Overview of ACSI Benchmark

OVERVIEW OF THE AMERICAN CUSTOMER SATISFACTION INDEX BENCHMARK

DESCRIPTION: The American Customer Satisfaction Index (ACSI) methodology measures satisfaction based on a standard index composed of three questions that are rated on a 10-point scale:

• Respondent’s overall satisfaction as of today

• Satisfaction compared to the respondent’s expectations

• Satisfaction compared to the respondent’s ideal

The ACSI organization measures satisfaction with the Federal government in the 4^th quarter (October-December) of each year. Respondents who have experienced some federal agency during the prior 12 months are asked the satisfaction questions about the specific federal agency and those experiences and not about the Federal government in general or all experiences they may have had over the period.

With a standard customer satisfaction index, any public or private organization using the ACSI methodology can compare their satisfaction index with other organizations. This is particularly beneficial for Federal agencies that want to “benchmark” their satisfaction index to other agencies or the private sector.

SAMPLING FRAME METHODOLOGY:

For a number of years, the overall Federal benchmark was based on the ACSI scores achieved by agencies conducting studies.  Since 2010, the Federal benchmark has been based on scores from a panel of approximately 1500 (half telephone respondents, half internet respondents). This approach ensures that the general population is represented better in the sample.  The ACSI does not release response rates by data collection mode. Since the ACSI is a national study surveying a large range of products and services simultaneously with the possibility for respondents to complete multiple industry/government segments, it is not possible to determine response rates for a particular subset of interviews.

The data collection for the ACSI Federal government study is an independent effort and not submitted as a generic IC of ICR 1090-0007. The ACSI Federal government study is not a "panel study" per se. While the ACSI uses a large "panel" of email addresses for approximately 50% of data collection done via the Internet, in technical terms a "panel study" is one where the exact same group of respondents is interviewed over time, wave after wave (perhaps every year, every quarter, etc.), again and again. ACSI uses a "panel" of email addresses to get valid email accounts and send out emails asking for participation in the survey. However, ACSI re-samples randomly each year, so theoretically there is a “brand new" group of 1500 or so participants each year.

Random-Digit Dialed (RDD) Sample (50% of respondents):

The RDD portion of the sample is very straightforward. ACSI calls landline telephone numbers by area code proportional to the geographic distribution of the population to guarantee representativeness. Numbers are called multiple times to try and get responses in order to minimize sample bias rather than just calling new numbers until the quota of 750 is filled.

Internet (50% of respondents)

ACSI purchases the email sample from a highly reputable commercial sampling company. This company maintains information on the demographics of the participants in a double opt-in Internet panel (e.g., gender, age, race/ethnicity, income, education, etc.). The specific methodology for creating the panel is proprietary to the commercial sampling company. Each year prior to selecting email addresses to send an invitation to participate, the panel list is corrected (i.e. "trimmed") so that it matches Census demographics. ACSI then draws a random sample from that "corrected" panel. This process better ensures that the sample pool has characteristics closely matching the citizen population based on demographics. As with the RDD sample, to minimize sample bias the ACSI attempts multiple re-contacts via email to convert each into a completed response prior to adding new sample to achieve the quota of 750.