Part B

Part B: Collection of Information Employing Statistical Methods

B. 1.1 Potential Respondent Universe

All visitors to the HUD.GOV website in December 2008/January 2009 have the potential to be a respondent. Approximately 10% of visitors will randomly be selected to be offered the opportunity to complete the online survey during their visit to the website.

B. 1.4 Expected Response Rate

The www.hud.gov website receives approximately 3.75 million visits each month. Since we anticipate offering the survey for a two-week period, this reduces the pool of visits/visitors to 1.875 million. Approximately 10% (187,500) will randomly be selected to receive the survey and of these we expect a 5% response rate or approximately 10,000 completed surveys. Should we not receive 10,000 completed surveys within the two-week period, we would extend the survey for up to an additional two weeks to ensure a robust sample. Existing literature suggests that web assessment surveys have response rates from eight to twenty percent, depending on the type of audience (see Appendix E). However, a similar survey was performed to assess the website of the National Library of Medicine (NLM) home page in 2002, which yielded a response rate of about five percent.¹

B.2.1 Procedure for Data Collection

The website survey will be live for a period of two weeks, but it will not exceed four weeks. Not every website visitor will be shown the survey. The survey will be randomly generated, both in terms of visitors being asked to take the survey and where in the site the invitation is presented to them. This invitation will be shown requesting feedback from the visitor (see Appendix F). There will be an option to decline the survey. If a visitor accepts the invitation to the survey, he/she will be shown the survey. Those declining to complete the survey will continue with their normal tasks without interruption.

B.3 Describe Methods to Maximize Responses Rates

The questionnaire shown in Appendixes B and C has been developed to provide a maximum response rate among potential participants. To achieve high response rates several steps will be taken:

First, the web survey is designed to be customer friendly. The navigation through the survey is simple and average response time will be less than two minutes.

Second, to reduce the burden on respondents, all users will be given the choice to not take the survey.

Third, the questions will be customized to the website. Questions pertinent to their visit to www.hud.gov only will be displayed to the users.

These procedures are consistent with contemporary strategies in the field of web survey research (see Appendix E). Many visitors to the HUD.GOV website have a personal or professional interest in the success of their ability to retrieve information or conduct business, which should enhance their willingness to respond.

B.4 Describe Tests of Procedures

Once OMB approval is obtained, and prior to programming the final survey questionnaires into the website, the HUD will pilot test it with 10 to 15 respondents.

B.5 Statistical Consultants and Information Collection Agents

The HUD.GOV Customer Satisfaction Website Assessment survey has been developed by HUD in conjunction with a HUD contractor, Mainstreet Technologies, Inc.. This small business is responsible for all data collection and data analysis.

1 Minutes Of The Board Of Regents Meeting on September 10-11, 2002. National Library Of Medicine. National Institutes Of Health, Department Of Health And Human Services