Ethnic Sampling

Ethnic Sampling for Social Research Advanced sampling for sub-populations

When targeting ethnic minorities, respondents from specific country of orogins or likely religions, general RDD sample does not perform well due to a low incidence rate. Another option are specific mobile phone provider for migrants, however the incidence rate is still low. Using our targeted ethnic mobile sample it is possible to survey respondents with high accuracy

Global live screening for migrants, ethnic minorities and likely followers of religions.

Hard to reach audience using SMS

Using an onomastic approach, by blending RDD with list assisted and social media data in order to target certain low incidence audiences within Developed countries

World Poll Gallup

The Gallup World Poll tracks the most important issues worldwide, such as food access, employment, leadership performance, and well-being. Studies were conducted in more than 160 countries that include 99% of the world’s adult population.

SMS Data Collection Sub-Saharan Africa

An experiment conducted together with Gallup that compared SMS data collection with Face-to-face data collection in order to outline mod differences between SMS and Face-to-face.


Tracking American consumer habits on a monthly basis, using a National representative Mobile sample for the entire Mainland US

Countries in Turmoil: Libya & Ukraine

During outbreaks of civil uprisings or wars, it is difficult to collect data face-to-face or online. CATI, however, remains a great way of collecting data for which we provided sampling frames for Libya and Ukraine.

Pew Global Attitude Study

Mobile Sample icon

Fully mobile based as this is the best mode to reach this targe group.

Global coverage for our ethnic sample

Can be used to collect data via CATI, SMS or WhatsApp


Sampling Methodology

Sample Solutions developed the enriched-RDD sampling approach in-house. We are using multiple steps to create a highly targeted sample that nevertheless contains probabilitic criteria such as the selection probability and frame size for social research. Here are the steps for the sample generation:

  1. 1. A raw mobile RDD sample is generated
  2. 2. The sample is screened for activity to show working/non-working numbers
  3. 3. We match the working numbers with social media data, public search queries and messenger services
  4. 4. We check wether there are indicators that a respondent is in the needed target group
  5. 5. The sample is split into three batches: matched, not-in-target group and no linkage possible

This appraoch allows researches to apply either a probabillty-based approach or quota based approach to sample these sub-populations.



The ethnic sample is overlayed with public data that can be found from various sources publicly. If data is not available publicly, we cannot retrieve it. Most social media and messenger networks have the option to set-up privacy options. When these are activated, we do not have a way to retrieve the data. Sample Solutions keeps track on the timestamp when the data has been processed and whichs sources have been used.

Data subjects have the option to retrieve their information, edit and delete them based on their data subject rights.

Ethnic Minorities

RDD Landline Sample can be stratified based on a country’s various non-geographical strata (network operators) or geographical strata (administrative divisions, geographical regions, Census regions, NUTS regions, etc.), with equal probability of selection within each strata. If the stratification is based on a geographical stratum, the sample is constructed based on the population’s distribution within the strata. Each possible telephone number per stratum is generated and stored together with an incremental id.

Country of Origin

We are using an onomastic approach to flag name-enriched RDD mobile sample with likely country of origin. Next to the onomastic approach we also check for any indicator on social media profiles or messenger services that might indicate a different country of origin.

In all cases we start with a true mobile RDD sample that is then screened using various Big Data sources.


We are using an onomastic approach to flag name-enriched RDD mobile sample with likely religion. This signficantly increases the chance to identify followers of eg. Islam, Judaism or other religions.

We only apply this to the mobile frame.

telephone-sampling process






We would source landline sample from white pages data where available and then apply an onomastic approach to the names. In most cases this yields older respondents and those that have been in the country for longer.

We can only target literate people as they use a messenger or social network. Overall education level is higher, more males and overall younger people. Since we screen an RDD sample there are indidivudals that cannot be matched, thus a coverage error is introduced.

No, the data is processed live meaning the generated mobile RDD numbers are compared live with various social media and messenger services.

Variables like region gender and age estimate can be added.

Our data is profiled / estimated. We have high evidence that a specific trait like religion, country of origin or ethnicity exists but we do not have hard facts or collected data.

Depending on the volumes and specifications, the default turnaround time for this sample is 2 to 3 working days.