Stage 9: Perform Basic Data Analysis
E-Prime collects only single participant data files. For group analysis, the single participant data files must be merged into a master file using the E-Merge application.
|Stage 9: Basic Data Analysis|
|1) Merging Data|
|2) Checking Data Condition Accuracy and Data Quality|
|3) Data Editing|
• Tracking modifications
|4) Analysis and Export of Data|
• Exporting raw data
• Exporting a subset of the data
• Creating and exporting tables
Stage 9, Step 1: Merging Data
E-Merge may be launched via the Start button. From the Start menu, choose E-Prime 3.0 to display the E-Prime 3.0 menu, and then, select E-Merge. Alternatively, E-Merge may be launched via the Tools menu in E-Studio. E-Merge opens to display a Quick Reference dialog for guidance through the steps involved in merging files.
To merge individual data files, navigate to the folder containing the data files using the Folder Tree, select the files to be merged in the File List view, and click the Merge button.
In many circumstances, data files may be saved in different folders. For example, data files collected by different research assistants may be saved in separate sub-folders within the same folder, or on a shared drive. The E-Merge application provides a specialized merge procedure to facilitate the process of merging files arranged in this way. The Recursive Merge operation merges files from a selected folder and all of that folder’s sub-folders. When a merge operation is launched, the user must specify whether the operation will be a Standard Merge or a Recursive Merge operation.
Data files may be collected on multiple machines, and then merged for analysis. When collecting data on multiple machines, take care to assign unique participant numbers. For example, when collecting data on four separate machines, a scheme could be used in which all participants run on the first machine are numbered 101-199, all participants run on the second machine are numbered 201-299, etc. For merge operations, it is important to keep experiment names and participant numbers unique, or conflicts will arise during a merge operation. There are two methods of merging the data for analysis.
Method 1: Copy all of the data from the separate machines to a single folder on one machine. Copying may be accomplished through file transfer, using read/write CDs, memory sticks or other portable devices, or e-mailing attached files to be saved into the common folder. The relevant files are those with the .edat extension.
Method 2: Have a shared disk space accessible by all of the machines, and save the data file to that shared space . Each station could record its data in a separate area on the shared disk, and E-Merge’s Recursive Merge feature could be used to aid in the process of merging these files from different folders. For example, a two-computer, two-experiment shared data folder might look like the folder tree below.
NOTE: Running an experiment on a network has significant performance implications and is likely to result in timing errors.
Stage 9, Step 2: Checking Data Condition Accuracy and Data Quality
Once the data has been merged to a master file, the file should be checked to verify the data it contains and that the values are within expected ranges. Rather than a checking of the logging of variables (i.e., this would have been done prior to data collection), the checking at this point is more in reference to the completeness and appropriateness of the data, the experimenter’s assignment of participant numbers, and assignment of participants to conditions. Particularly, if several assistants are collecting data for one or more experiments, it is easy to mistakenly assign the same subject number to several participants, or to fail to assign participants to a particular condition. E-DataAid is the application within E-Prime which allows the examination and modification of data files.
Stage 9, Step 3: Data Editing
Data may be examined and edited using the E-DataAid application. To launch E-DataAid, choose the E-Prime folder from the Start menu, and select E-DataAid.
E-DataAid may also be launched from the Tools menu within E-Studio. The E-DataAid application opens without loading a particular data file. A data file must be opened within E-DataAid using the Open command in the File menu, or by clicking the Open tool button.
E-DataAid works much like Excel, allowing the addition of columns (i.e., variables) and modification of cell values. For example, to edit the value in a cell, simply place the cursor in the appropriate cell, delete the current value, and type the new value. To select a column, click the column header, and to move that column, click and drag the header to the new location. Specific tool buttons, commands, and filtering abilities are included in E-DataAid to organize, rearrange, or reduce the amount of data displayed.
E-DataAid includes additional features specifically designed to track modifications made to an E-Prime data file. For example, all edits made to a data file (e.g., modification of cell values, added variables are displayed in red, indicating a modification to the file. In addition, an annotation of each modification is written to an Annotations record which is saved with the data file. The example Annotations record below (from a merged data file) indicates the single participant data files from which the data originated, and notes that the subject number was modified for one of the participants (specifically, Participant 1 was changed to Participant 10).
The Annotations record serves as a history of merges and modifications to allow the user to track not only the source files contributing to the master data file, but also any changes that were made to the data.
Stage 9, Step 4: Analysis and Export of Data
E-DataAid also allows the formatting and exporting of raw data, a subset of the data, and data tables (e.g., Mean RT x Condition).
Exporting raw data
To export the raw data, use the Export command in the File menu, or click the Export tool button. The Export command allows the output format to be specified for a particular statistical package.
Formatting for several common software packages (e.g., StatView, SPSS, Excel, etc.) is offered as an export option. The “Other” formatting option allows the user to manually specify the format of the data during export so that the output file may be imported into a package that is not specifically listed in the export options.
Exporting a subset of the data
During an export operation, all of the data displayed in the spreadsheet is exported. Therefore, to export only a subset of the data, the display must first be reorganized using the Arrange Columns and/or Filter commands in the Tools menu prior to the export operation.
The Arrange Columns command allows the user to choose whether to hide or display specific columns in the spreadsheet, and to organize the order of the displayed columns.
The Filter command allows the user to limit the display to certain values within a specific variable (e.g., only correct responses). To set a filter, choose the variable (i.e., column) name and the type of filter to apply (checklist or a range of values). Then select the specific values to include (i.e., apply a Checklist filter) or the boundaries for the range of values to include (i.e., apply a Range filter). Filters are inclusive, indicating the values that will be included in the display.
Once the columns are arranged and/or filters are applied, the export operation may proceed as usual, with only the displayed subset of the complete data file being exported.
Creating and exporting tables
A summary table of values (e.g., MeanRT x Condition) may be exported using the Analyze feature. The Analyze command may be invoked via the Tools menu, or by clicking the Analyze tool button.
The Analyze feature allows the user to specify the summary analysis to perform. An analysis is defined by dragging the appropriate factors from the Variables list to fields (i.e., Rows, Columns, or Data) to specify them as row, column, or data factors. Once the analysis is specified, click the Run button to generate the table, and (from the resulting Table dialog) the Export button to perform the export operation. With the Save Analysis and Load Analysis commands, the Analyze feature allows analysis specifications to be saved so that they may be reloaded and run again. Analyses are saved with the .anl extension.
In addition to an export, the table may be copied directly into Excel, or to the clipboard for saving in an alternative software package.