Raosoft, Inc.
Case Study on Survey Hosting for a Government Agency- Bureau of Public Debt

Introduction
Chronology
Special considerations
Conclusion
Figures

Introduction:

The Department of the Treasury's Bureau of Public Debt (BPD) launched an organization-wide Employee Survey targeted at all of the approximately 1800 employees. The survey consisted of 30 questions, most in matrix format. All responses were anonymous; at the same time respondents were tracked to determine who had completed the questionnaire. The survey ran for two weeks, and 76% of employees that received an invitation to the survey responded. For those interested, several graphs detailing the response rate and curve are included at the end of this case study.

Chronology:

Our project contact at the BPD authored the question set using EZSurvey and Raosoft provided design assistance on the electronic form. Raosoft, Inc. tested the survey instrument then installed it on the web server for additional quality checking. The plan was for the survey to run for two-weeks, with a reminder email sent after one week to those that had not responded. Data collection began on a Tuesday, and continued for 10 business days. Raosoft sent the first set of 1800 email invitations beginning at noon on the first day. EZSurvey's bulk email wizard was used to embed an ID and password in each message. Initial response was high, and as expected, tapered off within 3 days. On day 6 of the survey, Raosoft used the bulk email wizard to sort out which people had responded, and sent reminders to approximately 900 non-respondents. Again, the response rate jumped to the level seen on the first day. The re-send contributed an additional 25% of the total responses received. Upon completion of the survey, Raosoft divided the data set up into 12 subsets based on the respondent's office. The data was returned via email so that our project contact could develop an EZReport template for analysis.

Special Considerations:

Raosoft, Inc. was chosen to host the web form and distribute email invitations due to our experience managing this type of project, and to provide additional protection to the respondents' identities. Raosoft agreed not to return information to the BPD that would link a respondent to his or her answers. Organizing a re-send to non-respondents required that Raosoft use identifiers as part of the email messages sent. As an independent third party, Raosoft exercised due care to ensure respondents felt secure, and this success appears evident from the very high response rate. Because many employees take vacation time in July, the two-week run length was justified. This allowed more people to participate.

Conclusion:

The 2002 BPD Employee Survey represented a successful project hosted and managed in house at Raosoft, Inc. The 76% response rate indicates that employees were motivated to provide feedback and were comfortable doing so through a web based form.

Survey Response Rate Graphs:

	This graph shows the percentage of remaining non-responders that completed the survey on a given day. On day 6, reminders were sent to non-responders. The total response rate on the second sendout is nearly as high as the response rate on the first day.
	The second chart shows the percentage of the total population that responded on each day. The percentages responding on and after day 6 are higher than would be expected if the response rate had trailed off naturally.
	The plateau seen from days 3 to 5 is typical of single-mailout surveys. The re-send was an inexpensive way to increase effectiveness.