Data Quality & Methods
OMB’s guidance to agencies about how to construct post-transaction customer feedback surveys has instructed them to include a question measuring respondents’ trust or confidence in the High Impact Service Provider (HISP) entity.
OMB Circular A-11 Section 280 requires transactional feedback surveys to include an overall trust score of the entity in response to a Likert Scale question (preferred 5 point). Agencies are provided two sample statements and are instructed to make only minor edits. If a HISP would like to modify the statement, or use a format such as a “thumbs up” or 5-star rating iconography, they must discuss with OMB through the PRA approval process. The sample sentences provided are:
- “This interaction increased my trust in [HISP name].”
- “I trust [HISP] to fulfill our country’s commitment to [relevant population].”
In some cases, HISPs have been permitted to use the word “confidence” rather than trust.
HISPs have implemented OMB guidance to varying degrees, and OMB has taken a refreshed review of all surveys in FY23 Q2 to improve the quality and consistency of these surveys. To date, there are six versions of the trust/confidence question that have generated the data provided in this report.
- I trust - “I trust OPM Retirement Services to fulfill our country’s commitment to provide quality services to retirees and their families.”
- Increased my trust - “This interaction increased my trust in the IRS.”
- Recent experience and trust - “Based on my recent experience…I trust Social Security to meet the public’s needs.”
- Confidence rating - “Please rate your level of confidence in the Trademark Assistance Center.”
- Interaction and confidence - “This interaction increased my confidence in the Homeownership Center.”
- How confident - “How confident are you that you could or did enroll in a plan that meets your health and financial needs?”
The trust score includes all three versions of the first three questions listed above and the confidence question similarly includes all three versions of the last three confidence question.
HISPs are required to collect and report quarterly feedback one year after their service designation. Based on varying degrees of capability and expertise, as well as an ongoing refinement and improvement of designated services, many HISPs are still improving the quality of their surveys and data reported. As of Fall 2023, OMB now has dedicated staff working with HISPs to standardize and improve the quality of individual surveys, with a target of all reporting HISPs meeting a newly created Data Confidence Measure of at least 80% by the end of FY23. The Data Confidence Measure includes a review of the following criteria:
- Reporting required trust measure
- HISP met the requirement for submitting a tally of survey responses for Q1-Satisfaction and Q2-Trust in the 5-point Likert Scale breakdown. A survey is marked 0% if this criterion is not met.
- Unit of Observation - The point of survey administration is within the context of a transaction with a designated service. A survey is marked 0% if this criterion is not met.
- Response Rate - The response rate (number of responses divided by the number of customers offered the survey) is above 1%.
- Sample - The number of responses is greater than the minimum sample size (calculated at 95% confidence level and 5% margin of error) given the customer base.
- Survey Quality - In a careful review of all submitted survey questions, the survey must include high fidelity Satisfaction and Trust questions, and score above 66% when tallying for the number of high-fidelity questions over the number of questions that should be included. The set of questions that should be included is culled from the categories of Satisfaction, Trust, Effectiveness, Ease, Efficiency, Transparency, and Employee Interaction, and is dependent on the specific attributes of the corresponding designated service. In general, a high-quality survey should provide useful and relevant information on the appropriate drivers of customer experience for the service. A survey that is missing Q1-Satisfaction and Q2-Trust receives a mark of zero within this criterion.
If a survey meets both the Reporting and the Unit of Observation criteria, then all five criteria are weighted equally to arrive at a percentage for the Data Confidence Measure.