Task Details

 

The Symmetry span task is a complex span partner to the matrix span task. The participant is shown a series of grid locations one-by-one form the 4x4 grid in the center of the screen. The participant must remember the grids and the order in which they appeared. Afer each grid is shown the participant is shown an 8x8 grid that has a number of grids filled black to form a pattern. The pattern will either be symmetrical along the vertical axis or it will not and the participant must make this judgement using the left/right arrow keys before the next grid will be shown.

Screenshots

 

Presentation Phase

Image shows a 
			screenshot of what the program looks like while delivering the presentation phase of 
			the task - the to-be-remembered locations
Image shows a 
			screenshot of what the program looks like while delivering the presentation phase of 
			the task - A symmetrical pattern
Image shows a 
			screenshot of what the program looks like while delivering the presentation phase of 
			the task - A non-symmetrical pattern

Recall Phase

Image shows a 
			screenshot of what the program looks like while delivering the recall phase of the 
			task

The XML

 

You can download the default XML here. In this section I will point out the customisation options at your disposal by modifying values in the XML.

Under task settings there is a section of code where we tell the application where to find the executable that handles the digit span task. The property name/value keys here are where we can set task-specific properties.

Storage Phase

This set of customisable properties relate to the storage component of the task, therefore they are the same as described for the simple span counterpart (matrix span. These properties are found in the bean with id "symm_span".

Number of trials

You can select how many trials of each span size you want. The properties to edit here are:

  • 'spanTwoTrials'
  • 'spanThreeTrials'
  • 'spanFourTrials'
  • 'spanFiveTrials'
  • 'spanSixTrials'
  • 'spanSevenTrials'
  • 'spanEightTrials'
  • 'spanNineTrials'
The default values for these properties are "3" for spans 2 through 7 and "0" for spans 8/9. Therefore as it stands the module will administer 3 trials at each list length up to 7 and none at list length 8 or 9. Simply change the integer values within the quotes for each span*Trials property to get the desired number of trials.

Sequential/Randomised trial order

This property decides if the trials are administered in sequential order or not with regards to the list lengths. If sequential then all list length 2 trials are given, followed by all list length 3 trials, and so forth. However, it is often the case that a researcher may wish to administer these trials in a random order so that the participant is not aware on any given trial how many digits they will have to store/recall. The property to edit is:

  • 'randomisedTrials'
The default value is "0" which is the code for sequential order. The alternative is "1" which is the code for random order.

Matrix Size

It is possible to select how many grids you would like in the matrix. The property responsible is:

  • gridDimension
Whatever number(n) is set as the value will decide how many grids are in the matrix (n x n). the default is 4 which creates a 4x4 matrix hence 16 grids.

In addition there is an option to control the actual size (on screen) of the matrix.

  • ScalingFactor
The application will get the size of the screen and it is being executed on (in pixels) and initially create the matrix to be as large as it can be while still fitting the screen. It then applies the scaling factor (treat as a percentage). So if you want the matrix to be half size then this value can be set to 50. The optimum value will differ depending on the equipment used and researcher preference.

Processing Phase

This set of customisable properties relate to the processing component of the task where the participants are shown a grid with a pattern filled and must make a judgement on whether it is symmetrical. These properties are found in the bean with id "symm_processing".

Symmetry grid size

What do you want the dimensions of the matrix to be that delivers the pattern? The default value is "8" which produces an 8x8 grid with a pattern filled.

  • GridDimension
The actual size of the grid in pixels is determined in the same way as for the storage component. Therefore there is also a scaling factor property for the processing component.
  • ScalingFactor

Pattern fill range

You can also control how many squares are used to make the patterns in the symmetry grid.

  • MinSquares
  • MaxSquares
Working example: if I have an 8x8 grid ('GridDimension' is set to "8") then the application has a 64-square grid to work with. The application will fill this square with either a symmetrical or a non-symmetrical pattern (randomly decided each time, 50% prob). The number of squares filled to make the pattern is decided by randomly selecting a number between 'MinSquares' and 'MaxSquares'. The number refers to how many squares on one half of the grid to fill as after filling a random amount of squares on one half it will then mirror those (if a symmetrical trial) or it will randomly pick the same amount of squares for the other half (ensuring it doesn't end up symmetrical by chance). Therefore the value of 'MaxSquares' should always be less than half of the total number of squares in the grid.

Data Analysis

 

After running the symmetry span processing script (download from here) you will get a data frame with the following variables. Information on running the script and obtaining the processed data is here

Variable Name Description
fta.score This is the score the participant achieved when the full-trial accuracy scoring method is used.
prop.score The score obtained by the participant when the proportion correct method of scoring is used.
number.successes The score obtained by the participant when the number of correctly recalled elements method of scoring is used.
processing.accuracy The proportion of the symmetry judgements that were successfully answered.
processing.median.rt The median response time for answering during the processing phase.
max.span Maximum list length at which the participant successfully recalled all the elements in the trial.
span.2.corr Number of trials successfully answered (all grids recalled) at list length 2 - "NA" value if there were no trials at this list length administered
span.3.corr Number of trials successfully answered (all grids recalled) at list length 3 - "NA" value if there were no trials at this list length administered
span.4.corr Number of trials successfully answered (all grids recalled) at list length 4 - "NA" value if there were no trials at this list length administered
span.5.corr Number of trials successfully answered (all grids recalled) at list length 5 - "NA" value if there were no trials at this list length administered
span.6.corr Number of trials successfully answered (all grids recalled) at list length 6 - "NA" value if there were no trials at this list length administered
span.7.corr Number of trials successfully answered (all grids recalled) at list length 7 - "NA" value if there were no trials at this list length administered
span.8. corr Number of trials successfully answered (all grids recalled) at list length 8 - "NA" value if there were no trials at this list length administered
span.9.corr Number of trials successfully answered (all grids recalled) at list length 9 - "NA" value if there were no trials at this list length administered
user The value of the username of the account which produced the data