Knowledge Base

Export the Statistics Result

Aim

To export the statistics of the dependent variables in the results table to an external program, for example to export the values of average velocity to SPSS.

Prerequisites

• The statistics result is open on your screen.
• Select the Data profile and Analysis profile from the lists on the toolbar.
• If the table layout does not match the requirements of the program you want to export the result to, see The Statistics Result.

Procedure

1. Choose Analysis > Export > Statistics, or in the Experiment Explorer under Analysis > Export, click Statistics.
2. Under Export, choose to export Trial Statistics or Group Statistics.
3. Click the Browse button to select the location for your export file. Its suggested location is the Export Files folder located in your experiment folder.
4. Under File Type, select either ANSI text (*.txt), Excel (*.xlsx), or Unicode text - default (*.txt).
5. Next to Missing Value Representation, enter the character that you want to export when the value for a particular cell is missing.
6. For text export, enter the Delimiter character you want to use to separate columns in the exported file.
7. Click OK.

Options

• Missing Value Representation. Missing values in a statistic result are statistics that could not be calculated for that particular cell. For example, if the subject never visited a zone during trial 2, the result for the statistic Latency to First and for the dependent variable In Zone is "-". By default, missing values are exported as "-". Make sure the destination program recognizes the character you specify as a missing value.
• Merge Column Headers (selected by default in new experiments). Keep this option selected if you want to export to applications like Excel or SPSS. If you do not select this option, each of the four column headers in the results table will appear in separate rows in the export file. Note that the order of the text elements in the header depends on the order of the headers in the results table. You can change the order using the Layout function.

Notes

• The name of the export file is Statistics-[Experiment name]. If you export results multiple times to the same location, EthoVision attaches a number at the end of the file name: (1), (2), etc.
• Choose Unicode text to export according to the Unicode character set. This supports most world languages, including Chinese. Choose ANSI text to export the data as text files according to the ANSI character set. Choose this option only if the application in which you will open the exported file cannot handle the Unicode character set.
• The values exported have a variable number of decimals, depending on the number of integers. The total number of significant digits exported is always six (for example, 1.23456 or 1234.56), or seven if the unsigned number is smaller than one (0.123456). Because of this restriction, there may be a very small difference between the values in the results table and in the export file. This difference is in the order of 1/10000th of the measurement unit.
• To change the layout of the exported table, follow Modify the Layout of the Results Table and then export the results.

Export Settings for Specific Applications

• For R: Export to text. Make sure that R can recognize "-" as a missing value. Use the na.strings parameter to include all possible "no data" characters or strings, for example: loadfile <- read.table("exportfile.txt", header = TRUE, sep=";", na.strings = c("-", "NODATA", ....))
• For SPSS: Select the option Merge Column Headers.
• For Systat: Each case must be written in one line. Separator (for text files): space or comma.
• For SAS: For text files, use the DELIMITER option to specify the text delimiter if this is other than space. For Excel files, use the PROC IMPORT procedure with the SHEET option to specify the worksheet to import. Use the GETNAMES option to read variable names from the first row of the worksheet.

The Statistics Result

Learn About

• Trial Statistics Result
• Group Statistics Result

What Do You Want to Do?

• Modify the layout of the results table
• Sort the results table

Trial Statistics Result

Aim

The Trial Statistics tab gives the statistics for each trial separately.

Procedure

Choose Analysis > Results > Statistics and Charts. The Trial Statistics tab opens.

Default Layout

In the Trial Statistics table, each row corresponds to a subject tracked in one arena during a specific trial.

The row headers (A) tell you which Trial, Arena and Subject a specific cell corresponds to. Remember that, depending on your experiment design, one Trial may contain one or more Arenas, and one arena may contain one or more Subjects.

The column headers (B) show the type of dependent variable (e.g. velocity), the body point (e.g. center point), the type of statistic (e.g. mean) and the measurement unit.

Additional headers appear depending on your data selection:

• Selection results. If you define two or more Result boxes in the Data profile, the name of the boxes are shown in the first cell (C).
• Time bins. If you selected Results per time bin in your Data Profile, extra rows are added, one per time bin, and the headers show the corresponding time periods. For example: 0:00:00 - 0:00:20, 0:00:20 - 0:00:40, etc. (D).
• Results per zone. If you selected Results per zone in your Data Profile, the zone names are displayed on rows, together with the body points in those zones (E).

What you see in the result table depends on which Data profile and Analysis profile are selected on the toolbar.