Zappi Multicultural Representative Sample
At Zappi we aim to increase the usage of profiling questions to understand how respondents identify themselves and enable our customers to consistently explore data among underrepresented groups.
As part of these efforts, we've introduced a Multicultural Representative Audience in the US, designed to help businesses adopt an inclusive lens when running research with us.
This is a nationally representative audience that helps brands collect learnings that represent the opinions of consumers and make decisions that promote greater diversity and inclusion.
- Unlock multicultural insights for better decisions. Optimized screening with updated ethnicity profiling, designed to accurately reflect the country’s social reality.
- Robust quotas for targeted analysis. This audience uses a default sample size of n=500 that will provide a robust read across sub-segments.
- Available for Zappi’s most popular products. These include the solutions in the Creative Suite, ScreenIt, Zappi Concept Test, and Zappi Pack Test.
Improving the way we screen for ethnicity in the US
We evaluated the profiling questions used in our surveys to identify opportunities to improve our data collection techniques.
Some key learnings we identified are:
- The terms race and ethnicity have varied meanings for people across the country. Including these terms in the question wording can bias response. A more generic question, “which of the following best describes you?” pushes focus to the answer options and allows respondents to easily identify themselves.
- The term Hispanic applies to a large and diverse population, many of whom prefer alternative terms (i.e. Hispanic / Latino / Latina / Latinx). Countries of origin or ancestry are also very important to peoples’ identities. Mexico, Puerto Rico, and Cuba in particular have large populations within the US which should be uniquely identifiable.
- Providing a “not listed” response option is important for allowing respondents to express that their identity cannot be described by our current options.
- Including references to specific countries alongside regional answer codes makes it easier for respondents to identify themselves. For example: South Asian (including Bangladeshi, Bhutanese, Indian, Nepali, Pakistani, and Sri Lankan).
Our approach in the Multicultural Representative Sample (US)
We have leveraged learning which can be implemented using our current platform capabilities, and we will continue to iterate as the platform is further developed in order to deliver even more robust and representative insights over time.
One notable limitation at the moment is the need profile using a single code question - the respondent can only choose one option. This allows us to deliver more robust samples in target quotas. Future development will enable multi-code selection while delivering robust quotas.
Screener question [single code]
Which of the following best describes you?
- East Asian (including Chinese, Japanese, Korean, Mongolia, Tibetan and Taiwanese)
- Hispanic/Latino/Latina/Latinx - Mexican
- Hispanic/Latino/Latina/Latinx - Puerto Rican
- Hispanic/Latino/Latina/Latinx - Cuban
- Hispanic/Latino/Latina/Latinx - Other
- Indigenous American / First Nations
- Middle Eastern
- South Asian (including Bangladeshi, Bhutanese, Indian, Nepali, Pakistani and Sri Lankan)
- Southeast Asian (including Burmese, Cambodian, Filipino, Hmong, Indonesian, Laotian, Malaysian, Mien, Singaporean, Thai and Vietnamese)
- Not listed
- Prefer not to say
Does the Multicultural Representative Audience allow for further segmentation by acculturation among ethnic groups?
We do not measure each participant’s acculturation level within the standardized screener questions. A person’s acculturation level is typically determined through questions about their English language proficiency, years living in the US, and cultural influences. These also vary by ethnic group. Such additional questions may be added through Custom Questions for further subgroup analysis. However, incidence rates should be considered as it may not be feasible to achieve a robust sample for analysis among some niche audiences. Further, surveys using this audience are only run in English, meaning a proportion of Unacculturated populations will not be targeted with the survey.
How has the quota for Hispanic respondents been determined? And, does it account for acculturation levels within the Hispanic population?
Hispanic participants are identified in the screener through a question about ethnicity (Hispanic / Latino / Latina / Latinx - Mexican, Cuban, Puerto Rican, or Other). The quota for these participants represents 18% of the total sample, which is in line with the Hispanic population size determined by the US census.
Zappi's surveys using the Multicultural Representative Audience are targeted at English speakers who represent at least 74%* of the US Hispanic population and include people from all levels of acculturation, which is determined based on a variety of cultural factors. It's worth noting that participants in online panels are more likely to be identified as Acculturated or Bicultural Hispanics, which represent approximately 71%* of the US Hispanic population.
Given this, Unacculturated Hispanics may be underrepresented within Multicultural Representative Audience samples. However, given the predominance of English speaking and Acculturated/Bicultural Hispanics, we determined that the best approach was to set the Hispanic quota in line with the US census figure for the total Hispanic population. This allows us to robustly represent the opinions of the total Hispanic population.
* 74% of the US Hispanic population report being able to speak English “Very well” or “Well,” with a further 21% indicating they speak English, but “Not well.”