Improving your CATI Fieldwork – Part 2

In the previous week, we discussed various means for improving CATI Fieldwork for B2B projects using advanced sampling techniques. This week we will look at consumer projects which consist of various different types of sample files.

Consumer projects can run on client sample (e.g. customer satisfaction studies), RDD sample (e.g. public opinion polling) or lifestyle sample (e.g. brand awareness of a specific age group). All these sample types can be used individually or combined.

A. Consumer Sample

research and results 2016 highlights blogpost

  1. Sample Type

For each consumer project, it is essential to analyze the options for sampling. For general population studies, for example, it is feasible to use RDD sample since it is cost-effective and yields a large coverage while lifestyle sample would be more expensive.

Nevertheless, if certain quotas need to be met (quota-sampling) a mix of RDD sample and lifestyle sample may prove useful. Initial usage of RDD sample (dual-frame) can cover the majority of the population. In order to reach special groups at the end of a survey in specific areas or age groups, targeted lifestyle sample can be used.

2. Screening Techniques

Though a pulsed or verified sample is slightly more expensive than a raw sample, it is worth the investment as clean samples have a much higher return on investment. Not only the hit rate of the sample should be taken into account, but also the dialing time. Modern Predictive diallers can significantly reduce waiting times for agents,  nevertheless, many B2C projects are still not dialed predictive but rather through manual or automatic methods. Using a pre-dialled sample reduces waiting time.


Screening allows us to remove disconnected records while enrichment adds more value to the sample. Through enrichment additional information like names, age,  locations and more can be added to the sample.

Enrichment reduces the need for agents to ask for these variables during fieldwork and diminishes the chance for human error.  This also mitigates cost and also increases strike rate per interviewer as the average questionnaire length can be reduced.

4. Replicating

In projects where response rate is important, it might be useful to replicate sample and use a maximum amount of tries within a replicate before a new replicate is opened. By doing so it is possible to achieve a higher response rate while part of the sample remains untouched. However, at the same time, it might have an influence on the efficiency of the sample.

5. A/B Testing

In many cases, it is not known beforehand which approach for sampling yields the best possible outcome. Therefore, it is recommended to set quotas on the different sample sources. The same holds true for different sample providers. Various reasoning can be used for sampling an audience, nevertheless, only when you measure it, can you manage it and increase your incidence rate.

For instance, High-Net-Worth Individuals (HNWI): can be targeted based on postcode, lifestyle sample on income, lifestyle sample on secondary items (cars) or various other means. Beforehand it is not always possible to see which approach yields the best results, however, running a pilot with a split  RDD vs lifestyle sample can yield good insights.

6. Geocoding

In many consumer projects aspects such as rural/urban topology or size of the community (5000 people or more than 1,000,000 people), postcode, city or state are asked. Nevertheless, this kind of information can be precoded into the sample which saves money by reducing interviewing time – thus geocoding provides multiple advantages.

Additionally, statistical data for location and also demographics (income level, ethnographic variables etc.) can be appended in many cases.

To conclude…

Various means exist to improve CATI fieldwork efficiency by means of advanced consumer sample.

This article was posted in the category Sampling and contains the following keywords: B2C, CATI, Fieldwork, sampling, RDD Sample, Lifestyle Sample, Consumer Sample, dual-frame, B2C Sample, geocoding