Knowledge Base

Select Data for Analysis

Why Select Data?

In many cases you may want to visualize or analyze your entire dataset and obtain complete descriptive statistics for all your events in all your observations. If this is the case, do not select data, but continue with Visualize Data, or Calculate Statistics. If you want to analyze only a subset of your data, or analyze a number of behaviors together, carry out data selection first. To open the Data Selection window, choose Analyze > Select Data > New Data Profile.

There are four ways to analyze a selection of data:

• Analyze several events as if they are one
• Analyze some observations, subjects or events, not others
• Analyze events that occurred in specific time intervals
• Analyze data in regular time intervals

What Do You Want to Do?

Analyze several events as if they are one

• Example 1 – Visualize behaviors of the group Play type as one behavior.
• Example 2 – Analyze the behaviors Left hand moving and Right hand moving together as Any hand moving.

Merge the data, then visualize and run the analysis. See Merge Data.

Analyze some observations, subjects or events, not others

• Example 1 – Calculate the average duration of speech of the subject Child, not of the subject Mother.
• Example 2 – Visualize data in observations of female subjects, not of males.
• Example 3 – Calculate the rate of occurrence of events of the group User Error, not of others.

Filter the data, then run the analysis. See Filter Data.

Analyze events that occurred in specific time intervals

• Example 1 – Calculate the number of times the child smiled when the mother was present.
• Example 2 – Visualize the data from the moment the mother entered the room until ten minutes later.
• Example 3 – Analyze events when the subject's heart rate was higher than 100 beats per minute. This is possible when your license includes the External Data Module.

Select intervals, then run the analysis. See Select Intervals.

Analyze data in regular time intervals

• Example 1 – Split your observations in 10-minute intervals and calculate statistics for each of them.
• Example 2 – Split your observations in three equal intervals and calculate statistics for each of them.

Define Time bins, then run the analysis. See Select Regular Intervals.

The Data Selection Screen Explained

To open the data selection window choose Analyze > Select Data > New Data Profile and name the data profile. Alternatively, right-click the Data Profiles folder under Analyses in the Project Explorer and choose New.

The window that appears contains the Components pane on the left and a window on the right with the data selection sequence. The Data Selection sequence in the window on the right contains two boxes connected by an arrow.

• The Start box (left) contains all the data in all the observations currently stored in your project.
• The Results box (right) contains the data used for analysis.

The Start box is connected with a direct arrow to the Results box. This means that all data are used for analysis. Add boxes in the sequence to define which data to analyze. See Data Selection: General Procedure.

The Results Box Explained

The Results box shows the result of the data selection. When your project contains two observations and you defined three subjects and seven behaviors in the Coding Scheme, both the Start box and the Results box show three subjects and seven behaviors. The Results box also shows two intervals, one for every observation.

Filtering

When you filter by subjects, behaviors, modifiers, or duration (see Filter Data), the numbers in the Results box match the numbers you have selected in the filter box. Filtering has no effect on the analyzed observation time and no intervals are created. So the number of intervals in the Results box remains the same as in the Start box and represents the number of observations. When your data selection contains no data, this is shown in red on the Results box.

Intervals

When you select intervals by subjects, behaviors, modifiers, or duration (see Select Intervals), you create time intervals based on the selected data. No subjects, behaviors, or modifiers are filtered out, so these numbers remain unchanged in the Results box. The number of intervals changes; it shows the total number of intervals over all observations. If the data selection contains no data because the selected subject, behavior, or modifier has not been scored, the Results box shows 0 intervals in red.

Data Selection: General Procedure

Make space between the Start and Results box by moving the Results box to the right. To do so, click the header of the Results box so that the pointer becomes a four-arrow cross. Drag the box to the right. Then do the following:

1. Click one of the buttons in the Components pane to select the criteria to add to the data selection sequence. For example, to analyze only a selection of observations, click the button next to By Observation name.
2. Choose the values that specify your selection.
3. Click OK. A new box appears in the Data Selection window. Drag the box over the arrow between the Start and Results box. When the arrow turns white, release the mouse button.
4. If necessary, draw extra arrows to connect selection boxes. To do so, point to the center of the first box, press and hold the left mouse button and drag toward the center of the other box. Release the mouse button when the cursor has reached the center of the other box.
5. Click the Visualize button on the toolbar to check whether your data selection works as expected. The selected data are displayed in the white fields of the Time-Event Plot.

Tip: Click the Settings button on the Results box to change its name. This name is shown in the visualization and analysis results, so you always know which data selection the results come from.

Note: You cannot create connections from the Results box to any other box, or from a box to the Start box.

Merge Data

To treat two or more coding scheme elements (behaviors or modifiers) as one. For example, treat the behaviors Grab antennae, Grab mandibles, Grab head as one behavior Grabbing and calculate the total frequency or duration.

Prerequisites

• You carried out your observations.
• The Data Selection window is open (Choose Analyze > Select Data > New/Open Data Profile).

Procedure

1. Click the Merging button on the Start box.
2. Click the tab for the elements you want to merge.
3. To create a merged group:
   • Behaviors – Click the Add button at the bottom. The merged group appears at the bottom of the list.
   • Modifiers – Select the Modifier group name in which you want to merge Modifiers (see Notes about Merging). Then click Add. The merged group appears in the Modifier group.
4. Name the group and press Enter.
5. Drag and drop elements into the group.

How Merged Data Are Analyzed

Merged Behaviors

Behaviors A and B both have a duration of 5 seconds and they overlap for 1 second. After merging, the frequency of the merged behavior is 1 and its duration is 9 seconds.

Merged Modifiers

The top part in the picture below shows three occurrences of Behavior Walk, each with a duration of 3 seconds, scored with associated modifiers Slow, Normal and Fast. Merging the modifiers into one modifier Speed results in one occurrence of Behavior Walk with a duration of 9 seconds (bottom picture).

Notes about Merging

Behaviors

• Modifiers are not visualized and analyzed when you merge Behaviors.

Modifiers

• You can only merge Modifiers from the same Modifier group.
• If you create a Merged group of numerical modifiers, this group is considered as nominal. This means you cannot calculate average modifier values, but can calculate statistics like frequency and duration.

To Remove a Merged Group

To remove a merged group, select the group name and click the Remove button at the bottom. To remove all Merged groups in one of the tabs at once, click the Reset to default button.

What Happens with Merged Groups If I Change the Coding Scheme?

• If you add a new element to the coding scheme – If you want to include this element in the merged group, add it manually.
• If you delete an element in the coding scheme – The element is deleted from the merged group as well.

Filter Data

To choose a subset of data and exclude others from visualization or analysis. You can filter observations and filter events. Observations and subjects can be filtered directly, for example by choosing only the first five observations. You can also choose observations or subjects indirectly, by the value of independent variables. For example, to analyze only those observations in which the subject's age was below 10. See Filter by Independent Variable Values.

Filter Observations

• By observation name – to analyze some observations, not others. See Filter Observations.
• By Independent variable with scope observation – to analyze observations with specific independent variable values. See Filter by Independent Variable Values.

Filter Events

• By Independent variable with scope subject – to analyze events of subjects with a specific independent variable value. See Filter by Independent Variable Values.
• By Subjects – to analyze events for some subjects, not others. See Filter Subjects.
• By Behaviors – to analyze some behaviors, not others. See Filter Behaviors.
• By Modifiers – to analyze some modifiers, not others. See Filter Modifiers.
• By Duration – to analyze events with a minimum or maximum duration. See Filter by Duration.

Filter Observations

To choose some observations for analysis and exclude others.

Prerequisites

• You carried out your observations.
• The Data Selection window is open (Choose Analyze > Select Data > New/Open Data Profile).

Procedure

1. Click By Observation name.
2. Select the observations and event logs to analyze. If your observations contain multiple event logs and you want to analyze only a few, unfold the observation item and select the event logs.
3. Insert the box in the sequence between the Start box and the Results box.

Filter by Independent Variable Values

To choose observations, event logs or subjects according to the value of an independent variable.

Prerequisites

• You carried out your observations.
• The Data Selection window is open (Choose Analyze > Select Data > New/Open Data Profile).

Independent Variables and Scope

Each independent variable has a variable scope. System variables, like Start time, always have scope Observation and therefore filters for these variables appear under Filter Observations. User-defined variables can have scope Observation, Event Log, or Subject. Depending on the scope you can filter observations, event logs or subjects according to the variable values. For example, the user-defined independent variable Patients file has the scope Observation, while Gender has the scope Subject. Therefore, By Patients file appears under Filter Observations and By Gender under Filter Events.