Process for Developing Online Surveys using phpESP - php Easy Survey Package

Note: This service is for Library internal projects only.
Questions for starting projects should be sent to Sam McDonald at rulweb@rci.rutgers.edu

What is phpESP?

phpESP is a online survey support application that allows the rapid deployment of an online survey. The survey can be created, modified, tested, and activated by the survey creator. It can support large surveys (100 questions) and large respondent groups (4,000+) without a problem. The collected data can be viewed online in various simple cross-analysis and cross-tabulations. For more advanced analysis the data can be downloaded as a CSV file (Comma Separated Values) and imported into Excel or Access.

Process for Developing a Survey

1. Secure approval for campuswide and systemwide surveys

Approval for a campuswide survey needs to be secured from the appropriate campus director and for systemwide surveys from the appropriate associate university librarian and council.

2. Let the webmaster know you are designing a survey

The webmaster will want to know:

• who the contact person/creator will be
• when the survey is to go live
• how long the survey will be up for

3. Designing your survey

Design your survey using Word or another text program. Please read the section 'Survey Question Types' and 'Survey Limitations' below.

4. Creating your survey

Please leave enough time so that the survey can be created and demonstrated after creation and before deployment.

Contact the webmaster and set up an appointment to configure and create your survey.

In preparation, please have prepared in advance:

• Your designed survey in digital format (please send a copy to the webmaster, a draft is sufficient)
• A survey name (title)
• A sub-title (if applicable)
• The top matter / introduction / instructions.
• A concept for an end page. The survey can go to a custom 'Thank you' page or any other URL. The webmaster will help you design and mount this page with your prepared text.
• Publicity preparation:
  • If the survey is to be linked from the front page then the survey icon/link presence on the front page will have to be reviewed by the Web Advisory Commitee. Please allow plenty of time so that WAC can comment.
  • Note: The URL will be rather long and not initially suitable for typing in. A shorter special URL can be created but may take a week to configure.
  • News: Please contact the Communications coordinator to prepare your news well in advance of the survey live date.
  • Other Web Pages: Let the webmaster know if the survey is to be linked from other pages (eg. Art Survey, from the Art Research Guide).

The webmaster will help you program your survey questions into phpESP.

4. Testing and tweaking your survey

The webmaster will show you how to demonstrate, test, and modify your survey in advance of its activation.

5. Activating your survey

The survey can easily be activated through the survey management menus, but, it probably will need to be coordinated with links from other pages. You can set up a plan with the webmaster, communications coordinator, and others to launch the survey. Note: A survey can be toggled from test mode to development mode, but after a survey has been activated, it cannot be changed.

6. Concluding your survey

The survey can easily be ended through the survey management menus. Like activation, however, some coordination may be necessary to remove links to it.

After the survey has concluded, the cross-analysis and cross-tabulation tools can be used for review. If more analysis is needed, the results can be downloaded as a CSV files and imported into other tools.

Survey question types

• Yes / no (radio buttons)
• Single line text entry (fill-in-the blank) (1 line)
• Essay (large text box - defined by columns and rows eg. 40 x 8)
• Choose 1 of X (radio buttons) (may have an "Other" fill-in-the blank associated with it)
• Choose 1 of X (pull down)
• Choose Y of X (check boxes) (may have an "Other" fill-in-the blank associated with it)
• Rate (grid of radio buttons each row is question. can rate 1 - X (with N/A))(AKA Likert scale)
• Enter date (structured data, validated) (e.g. 4/21/2002)
• Enter number (prevents entry of non-number characters)

Survey limitations

• Cannot limit multiple voting. Meaning theoretically users can take the survey multiple times which might result in "ballot stuffing."
• For Likert scale questions, the range is by numbers 1,2,3,....,X, (N/A). The question itself must have the response definitions, eg. 1 = strongly disagree,..., 5 = strongly agree. Explanatory text cannot be inserted between questions.
• Likert scale questions cannot be a required element.
• Does not explicitly support 2-part questions (i.e. 3(a). and 3(b).) The survey will have to be designed to take this limitation of flow into consideration.
• The program allows radio buttons and check box responses to have "Other" as an option with a text field. The "Other" label, however, cannot be changed to a similar label such as "Why."
• There is no convenient way to allow a respondent to rank items.
• Surveys cannot be dynamic. For example, if you answered question "3" as "no", the question "7" never appears because it does not pertain.
• User can't confirm answers before submission. When the user clicks "Submit Survey" they do not a have a review option.

Help designing surveys

Questionnaires & Survey Design by StatPac Inc.

- 20 pages. Simple and direct. A chapter from their acclaimed Survival Statistics book.

More information

The source pages for phpESP can be found at:
Sourceforge page [http://phpesp.sourceforge.net/]
Sourceforge Project Page [http://sourceforge.net/projects/phpesp/].