Conducting Hispanic Research via the Web [INSIGHT]

By Peter Roslow – President, Roslow Research Group / Adjunct Asst. Professor, Hofstra University

A number of research studies – from Pew and others – have confirmed that the incidence of use and frequency of use of the Web among U.S. Latinos rivals – and sometimes even surpasses – the levels of other Americans.  Yet survey research data collection via the Web continues to pose challenges when the target is Hispanics.  Especially if the survey target includes Spanish-dominant and/or less-acculturated Hispanics, building adequate samples has historically been problematic.  

For many reasons, Hispanics are less likely to participate in surveys …and while this is true regardless of the methodological approach, it is even more true when the data collection is handled online.  The growing popularity of conducting surveys via the Web has resulted in an ever-expanding number of online research panel sources.  But without exception, Hispanics are under-represented in these online panels.

Marketers love to use the Web for survey data collection for several reasons, but two predominate:  

(1)    It’s cheaper – Indeed, it can be as much as 40 or 50% less costly than telephone or intercept approaches; and this advantage can become even more pronounced as sample size grows.

(2)    It’s faster – Even versus computer assisted telephone approaches, survey turnaround – from date interviewing commences to date that findings are available – is accelerated.

Of course, regardless of the target, one always worries about representativeness with online surveys.  Because of the opt-in structure of panel samples; because of the fact that there is no such thing as a sample frame from which one can randomly draw potential respondents, online surveys can never be construed as approaching “random probability” in nature.  But the cost and speed advantages are usually so substantial, these often outweigh the drawbacks for many.

The bad news is that historically, because of the challenges with surveying Latinos, the two primary advantages largely disappeared.  Put simply, to do a quality job of surveying Latinos online, it took longer and cost more.

The good news, however, is that this is changing.  Successfully surveying Latinos online is becoming more doable every day.  While compared to online data collection among general market consumers, Hispanic surveys still take longer and cost proportionally more, there can still be substantive time and cost savings versus other methods of surveying Latinos.

As every Hispanic research specialist knows, there are a number of unique requirements when it comes to properly surveying Latinos.  And these specialized procedures equally apply when the methodology utilizes online data collection.  And an additional key, we have learned, is to access multiple research panels for drawing samples.  

Some panels are better at attracting more acculturated Latinos, some comparatively more effective with the less acculturated.  Some panels work better with specific age groups.  If the target needs to include Hispanic teens, for example, you’ll have more success with certain panel samples over others.  Conversely, if the target is older Hispanic adults, other sample sources may need to be utilized.  The bottom line is to be flexible in your design and execution, and to not limit yourself in terms of which (or how many) research panels you incorporate for sampling purposes.

As an example, a recently completed online survey required a sample size of several thousand Latinos, and included targeting adequately large age/gender and language sub-samples.  With the assistance of Opinion Access Corp. (a research supplier whom we have partnered with on a number of studies of this type), we were able to draw samples from six different online panel sources so that the age/gender/acculturation sample targets were met.

There are other limitations when using online data collection for Hispanic studies.  For example, because panel participation by Latinos is relatively low, and since one typically “burns through” sample very quickly, using such an approach for ongoing and/or periodic tracking studies is less viable.  But at the end of the day the fact remains that – depending on survey objectives and specifications – the Web CAN be effectively and efficiently utilized for much Hispanic research.

 

Skip to content