A – Instructions for Participants

B – Raw Data

Appendix A: Instructions for Participants

Instructions

During the experiment you will be presented with a series of trials. On each trial you will see one of the characters G, J, R, 2, 5 or 7. The character will be either the right way round or back to front. Your task is to indicate which way round the character is facing, by pressing one of two keys. If the character faces the right way round, press ‘M’, but if the character is mirror reversed press ‘Z’.

The characters will be presented in six different orientations; in each orientation, a character can be either the right way round, or mirror reversed. They will appear at angles of 0 degrees (right way up), 60 degrees, 120 degrees, 180 degrees (upside down), 240 degrees and 300 degrees. All angles are measured in an anti-clockwise direction.

You should respond as quickly as you can, but try to avoid making errors. The trials are arranged in blocks of 12, with each block being started by pressing the space bar. Once you have made a response, there is an inter-trial interval of 500 milliseconds before the next stimulus appears.

Summarising, the steps are as follows: