Online Research Panels and the Hispanic Market [INSIGHT]

By Raul Lopez – Principle and Chief Research Officer / New American Dimensions

When I worked on my first survey with Hispanic respondents, circa 1980, the best-in-class approach to reach a fully representative sample was to go door-to-door.  Of course, this was pre-cellphones, pre-internet and when Hispanic landline telephone penetration was around 70 percent.  The market was growing and evolving rapidly (and still is, but in different ways).  By the later part of the 80s we were able to complete fairly representative surveys via landline.  We’ve come a long way since those early years in Hispanic marketing, and for the most part, the Hispanic consumer is as savvy and sophisticated as the general market consumer, but there are still sampling nuances to take into account.

Today, the Hispanic online presence is on par with the general market.   Hispanics are populating the major research panels in greater numbers, and new panels claiming to have Hispanic respondents pop up regularly.  While online access and use among the least acculturated Hispanics (and most likely Spanish preferred) is also increasing rapidly, this group in particular is still poorly represented in most online research panels.  It comes down to this:  there are a great many Hispanics online and a growing number in research panels, but sometimes not enough respondents to fill specific study quotas.  Additionally, they may not always be representative.

For many studies, especially those that are more general in nature and/or nationally targeted, enough online, less acculturated Hispanics are available to develop a fully representative sample.  For those studies, a best practice is to draw from multiple online panels, with participants invited to the survey via a link.  In aggregate, multiple samples recruit from diffferent sources, including the web and Spanish-language TV, assuring a sample that is well balanced on acculturation and other key demographics and, thus, more representative of the targeted groups.  Importantly, it is best to use de-duping software to eliminate panel “groupies” that subscribe to multiple research panels.  

Using only one panel source can lead to research results that are not representative of your target.  This is especially true in the Hispanic market, which is characterized by variables not encountered in the general market, such as acculturation level and country of origin.  I often speak with clients who conducted Hispanic surveys using an online sample from one panel, only to discover odd results, like 90% of their “unacculturated” respondents chose to complete the survey in English or mean results for key demographic measures that are way off compared to previous data.  Multiple panel use along with careful monitoring of key measures can ameliorate these types of issues.

In other cases, where the project has more specific respondent requirements, it becomes difficult to reach a representative number of less acculturated Hispanics via online panels AND meet the specific respondent specifications, even when using multiple panels.  For example, a national survey of 1,000 Hispanic grocery shopper respondents which is representative when looking at acculturation and country of origin is often quite feasible using multiple online panels.  However, take that same study and add more quotas – a mix by market, category requirements, and no children –  and the feasibility of completing those quotas 100% via online panel is severely impacted, especially with the less acculturated segment.

In those cases, a mixed methodology approach is recommended, where a significant portion of the sample is completed online (for example 80%) and the remainder (20%) is completed via an ITC (Intercept to CAPI) approach.  The percentage split between online and ITC will vary by study depending upon the individual study specifications.

The potential respondent is approached by an interviewer carrying a data transmission enabled tablet with a link to the online survey.  This way we can work anywhere even if there is no Wi-Fi available.  If the respondent qualifies and agrees to do the interview, they will complete the exact same interview as the online respondents; the only difference is that we came to them versus the online people who are invited via an email from the panel manager.  This eliminates instrument bias since the survey is identical.

Another benefit to this method is that it allows us to reach respondents who may not have the experience or ability to complete self-administered online questionnaires, since there is an interviewer present to assist them.  For respondents who are not computer-literate, this approach can reduce anxiety, improve response quality, and assure that all questions are answered.  

A mixed methodology is slightly more expensive than one using only online panels, so we only recommend this in cases where reaching a representative sample of less acculturated respondents is essential to the study AND a representative number is not available from online panels.

Every study is different and its unique set of parameters need to be addressed individually.  However, by partnering closely with clients, and having a thorough understanding of a study’s objectives, experienced multicultural researchers are able to identify the best approach at the most economical price.