2010-09-13
QPL Reference Manual

The QPL system provides several ways to see the results of your questionnaire project. When you log into your web site as an administrator, you can view one of the built-in reports, such as the history log or the summary statistics report, or create your own report using an SQL query. You may also export the respondent data from the web site in any of several formats while using the administrative web page, and then analyze it using a SAS or SPSS program that the QPL system can generate from your questionnaire program.

At GAO, we have developed a special process to display a questionnaire and the summary analysis (generated by SAS) as web pages, which we have call an "e-supplement." An e-supplement may be posted on GAO's public web site as an addendum to an audit or evaluation report. The web questionnaire form looks like the original form but can not be used collect data. Links have been added to each question that jump to the response summary table in a second web page. The summary tables web page also has links back to the orginial question.

If your project used open-ended or string questions, you can use the content analysis functions on the administrative pages to categorize and tabulate the responses. These functions include several on-line reports you can use to see the results of this analysis as well as an export function that lets you download the tagging data for further analysis.

Finally, you may also ask you your network administrator to dump the underlying MySQL database and use that to rebuild the database on your local system and that to analysize your survey results.

Clean Skip Path

Starting with QPL 6.0.26, an option has been added the lets you create a copy of your questionnaire data where the questions that should not have been answered are set to NULL, or missing, values. This can be useful because because it makes it easier to interpret the results when viewing them with the questionnaire summary statistics report or when analyzing the exported data with your statistical analysis software, both of which can exclude cases with missing values from the analysis.

The system follows the path each respondent took through the questionnaire according to the GOTO and NEXT skip instructions you used to guide the respondent to the next appropriate question, and then sets the answers to the questions that were not on this path to NULL, or missing, as it copies your data.

This copy is referred to as the "clean" data set on the administrative web site pages. The original data is referred to as the "raw" data. Once you have used the option, Export data > Build clean data set, to create a clean copy, then options will appear in the various report and data export functions that let you choose the raw or clean data set for the report or export.

While the raw data continues to be updated as respondents continue to use the web site, the clean data set is only updated when you re-run the Build clean data set option again.

QPL 6.1.3 extends this functionality further by also creating a "dirty" version of your questionnaire data. This is the opposite of the clean data. Here, All of the answers on the skip path are set to missing values. This lets you quickly see where respondents are answering questions that they should not answer, according to your skip instructions. If all of the answers you see are only the default answers, then you may confidently use the clean version of the data set for your analysis. Like the clean data set, the dirty data set may be viewed with the on-line Summary Statistics report or exported as a data file. It will have the label, "dirty," appended to the file name.

Questionnaire Summary Statistics Report

The Questionnaire summary statistics lets you monitor the progress of your project. You can use various options when building your web site to control what questions are included in the report. This report is not, however, intended to be used for reporting final statistics about a project. The report includes all the cases (excluding only cases that are specifically programmed by the System Administrator on the Default Project Settings page (by default, it is set to not include cases create by test accounts), and does not offer any way to weight samples or do cross-tabulations.

If you build a clean copy of your data, as noted above, you will have the option of running this report using the raw, clean, or dirty versions of the data.

You may also create a Custom Report to create special versions of the statistics report by including only certain sub-sets of your respondents in the report.

Export Data Files

The export files are available in three different formats: fixed format, tab delimited, and an askSam import file. You should use one of these files locally to do a thorough analysis of your data. These files, of course, are just a snap-shot of the results at a particular time during the data gathering period. But you can export a new copy of a file for analysis at any time. Every version of your export file will contain all of the data from the project data base on the web server (i.e., not just the new cases). The cases included in the export data files are set by the System Administrator on the Default Project Settings page and by default will not include records created by test accounts.

The fixed format file is intended to be used with the SAS or SPSS statistical analysis packages. The QPL system can generate SAS and SPSS programs that can read and label the fixed format file and save it as a system file. (Note: The responses to OPENEND questions are not included in this file. Instead, a numeric field that contains a count of the number that a characters a respondent typed into an OPENEND question is used.) This file may be generated with or with out (for compatibility with previous versions of QPL) the internal system variables which show who created each record and when each record was last updated, among other things.

The tab-delimited file is a generic data format which may be loaded into a variety of software packages, such as Microsoft Excel or Access. You can select responses for particular questions, including system variables, to be included in the tab-delimited file.

The askSam import file is an ASCII text file that has been specially formatted to make it easy to import into the askSam text retrieval/data base system. askSam is useful for manipulating and analyzing long text responses to OPENEND questions. It will also contain any content analysis tags your reviewers may have applied to STRING and OPENEND question responses.

If you have created a clean copy of your data file, you will have the option of exporting a raw, clean, or dirty version of each of these files.

Content Analysis File

Content analysis functions were added to QPL with Version 5.04.001. Using these functions, you can perform a systematic analysis of the responses made to STRING and OPENEND questions by using multiple reviewers to read each response and select one or more pre-defined codes or tags that fit the response. The functions will produce frequency statistics on how often each code was applied to responses for each question, inter-rater reliability statistics, and summary reports of responses and tags by reviewer or merged across selected reviewers.

You may define up to 256 tags (or up to 3,200 tags if the project was built using the /c CONVERT modifying option to increase the maximum number of tags, ex. /c 3200) that may be applied to any or all of the STRING or OPENEND question responses. And you may use any number of reviewers to read and apply the tags.

MySQL Dump File

Generally, you would only use a dump of the MySQL data base file from the web site if you were interested in analyzing any of the built-in system variables, such as information that was stored in the project log table or user account table. You can, however, use MySQL to export subsets of a table as a tab delimited file that may be imported into other analysis packages such as Microsoft Excel or Access.