Method
Design
The independent variable was the ‘type of object’ displayed (with two levels, SS and SD objects), and the major dependant variable was the average correct ‘recognition time’ (in milliseconds (ms)). On completion of the experiment an additional dependant variable was ‘error rates’ for SS and SD objects. The design utilised to test the hypothesis and analyse the data was a within subjects paired sample, related t-test where outliers (time taken over 1000 ms) were removed for more accurate results.
Participants
Each student carrying out this study selected 2 participants, the total number of participants was 272 mixed gender associates, all of which were naïve to the psychology of perception and agreed to take part on a voluntary basis.
Apparatus
The running of this experiment was conducted using a computer software programme named SUPERLAB; this programme is designed to run simple psychology experiments. Using this software each participant was shown, on screen, 32 randomised line drawings (see appendix 1) of SS and SD objects (16 of each condition, one at a time) and their response times were calculated. A response table was used by the experimenter to record correct and incorrect responses for further analysis (see appendix 2) and a memory stick was used to save the results.
Procedure
An experimental instructions sheet was read (see appendix 3). A memory stick was inserted into the computer in order to save the collected data and the SUPERLAB programme was started up (see appendix 4). Each participant was read a set of instructions (see appendix 5) before the experiment explaining the procedure they were expected to follow. When each subject was ready they were given a small practice block to make certain they understood the task. On satisfactory completion of the practice run subjects were asked to press the space bar and wait for the display of the ‘fixation cross’. This was the focus point of which participants were asked to gaze upon and wait for the presentation of each line drawing. As soon as a line drawing appeared the computer timer started, each subject was instructed to press the ‘Y’ key the moment they recognised the object (the timer stopped and the screen went blank) and outwardly name the object. Each name given was noted and recorded in a response table for further analysis (see appendix 2). Again when ready, the participant pressed the space bar for the next object and this course of action was repeated for all 32 experimental objects (randomly presented 16 SS and 16 SD). Each participant was asked if there were any line drawings they found particularly difficult to recognise and name. The data was saved (see appendix 6) as an excel document where it was put in chronological order and all naming errors were removed to calculate the average recognition time for SS and SD objects. All naming errors were reported at the bottom of the excel document. The completed document was submitted for collation with the remainder of the years’ results (see appendix 7).
Results
Outliers were removed for more accurate results. The mean and standard deviations for both recognition times and errors for SS and SD objects are shown in table 1.
Table 1. Mean correct response times and mean error rates, and standard deviations (StDev) for SS and SD objects.
Firstly, only recognition times with the correct response were subjected to analysis. Secondly a statistical analysis was carried out on errors. A paired sample t-test was utilised for both analyses. The factor was object type, with two levels, SS and SD objects. In a one-tailed, paired sample t-test for SS and SD times, the correct recognition times for SS objects were found to be significantly longer than for SD objects ( t(217) = 0.373, p<0.05). In a two-tailed paired sample t-test for SS and SD errors, it was found there was a significant difference in errors between the two different types of objects (t(217) = 1.650, p<0.05) with more errors for SD objects. Results of errors are shown in table 2a and 2b.
Table 2a. SS Frequency Table of Errors
Table 2b. SD Frequency Table of Errors
The following data is the descriptive statistics created by the use of SPSS for windows.
Discussion
According to the results of this study there was a significant difference between the recognition times of SS and SD objects. These results supported the hypothesis stating there will be a significant difference in that recognition time for a correct response to SS objects will be longer than that for SD objects. This in turn supports the work of Humphreys et al’s. (1988) cascade modal maintaining natural objects appear, to the eye, to have more parts in common and be more visually similar than artefact items therefore natural objects will take longer to identify.
When participants were asked if any objects presented particularly difficult to recognise and name it was noted that one participant expressed they later thought the spider, which they named correctly, may have been a jenny long legs and the duck, which they named (seagull) incorrectly, may have been a duck. The results show recognition times for SS objects as a higher measurement of variation (Std. Deviation = 168.236) than that for SD objects (Std. Deviation = 157.791) suggesting that time for SS recognition was more varied and spread out compared to that of SD objects. These findings are consistent with Gerlach et al. (1999), in that potentially, competition arising within the structural description system for recognising natural objects makes the process of identification longer for SS compared to SD objects.
In concurrence with Collins and Quillian’s (1969; cited in Hayes, 2000) judgment, the process of categorisation and elimination when identifying SS objects is most probably the cause of recognition time being longer for SS than SD objects. Following on from this, in relation to the results of error rates it appears reasonable to assume that an explanation for SD having more than SS objects could be that the time delay for recognising SS objects may aid in arriving at a more accurate identification.
Considering the view of Coren, et al. (1994) it seems fair to say that semantic knowledge, or lack of it, may influence or inhibit our abilities in object recognition and could affect results. The universality of SS and SD objects is not clear which could introduce conflicting opinions on the correct names of particular objects. For example, what seemed to be the most controversial picture, the penknife, had so many of what could only be agreed as correct answers (see frequency table in results section) in naming the picture but termed wrong according to the response sheet. Distinguishing between the wrong name because either that is what the subject believed they saw or what they believed the item should be called or perhaps as others may argue, it is a semantic error, seems difficult and therefore, in the frequency table columns ‘correct’ and ‘semantic’ are joined together.
Many other potential factors that may have hindered the results are general subject errors such as making mistakes like pressing the ‘Y’ key before truly recognising the object, it is thought that this may be the cause of the outlier displayed in the box plot of SD errors (see descriptive statistics in results) as the recognition time was spuriously short. Perhaps it should have been stressed to the participants that specific names should be given and not superordinate names e.g., tool instead of hammer or insect instead of spider.
Rossion, B., & Pourtois, G. (2001) claimed that the addition of texture and colour significantly improves naming of objects. Based on this claim it would be interesting to repeat the present study with the introduction of colour and texture to the pictures for further investigation.
In conclusion, and in concurrence with Humphreys et al’s. (1988) cascade model, this particular study has been successful in providing further support that structurally similar objects are more difficult and take longer to recognise and name than structurally distinct objects.
Reference
Coren, S., Ward, L. M. & Enns, J. T. (1994) Sensation and Perception (4th ed.). London: Harcourt Brace & Company.
Gerlach, C., Law, I., Gade, A. & Paulson, O. B. (1999). Perceptual differentiation and category effects in normal object recognition: A PET study. Brain, 122, 2159-2170
Hayes, N. (2000). Foundations of Psychology (3rd ed.). UK: Thomson Learning.
Humphreys, G. W., Riddoch, M. J. & Quinlan, P. T. (1988). Cascade Processes in Picture Identification. Cognitive Neuropsychology, 5, 67-103.
Humphreys, G. W. & Forde, E. M. E. (2000). Hierarchies, similarity and interactivity in object recognition: On the multiplicity of ‘category-specific’ deficits in neuropsychological populations. Behavioural and Brain Sciences, 24, 453–509.
Rossion, B. & Pourtois, G. (2003). Revisiting Snodgrass and Vanderwart’s object pictorial set: the role of surface detail in basic-level object recognition. Perception, 33, 217–236.
Snodgrass, J. G. & Vanderwart, M. (1980). A standardized set of 260 pictures: norms for name agreement, image agreement, familiarity and visual complexity. Journal of Experimental Psychology: Human Learning and Memory, 6, 174–215.
Appendix 2
Use this table to check the answers of your subject on the trials.
Appendix 3
Your (experimenter) instructions:
Get the SUPERLAB program up and running and load the practice trials (B1) followed by the randomised SS and SD block (B2) into the program. Read the instructions to the subject and answer any questions they might have. Run the practice trials - try to get the subject to place their finger over the "Y" key before they press the space bar to start the trial. Agree on a starting signal with your subject, e.g. you could say either "start" or "begin" etc after which the subject should press the space bar. If you feel the subject is not following the instructions i.e. pressing the “Y” on recognition and then immediately stating out loud the name of the object, then try to point out what they are doing wrong. At the end of the practice trials you will see a screen saying "Start of Experiment”. Ask your subject once more if they have any questions. After which you can continue and run through the 32 SS and SD pictures. Thank your subject for participating and try to answer any questions they may have about the experiment. You might also ask them if there were any pictures they had problems with and if so, why those pictures were problematic. You should note down any of the subject’s responses.
Appendix 4
Using SUPERLAB
Superlab is available on all the F block lab computers (F333 & F337). However, as there are sometimes changes in the way the computer lab is serviced and administered by the computing dept the files required for particular experiments may be located in a different location to the one stated in the handout. Try using the address given in the textbox below. If you cannot find the correct folder on the machine hard drive (C:) or any of the network drives you can access, then check the Perception blackboard site for any updated instructions.
Boot up the computer and once Windows has started, put a FLOPPY DISK into the a: drive. Select the START button at the bottom left of the screen. Go to PROGRAMS and then select SUPERLAB PRO for Windows. The program should start and you will see a Superlab window which is divided into three sections
The window should now have some new information in it. Go to the BLOCKS column in the window (left hand column). Using the mouse to highlight B1 B2 (ignore B3). (click on B1, hold down the SHIFT button on the keyboard and then click on B2). Both rows should now be highlighted in blue.
Select EXPERIMENT from the menu at the top of the window. Click on RUN.
In the new window click in the INPUT SUBJECTS NAME box. Type in a name /initials or a unique number for the subject you are testing.
Ensure that both “Run collected data” and “Run the 2 selected blocks only” are both ticked.
Click on RUN
In the “save collected data file” window, provide a file name for the data to be stored in (remember to make a note of what you call it). Select the Floppy (a:) drive for the destination to save to. NB you will need to put a floppy disk into the machine to save your data onto.
Click on OK to start the experiment.
Appendix 5
EXPERIMENTERS SHEET
Name: ______________________________________
Instructions for subjects (to be read aloud to the subject before they begin)
"Thank you for agreeing to participate in our experiment. The task you will be asked to complete is very straightforward and should only take you a few moments to complete. You will be presented with a series of 32 line drawings of common objects. The pictures will be shown one at a time. At the start of each trial you will be asked to place a finger over the "Y" key on the keyboard and then when you are ready press the “space-bar” on the computer in front of you. The screen should now show a small cross at the centre of the screen. Keep your gaze on the cross and after a few seconds it will disappear and a line drawing of an object will appear. Your task is to press the "Y" key as soon as you recognise the object. At this point the screen will go blank and you must immediately say out loud the name of the object you have recognised. It is very important that you try to remember to press the button first than name the object. If you name the object first, before pressing the ”Y” key. Or if you press the “Y” key and then hesitate before naming the object, we cannot use the data for that trial in our analysis. There will be 32 pictures to name during the experiment. To help you get used to the task we will have 4 practice trials to begin with. If you have any questions please ask me now."
Appendix 6
Sorting & using the data from the saved file:
Having run a subject, you will now have a file {yourname} saved on your floppy disk.
Open up Excel on the computer and load that file into Excel (you may have to type in the actual file name or use “*.dat” to find the file on the floppy disc as Excel does not recognise the “.dat” extension to file names. Obviously the file was not written in Excel, so Excel will offer to convert it into an Excel file. Accept the changes/options that Excel provides (click on “next”) and the file should now open up in Excel.
As the experiment used a randomised presentation of the SS and SD pictures, we will now un-randomise them using the sort function in Excel. The data listed under the “value” heading in your file are numbered 01-32. Values 01 to 16 correspond to SS01 to SS16, while Values 17-32 correspond with SD01 to SD16. By sorting all the data in the file into order using the Value numbers we will have all the SS trials listed above all the SD trials. Select the first number underneath the words “Trial Name”, hold the left mouse button down drag the mouse to the bottom right hand number to highlight all the data. Then click on the sort button (an A above a Z with an arrow pointing downwards). The resulting file should now be sorted by the Trial Name variable. I.e. the numbers 1 to 32 should be in order down this column.
The next task is to remove any “errors” from our data. Check your experimenter’s sheet to find any trial that you noted as an error. Then find the corresponding Trial Name in the data file and delete all the information in that row. Do this for all the trials you have noted as errors. E.g. in the 10th presentation the subject said “bird” instead of the correct “owl”. From the experimenters sheet we know that the OWL is coded as SS08, we can therefore go to the Trial Name column find number 8 and delete all that row of information. Similarly, you suspect that the subject said the word “sock” well before pressing the “Y” key. From the experimenters sheet we see that sock is coded as SD10. As SD01 is actually Trial Name 17, we can count out the 10th trial for the SD’s with Trial Name 17 being number one, the 10th SD trial is therefore Trial Name 26. This row can also be removed.
Submission Layout:
YOUR NAME
Details of your results (give the time in Milliseconds)
SUBJECT A SUBJECT B
Time for SS Time for SS
Errors for SS Errors for SS
Time for SD Time for SD
Errors for SD Errors for SD
A list of any items wrongly identified:
e.g. ‘s
SS05 (spider) given as “Insect”
SS02 (duck) given as “Goose”