According to the Pew Research Center, the U.S. Hispanic population reached 57 million in 2015, or 18% of the total population. Hispanics comprise large percentages of the populations of several U.S. states, including New Mexico, California, Texas, and Arizona.

Hispanics in the U.S. are influencing many aspects of American life — from culture to voting — and they represent a significant target market for advertisers. As the Hispanic population grows in size and influence, it is more important than ever to accurately measure the consumer behaviour attributes of this group.

For a variety of reasons, Hispanics are a challenging population to reach with standard survey research methods. Our experience has shown that Hispanics are consistently under-represented in surveys in the United States, and this is especially true for Spanish-dominant Hispanics.

What can researchers do to ensure Hispanic populations are accurately represented in surveys?

At Nielsen Scarborough, we employ a number of procedures to ensure accurate representation of Hispanics in our local-market surveys. We call these procedures differential survey treatments, or DSTs.

For example, we employ bilingual interviewers and all telephone interviews are conducted in English or Spanish, depending upon the respondent’s preference. We also provide Spanish-language translations of our survey materials, and a Spanish-only respondent phone line and Web site are available to answer questions and provide guidance about the survey. But this is just the beginning.

Accurate representation starts with the sample. In markets where we conduct telephone interviews, we use a dual-frame sample design or two sample frames: a random digit dial (RDD) frame and an address-based (ABS) frame.

The RDD sample frame is a list of telephone numbers. Our sample provider develops the list by identifying the 100 blocks (100 consecutive telephone numbers) in every active exchange in the market that has a minimum number of active, listed telephone households. This technique yields coverage of more than 95% of the landline households in the market, both listed and unlisted.

The ABS frame is a list of mailing addresses that may or may not be associated with telephone numbers. The ABS frame is used to represent cell-phone only (CPO) households and improve young demographic proportionality.

This is key to reaching Hispanic respondents because, according to GfK MRI, adults of Hispanic or Latino origin or descent have the highest incidence of living without landline telephones, with 67% reporting as cell phone only.

Incidence levels are another aspect of sample design in relation to Hispanics. Because Hispanics account for about 18% of the U.S. population, simple random sampling can be inefficient as the majority of households contacted will not include an eligible Hispanic respondent.

To increase the efficiency of reaching Hispanic respondents, we develop high-density Hispanic areas (HDHA) as separate sampling units within the sample plan. These are groupings of ZIP codes with high incidences of Hispanic adults. HDHAs ensure that more Hispanic respondents and, importantly, more Spanish language-dominant Hispanics, are included within the sample.

During the data collection process, we employ a number of additional DSTs to ensure more Hispanic respondents participate in our surveys. We use a respondent contact regime that includes extra call backs, mailings, and reminders to Hispanic respondents.

We also utilise a series of incentives and incentive promises to encourage Hispanic respondents to complete and return our mailed surveys or complete the survey online.

Once the data is collected, we sample balance for respondents who report being of Hispanic or Latino origin. We also sample balance for Spanish-language dominance using universe estimates provided by Nielsen. These estimates are provided as percentages of the Hispanic population, and we apply these percentages to the estimate of total Hispanic adults aged 18 years and older. This allows users of the data to analyse Hispanic populations by language preference.

How important is language preference? Earlier this year, Amazon quietly rolled out a Spanish-language version of Amazon.com. Amazon likely assessed the Spanish-language e-commerce opportunity and discovered there is a large population of highly engaged online shoppers who prefer Spanish.

Within the survey itself, we collect the following information from all Hispanic respondents:

• Hispanic or Latino origin.
• Hispanic by race (white, black, or other).
• Country of birth (in the United States or outside of the United States).
• Cultural descent: Mexican, Mexican-American, Chicano, Central American, South American, Cuban, Puerto Rican, Other Caribbean, or other.
• How long lived in the United States.
• Language preference (English and Spanish) in the home, away from home, and preferred language.

This information is key to understanding the differences between Hispanic populations in local markets. It is a mistake to look at the U.S. Hispanic population as one homogenous, monolithic group. Hispanics in Albuquerque are quite different from Hispanics in Denver or Miami as acculturation levels, country of cultural identification, and language preference can all significantly impact consumer behaviour.

These DSTs have served us well over the years and have ensured the accurate representation of Hispanics within our local-market surveys. But one thing every survey researcher knows is that you must evolve your methods to keep pace with the changing aspects of the populations you are trying to understand.

As the population becomes more multi-cultural, researchers must innovate to meet the challenge.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .