In the task editor, the trial system is located on the left side of the screen. In general, the trial system is used to create several repetitions (trials) of a task . The trial system consists of two parts, the "Factortree" and the "ConditionTable". When the "Factortree" is edited by changing trialgroups, factors or levels, the "ConditionTable" is automatically updated based on the settings in the "Factortree."
The factortree can be used to change and modify the top hierarchical structure of a task. Specifically, it is there to modify the 3 top elements of the task hierarchy:
- Trialgroups (contain factors and frames)
- Factors (contain levels)
- Levels (are crossed to create conditions)
The names of factors and their levels must be different from other factors and levels. To keep your study organized and the event logic clear, avoid similar-sounding labels such as "level 1" and "level 1.2" for your conditions.
Avoid creating large quantities of levels that only have one trial in each. This causes the study to be very crowded and disorganized. A better structure would be to have a few levels (or conditions) with multiple trials in each.
The trialgroup acts as a container for a group of trials within the task which are roughly similar. Most importantly, all trials within a trialgroup have the same series of frames, with the same objects and events on them (although properties of objects can events can change from trial to trial). The trialgroup sets the overall structure of a trial. Each task needs at least one trial group.
As with most experiments, very different types of trials should be intermixed (e.g normal and catch trials). Multiple trialgroups within a task can be used to create and intermix very different trials.
By adding factors to a trialgroup, the user can create different versions of a trial. Typically, the name of the factor details which aspect of the trial will be varied (e.g. position, size, image catetory etc). The levels of the factor then describe the characteristics of the change (left/right, small/medium/large, etc). The more levels a factor has, the more (trial) variations exist and a larger stimulus space is created. This also means that a factor must have at least 2 levels to extend the overall stimulus space. Importantly, each factor is linked to a factor variable. This is useful for two main reasons. First, events can then be be customized to certain conditions (you can instruct the experiment do something ONLY IF Factor "Image Category" equals level "house"). Second, each trial records the value of the given factor, so that later the experimenter can easily separate different conditions when looking at the dataset.
When adding a factor to a trialgroup, you can choose to reuse an existing factor variable or create a new variable. Factor variables can be linked within or across tasks, but it is important to know that a change of a linked factor variable will trigger changes in all tasks and trialgroups where the factor is used.
Factors can be either fixed or random. A trialgroup must have at least one fixed factor and each factor must has at least one level. This distinction does not affect how many trial variations are defined, but it will affect how many trials a subject sees when the task is executed. Fixed factors generate more trials when the task is executed, while random factors only generate more trial possibilities to choose from for each trial. For example, if the factor "position" with the levels "left" and "right", is fixed, a subject would see a "left" variation AND a "right" variation of the trial. However, if the factor "position" were random, the subject would see a "left" variation OR a "right" variation of the trial. For more information on factor randomization, please see the section on trial randomization.
Levels are the possible values for each factor and determine the characteristics of each factor. The factor "images" could, for instance, have the three levels: "houses, faces, cars". If this factor is the only factor in the trial group, a condition is created for each level, which is displayed in the "Condition Table" at the bottom left of the editing screen. If several factors are created, each level of a factor is combined with all the levels of the other factors within the factor group (the factors cross). In a 3-factor design with factor "images" having 3 levels, factor "position" with 2 levels, and factor "size" with also 2 levels, there would be a cross of 3 x 2 x 2 = 12 conditions in total.
Usage: To define the exact influence a factor level will introduce, the user can click on the factor level in the Factortree and then modify the objects on the current frame. The changes will affect all trials and conditions in which the respective factor has the chosen level. This approach is particularly useful when a stimulus variation is dependent on the level of factor, but not on a condition or a single trial.
The "Condition Table" can be used to modify the middle layer(s) of the task hierarchy, specifically, conditions and trials. Due to the hierarchical nature of a task, the crossing of the factor levels generates the conditions separately for each trialgroup. These are then shown in the Condition Table at the bottom left of the task editor. Whenever there is a change in the factor tree, the Condition Table will update automatically. For each condition, the number of trials can be defined separately. It can also be set to 0 if some conditions should not be shown at all.
The Condition Table is also grouped according to trialgroups and each trialgroup has exactly one "default-trial" in the Condition Table. The default trial is located in the first row for each trial group, above the conditions. When the default trial is selected, all of the changes made are applied to all trials in the trial group. Any (trial, condition, level) variations, which were made before selecting the default trial will be overwritten by changes in the default trial. Overall, modifying the default trial resets all trials in a trialgroup to the same "default" value and all trials within a trial group are selected and edited simultaneously. This is useful to define the rough overall frame structure.
It is most reasonable to place new elements first in the default trial and then adjust for desired variations using the condition or trial selection.
The conditions are determined by the number and crossing of factor levels. Each condition is clearly defined by the values (factor levels) of all factors in its trial group. For instance, one condition could be defined by the following three level values: face image, left, small; while another condition could be defined by the values: face image, left, large, and so on.
To make changes to the individual conditions, users can click on the condition and then make changes to objects on the canvas. The changes made will only apply to the selected condition (i.e. all the trials in that condition). This is a precise way to make changes to collections of trials at once.
Trials are located within the conditions and are the most basic unit in the trial system hierarchy. Each condition is created by default with one trial. The number of trials per condition can be adjusted in the Condition Table.
If there are several trials within one condition, the condition can be opened (expanded) in the "Condition Table" and individual trials can be selected. Consequent changes to the canvas and objects will then only apply to this particular trial.
Generate Trials by Adding Stimuli
The trial structure can also be created automatically via the file manager and the integrated multi-select function. This is useful if the variation is produced by different stimulus content (different images, videos, sounds etc.). To do so, an image, video or sound object (container) must be selected, then the file manager must be opened via the folder icon in the properties panel. Then select several (2 or more) pictures/videos/audio files and click on the "Assign" icon (grid symbol with arrow) in the lower right corner.
A new menu will open up asking you how to use these multiple selected files. Here you have 3 main options how to assign / use the selected files, with some sub-options:
Trials: Assign the Files to Trials
- Assign the files to existing trials: The image/video/audio files will be assigned to already existing trials in the selected condition.
- Create as many trials as files selected: The image/video/audio files will be matched to the number of trials in the selected condition. There will be exactly 1 trial for each selected file. Former trials in the condition will be removed.
- Create a new trial for each selected file: For each selected image/video/audio file, a new trial will be added to the selected condition. The new trials will be appended to existing ones.
Conditions: Assign the Files to Conditions
- This option assigns the files to a selected factor group and existing conditions.
Factor Levels: Assign the Files to Factor Levels
- Assign files to existing factor levels: The images/video/audio files will be assigned to already existing levels in the selected factor.
- Create new levels within an existing factor: For each image/video/audio file, a new level will be added to the selected factor. The new levels will be appended to existing ones.
- Create a new factor and new levels: A new factor will be created. For each image/video/audio file, a new level will be added to this newly created factor. There will be exactly 1 factor level for each selected file.
After choosing the overall assignment method, you can determine which file should be assigned to which trial, condition, or factor level. The respective menu will display two lists, one for the selected files (right), and another for the assignment items (left). You can then drag & drop individual files from the right to the left, manually determining which file should be associated to which trial, condition, or factor level. However, you can also use the "Auto Assign" button, which will assign all files automatically in a random order. This way you can setup your stimulus associations quickly.
Working with Trials
After the factorgroups, factors, levels, conditions, and trials are set up and frames and objects are added, the next step for the user is to modify the object properties such that different trials will show either different stimuli, the same stimuli in different positions, with different sizes, or other such modifications. Here it is of central importance to select the appropriate modification level for each modification made. Overall, there are 4 levels of modification:
- Trial-Group-Modifications using the "Default Trial": Changes all trials within the selected trialgroup.
- Factor-Level-Modification: Changes all conditions (and trials within these) in which the selected factor possesses the selected factor level.
- Condition-Modification: Changes all trials within the selected condition.
- Trial-Modification: Changes only the selected trial.