Experiment: An aiming task to test Fitt’s Law
Rationale: Research has well established that a negative relationship exists between speed and accuracy. Fitts (1954) formulated the “Index of Difficulty (ID)” equation which is used for the mathematical description of the speed-accuracy trade-off. The ID is dependent on the amplitude of the movement (A) and the width of the target (W), and is defined as:
ID = log2(2A/W)
Fitt’s Law (Fitts, 1954; Fitts & Peterson, 1964) describes the relationship between movement time (MT) and ID, and is a robust predictor of movement time, defined as:
MT = a + bID
where the intercept (a) and slope (b) are empirically defined constants that depend on the task characteristics (Langolf, Chaffin, & Foulke, 1976).
Procedure: Draw a line back and forth between two targets for a 10 second period. Target sizes are either Large (1.5 x 3 cm) or Small (0.5 x 3 cm). The movement amplitude is 5 cm between target centers. The order of conditions is randomized, and trials are blocked.
Analysis: Movements are segmented into primary and secondary submovements. For each movement, duration, amplitudes of the primary and secondary submovements and normalized jerk are measured and plotted as a function of condition.
Step 1: Setup targets: Targets are setup as Elements. An Element can be a geometrical shape, a drawing or handwriting pattern, or a bitmap. Elements may serve as targets or templates.
· In our experiment, we will need a total of four targets: left and right large targets (FLL and FLR), and left and right small targets (FSL and FSR). Each target is setup as an individual element.
· Left experiment tree window > Elements > right click > Create New..
General: Specify a 3-digit, alpha-numeric ID. In our experiment, the targets will be a shape and also a target. We will initially setup the Large Left target. Check the option ‘This element will be a target’
Dimensions: Specify physical features of the targets.
· Shape -> Rectangle
· Center point X-Coord and Center Y-coord -> Specify the location of the target on the digitizer (and correspondingly on the computer monitor
· 1/2 Length of X and 1/2 Length of Y -> Specify the width of the target and the height of the target respectively. In our experiment, the large targets are 1.5 cm wide x 3 cm high. We will not allow any error, as we want the subjects to accurately hit the targets.
Appearance: Here, we define the color of the targets, and whether the targets change color dependent upon successful or unsuccessful hitting of the target. The targets will initially be red in color, which we define in “Fill Color”.
· In this experiment, the targets will remain red throughout the trial, regardless if the participant hits the target or not. Another option is to have the target turn another color depending on whether it was successfully hit or not. Define these criteria by selecting “Successful Target” in the Appearance / Category menu, and clicking on an alternate color in the “Fill Color” box, then repeat for “Incorrect Target”.
· Click on OK and the first target is defined!
· To see what your target element looks like, right click on the element, and “Display in Recording Window”.
· To create the other elements, You can right click on the element > right click > create new.. Or duplicate it by element > right click > duplicate..
· Give each their own 3-digit, alpha-numeric ID, and modify the dimensions for each specific target.
· In our case, we would change the Center Point X-coord to 15 for the large right-side target. The 1/2 Length of X would be changed to 0.25 for the small targets.
Note: In the current version of MovAlyzeR (2.7.0.0), the screen will not go blank at the end of the recording period or the end of the trial. This results in the targets always being displayed, and the start of the recording period is cued by a “beep”. In our experiment, we wish to have the screen go blank between trials, thus we must set up an element of a blank screen. This is done the same way as setting up the previous elements, and we simply make the color of the target white so it cannot be distinguished from the display background.
Step 2: Set up Stimuli: A Stimulus consists of one or several elements. Stimuli are what will be presented to the subject in each condition. For our Fitt’s Law experiment, we will have two stimuli, one for the large targets (FLG) and one for the small (FSM).
· To create a new stimulus, right click on Stimuli > Create New..
· Give the new stimulus a 3-digit, alpha-numeric name.
· Click on the + sign to the right of the “Stimulus Elements” box, and a list of the elements previously created appears in the “Add Elements” box.
· Highlight an element and click on the “Add” sign at the bottom of the menu to add the element to the Stimulus.
· Add the two large target elements, and the blank screen element, to make up the large target Stimulus.
· Click on OK and the first Stimulus is done!
· Repeat the Setup Stimuli steps to create the small target Stimulus (FSM), but adding the small target elements and the blank screen element.
· You should now have 5 Elements, and 2 Stimuli, created.
· View the Stimuli to make sure they look ok by right clicking on the Stimuli, “Test in Recording Window”. The large target Stimulus should look like this:
Step 3. Setup Conditions: Conditions are made up of Stimuli. In the Conditions menu we will define the Stimuli to be presented, in what order, and what type of data we expect.
· To create conditions, Left window experiment tree > right click on Conditions > “Create New”.
General: Give the new condition a 3-digit, alpha-numeric name and a brief description
Stimuli: Stimuli per condition to be displayed during the recording/precue/warning.
· Instruction -> In this menu we can write the Instruction to be displayed
Warning Stimulus -> None
· Precue Stimulus -> In our Fitt’s experiment, we want the targets to be displayed on the computer monitor prior to the recording the data, so the subjects know where to position their pen. We will use the same stimuli for both precue and imperative stimulus.
· Duration -> 0.5 sec, Latency -> 0.5 sec, the Precue will be shown on the screen for half a second, which we specify and then disappear. Half a second later, the imperative stimulus is displayed.
· Imperative Stimulus -> Select the same stimulus as for precue. Notice that the targets will appear to blink off then on again, because we specified a latency period of 0.5 second between the Precue and Imperative Stimuli. If we wanted the targets to continue to remain on, we would have inserted “0” in the “Latency” box for precue stimuli.
NOTE: For the current experiment, we make the targets appear to flash so that the subject may position his/her pen above the target during the Precue period, then after the target blink the subject lowers the pen on to the tablet and begins to perform the drawing movement between the targets. Thus the target blink serves to inform the subject when to begin the movements.
At the beginning of the next trial we initially want the screen to go blank before the new targets appear. We thus insert the blank screen as the “Warning Stimulus” for the duration of 1 second with a latency of zero, meaning the display will go blank for 1 second, followed immediately by the Precue stimulus.
Consistency Checking: We specify stroke parameters we expect in the Condition.
· Minimum and Maximum strokes -> We set the range of stroke number to 1-100
· Stroke Length -> The required length of the stroke
· Range Length -> The range of acceptable stroke length error, and used to test the consistency of the strokes.
· Stroke Dir and Range Dir -> Required Length and error of direction Setting these values to “0” means consistency checking of the strokes will not be performed. If values are set, MovAlyzeR will check when processing the data to make sure each stroke falls within the defined length, direction, and range. If the stroke does not fall within the set parameters, an error will be given during the data processing stage.
Word Extraction: In the current experiment, no strokes will be extracted for analysis, so we simply check the options “Trials will be Recorded” and “Processed”. Word extraction is often used for handwriting experiments.
· Click on OK, and the large-target Condition is defined!
· Repeat the Setup Condition steps to create the small-target condition, using the small target stimuli for the Precue and Imperative stimuli. The conditions will be added to the Master list of conditions per user.
Step 4. Setup Subjects: Subjects are the people that participate in your experiment. Subject profiles are confidential and may only be viewed by the User that set up the subject.
· Left experiment tree window > Right click on Subjects > “Create New” to setup a new subject.
· If you have not specified a new password for the current User, the “Enter User Password” screen will appear before a new subject can be setup. The default password is userpass.
· Subject demographics, etc., can be added in the Subject menu. Give the subject a unique, 3-digit, alpha-numeric ID. Click OK to add the subject to the Master list of groups per user.
Step 5. Setup Groups: Groups are made up of Subjects. Groups may be made up of Pre-test, Post-test; older adults, younger adults, etc.
· Left experiment tree window > Right click on Groups > “Create New” to setup a new group.
· Give the group a 3-digit, alpha-numeric ID. Click OK to add the Group to the Master list of groups per user.
Step 6. Setup Experiment: An experiment is made up of all the preceding variables, more specifically, groups, subjects, conditions, stimuli, and elements.
· Left experiment tree window > Right click on Experiments > Create New.
· Give the experiment a 3-digit, alpha-numeric ID. Click OK to add the Experiment.
· Add a group to the experiment by right clicking on the “Fitt’s Law Groups” folder under the Fitt’s Law: FIT experiment, and “Add groups to experiment”. Select the group from the master list and click ‘add’
· Right click on the Group name, and “Add subject to Group”. Select the subject to add to the experiment and click ‘add’
· Add conditions to the experiment by right clicking on Fitt’s Law Conditions, then select the conditions to add; in our case, the FLC and FSC conditions.
· Once the Conditions have been added to the experiment, the number of trials for each condition must be specified. Right click on a condition listed under the general experiment (in our case, right click on a condition listed under “FIT Conditions” and menu for the condition will appear. Select the “Number of trials” option and insert the desired number, for our example we will select two trials per condition.
Step 7. Setting Experiment Parameters: The specifics of the experiment are defined in Experiment Settings, which is found by right clicking on “Fitt’s Law: FIT” under the Experiments folder. There are quite a few sub-menus under this heading, which define the capabilities of the MovAlyzeR program.
Running Experiment – Trials: We define the particulars of the trials
· Start -> This refers to the amount of time after a trial begins in which the movement needs to start. This can be thought of as the maximum reaction time allowed. Remember, however, that this period needs to include the time it takes to put the pen on the digitizer. Since the Fitt’s task is not a reaction time task, we will set Start to 2 seconds.
· Recording -> Specifies the amount of time you wish to record the movement. In the current experiment, participants will perform cyclical movements back and forth between targets for 5 s.
· Pen-lift -> Corresponds to the maximum amount of time allowed for the pen to be lifted off the tablet during the trial. This setting is most often used with handwriting experiments where multiple words are written, necessitating a pen lift between words. In other words, if the pen is lifted off the tablet ≥ 3 seconds during the trial, the trial will end and an error message will be given. This setting should be set at ≥ 3 seconds.
· Trial-to-trial -> Refers to the amount of time between the end of one trial and the beginning of the next trial.
· Process immediately after recording -> Corresponds to the data processing of each trial after it is performed. This can be checked for all studies other than those that entail word extraction, in which case it should be un-checked.
· Make recording window real size/Maximize recording window -> Pertains to the area of the computer monitor on which the stimuli will be displayed. If the stimuli should take up the entire screen, click on “Maximize recording window”. If the display on the screen should mirror the size of the digitizer recording tablet, “Make recording window real size” should be checked.
Recording – Procedure: The order of conditions and trials are determined in this menu.
· In the current example, we will randomize the condition order, thus that box is checked.
· Checking the “Randomize replications as well” box will randomly present trials from all conditions.
This acts as a complete randomization of the experimental design. If you do not check this box, but only have the “Randomize condition order” box checked, the order of conditions will be randomly selected, however, the trials within that condition will be blocked, such that all trials from Condition 1 will be presented together, all trials from Condition 2 will be presented together, etc.
NOTE: Checking on the “Randomize replications” box without checking the “Randomize condition” box will present the trials from Condition 1 first in a blocked manner, followed by trials in Condition 2 in a blocked fashion, etc. Thus, this essentially is not a relevant option as it is the same as Non-randomization of conditions.
Instruction constantly visible during recording: Check this if Instructions should remain visible on the screen throughout the duration of the trial.
View/Answer Questionnaire: Relates to a Questionnaire you may wish your subjects to complete before the start of the experiment, for example, a health-history questionnaire. The Questionnaire template is found under the main menu, “Settings” menu.
Processing - Time Functions settings: Refer to the filter and sampling rate frequencies.
· Filter Frequency -> The standard filtering frequency for experiments using a pen on a graphics tablet is 10-12 hz.
· Read only specified decimate-th sample -> We can also indirectly modify the sampling rate by choosing to read only a specific number of samples. Since we wish to examine all samples at the digitizer’s sampling rate of 206 Hz, we will leave this as “1”.
· Unrotate Automatically -> When using a horizontal, point-to-point movement task, it is often best to rotate the trajectory to the vertical plane in order to increase accuracy of movement variables using the MovAlyzeR. Rotating the trajectory 90º is equivalent to 1.57 radians, and we will do so for our experiment.
Processing – Segmentation: This menu allows for the parsing of the trajectory into primary and secondary submovements.
· Submovment Analysis -> The classic Fitt’s Law experiments demonstrated that smaller targets resulted in increased secondary or corrective submovements compared to large targets. In replicating these experiments, we will also wish to examine submovements and thus we check the “Submovement analysis” box.
Step 8. Running Experiment/Recording data:
· Right click on a subject listed under the specific experiment, in our case, FIT, and we’re going to run our subject, DIA.
· Click on “Run Experiment” and the trial will start.
· Once the experiment is finished running, the data will be immediately processed and a ANALYSIS CHART graph will appear.
· Analysis charts: Different dependent variables may be chosen for graphs by selecting the variable from the “Grouping” menu in the analysis chart and clicking the “Refresh” button. In the graphs, “Submovement 1” = the primary submovement, “Submovement 2” = the secondary submovement, and “Submovement 3” = the overall (primary and secondary combined) movement.
· In the example here, Condition 1 = the small targets, Condition 2 = the large targets (to find out order, go to “Action Settings/View Identities” and selection “Conditions”.
· It’s evident that the all submovements and the total movement duration are increased to small targets. Thus, the whole movement is slowed down when executed to small targets. This is also demonstrated by viewing the vertical velocity and acceleration graphs.
· Raw data charts: Charts of an individual trial may be viewed by right clicking on a single trial under the subjects directory within the experiment. Below is the data from one large target trial, rotated to the vertical. A number of charting options are available, including acceleration, velocity, etc.
· Data must be re-processed after altering experimental settings. To do this, right click on the experiment (FIT) and the reprocessing, summarization, and analysis menus appear.
· Summarization: The “Summarization” menu serves two primary functions. After collecting a number of subjects, you may select which subjects to summarize for graphing, and they are selected in “Summarize”. Data may be viewed as saved as an “.inc” file, which is in ascii format and may be cut and pasted into Excel, SPSS, or any other spreadsheet. This is also available from the “Summarize” menu, in “View extracted data”, then go to “Edit” and “Save as”.
The data is now ready for analysis in another software package!