CEP Experiment Editor

In an attempt to ease the creation of your own studies, we decided to develop a graphical editor that allows for the creation of experiment configurations. Subsequently, we provide a brief description of the CEP Experiment Editor.

The Editor

After launching the editor and initializing a new study, the editor is not empty but provides a minimal experiment configuration to get started. This configuration usually consists  of a start node, an end node and a single Show Message Activity. We will later have a closer look at the different types of activities that are supported by CEP Experiment Editor. Let’s have a look at the four different sections of CEP Experiment Editor first though.

editorInitial

The main working area is the Experiment area, where you create your experiment by dragging elements from the Nodes area on the right. For each element selected in the Experiment are, the corresponding configuration properties are displayed in the Properties view on the bottom of the screen. These properties contain the text of a message box or the selection of a psychological instrument. Similarly, if no activity of the experiment is selected, i. e. by clicking on the white space, you will see the properties of the experiment itself.
Additionally, another tab next to the properties view provides the validation of the experiment. Here you can see if your model consists errors. By double-clicking on an error message, you can jump right to the problem.

In the Questions / Models view on the right of the screen, items that can be reused in different activities during an experiment are collected. These elements must be created before they can be used in a node, i.e., an experiment activity. A simple example are questions that will be used in a survey.

Experiment Structure

Every experiment has exactly one Start Event and exactly one End Event. The flow is sequential, which means that it goes from node to node according the edges. Every node (except the Start Event , the End Event, and the XOR) must have exactly one incoming and one outgoing edge. Cycles are not allowed.

Single Branch

The simplest possible experiment structure is one branch.

If there is only one branch, the edge after the Start Event must be named. The edge description, i.e., 1324 in this example, constitutes the code that has to be entered when a subject starts the experiment.

Multiple Branches

More complex experiments are usually split into different branches, e.g., control group vs. treatment. This can be realized by using XOR nodes to split the control flow in two or more branches. This influences how edges have to be named throughout the experiment configuration. It is not allowed to name the edge after the Start Event, but every edge the goes out of a XOR must be named. The subject has to enter the code of the respective group at the beginning of the experiment. This way, group assignments can be realized by distributing different codes to your subjects. In order to join two branches, another XOR is needed.

Multiple Branches with Repeated Splitting

If there are more than one splits, the code that has to be entered by your subjects turns into a concatenation of the edge names on the path that should be followed.

If your subject should execute activity 1 and then activity 3, the subject must enter “1245” (“12” and “45”) at the beginning of the experiment. If your subject should complete activity 2 and then activity 3, the code to enter is “2345” (“23” and “45”). This way, subjects can be assigned to all possible combinations.

Types of Experiment Activities (Nodes)

This section gives you a short overview of the different types of activities provided by Cheetah Experimental Platform that can be configured using the Experimental Editor.

Directing the Control Flow of Your Experiment

This set of nodes let you direct the control flow of your experiment. Additionally, you can display messages and interrupt the execution of the experiment in order to perform a task not covered by Cheetah Experimental Platform.

Start Event

Every experiment must have exactly one Start Event. This has no incoming edge and exactly one one outgoing edge. The outgoing edge must be named if there is only one branch, as this constitutes the code for starting the experiment.

End Event

Every experiment must have exactly one End Event. This has non outgoing edge and exactly one incoming edge.

XOR

XOR gateways are utilized to split the experiment execution in different branches. In order to do this, the edges have to be named as described in the previous section. By distributing different codes to you subjects you can perform a group assignment according to your experiment’s needs.

Show Message

This allows to show a message to the subject at the position you choose in the experiment.

  • Message: The message you want to post.
  • Name: name of the node (internal in the Experiment Editor).
  • The title of the message in the experiment.

Ask for code

Sometimes it is necessary to stop the execution of the experiment until a certain condition is fulfilled. For example, when performing a study using eye tracking you might need to calibrate the eye tracker before a certain activity is executed. This node asks the subject to enter a code before they can continue, forcing them to contact the supervisor of the experiment.

  • Expected code: the code the subject must enter to continue.
  • Name: name of the node (internal in the Experiment Editor).

Feedback

This node is usually the last activity in your experiment as it asks the subject to provide some feedback on your experiment. The feedback can be entered as free text.

  • Name: name of the node (internal in the Experiment Editor).

Modeling Activities

Since we are interested in process modeling, the following set of activities described the modeling tasks and notations supported by Cheetah Experimental Platform.

BPMN Modeling

This node represents a BPMN modeling task during your experiment. Several configurations options are available as described here briefly. For example, there might be an initial model the subject can edit. The properties view contains the following elements.

  • Description: this is the description of the modeling task, i.e. what should the subject should do?
  • Initial Model: this property specifies the initial model which can be edited by the subject. If no initial model is required, this field can stay empty.
  • Modeling Process: This is the ID of the BPMN modeling task. A unique id is required for the analysis of the modeling task.
  • Name: name of the node (internal in the Experiment Editor).

Declare Modeling

Similar to the BPMN modeling activity, the experiment can contain a modeling activity using the declare language with the following properties.

  • Constraints: You can specify the constraints that should be available to the subject during the modeling task.
  • Initial Model: this property specifies the initial model which can be edited by the subject. If no initial model is required, this field can stay empty.
  • Modeling Process: This is the ID of the Declare modeling task. A unique id is required for the analysis of the modeling task.
  • Name: name of the node (internal in the Experiment Editor).

Change Pattern

Similarly, it is possible to use change patterns, instead of change primitives, for creating a BPMN model.

  • Description: this is the description of the modeling task, i.e. what should the subject should do?
  • Initial Model: this property specifies the initial model which can be edited by the subject. If no initial model is required, this field can stay empty.
  • Modeling Process: This is the ID of the Change Patterns modeling task. A unique id is required for the analysis of the modeling task.
  • Name: name of the node (internal in the Experiment Editor).
  • Patterns: This property allows you to select the change patterns set that should be available during your experiment.
  • Undo feature: This property specified whether your subjects are allowed to undo their change pattern applications.

Tutorial

During the execution of an experiment, it is essential that all subjects are well informed about the features of your modeling environment since struggles with the tool can blur your effects, making a scientific analysis difficult. To counter this threat, we incorporated interactive tutorials in CEP. This node can be selected in order to incorporate an interactive tutorial in your experiment. The following properties can be specified.

  • Name: name of the node (internal in the Experiment Editor).
  • Tutorial: this property specifies the tutorial that should be incorporated, e.g., BPMN.

Asking your subjects questions

Finally, experiments are usually not only about modeling but also about answering questions. This set of activities contains various types of questionnaires that can be configured using Cheetah Experiment Editor.

Comprehension

Comprehension activities present the subject with a model (or an image in general) and juxtapose the model with a question to be answered. This is especially useful if the impact of certain modeling construct on the process model’s understandability are assessed.

  • Model / Image: Select an item (model or image) from the Questions / Models / etc. view.
  • Name: name of the node (internal in the Experiment editor).
  • Question: Select the question from the Questions / Models / etc. view that you would like to ask.

Survey

The difference between a comprehension and survey activity is that surveys are usually not accompanied by models or images and that surveys are intended to ask a series of questions instead of a single question. This ways, surveys constitute the foundation of several psychological questionnaires provided by Cheetah Experimental Platform.

  • Name: name of the node (internal in the Experiment editor).
  • Question: Select at least one question from the Questions / Models / etc. view. It is also possible to load folders of questions.

Psychological Tests

Cheetah Experimental Platform provides a list of psychological tests that can be executed. This node allows you to select the desired test.

  • Name: name of the node (internal in the Experiment Editor).
  • Psychological Tests: Select the test to be used in the experiment.

Interested?

Are you interested in using Cheetah Experimental Platform for your research? Do not hesitate to contact somebody of our team! We are more than happy to assist you with Cheetah Experimental Platform.