Sampling is one of the most important items in having a successful market research project and accurate results. Recently, we had the chance to ask an expert about all things sampling. Rudly Raphael is the President of qSample and he took some time out of his busy schedule to answer some of the most trending questions on sampling, representativeness, managing panels and how to conduct some quality control. Rudly is a Harvard University Grad in Information Sciences and has been working with market research, specifically panel management for the past 15 years. Today, qSample steps in and offers over 10 specialty and well taken care of panels to hit research targets for companies like us, The Ivy League Magazine Network, TruRx and many more across different industries and studies. Whether you are managing your own sample or using third party sampling for your research projects - these are some great insights from a field expert to consider. Continue reading to see the full Q&A with Rudly.
Q1: How did you originally get into sampling?
Rudly: By definition, a sample is a subset of a population. I always had a keen interest working with databases since I can remember and that interest made it easier to transition to my first market research job where I was responsible for developing an in-house panel tool, which ultimately led to me delving more into the sampling business.
Q2: What are some guidelines for making sure you have a representative sample?
Rudly: Working with a panel vendor that has a well recruited panel helps. However, the provider can ensure this by applying what we call in the industry "quotas" to key demographic survey questions such as age, gender, income to make sure the data is representative of the population they're surveying.
Q3: How do you go about sampling in hard to reach groups and demographics? (i.e. males under 30)
Rudly: This is certainly a challenged faced by most panel providers, but it is something that is often addressed at the point of recruitment. There must be some sound strategies put in place even post recruitment - proper incentives, survey invitations that fit panel member's lifestyles, understanding their behaviors while on the panel, etc.
Q4: What's a good and realistically achievable sample size to aim for?
Rudly: It depends. Your population has a lot to do with it. For a national survey, you may want to aim for a sample size of 1000 or more. However, the key to good research is not necessarily the sample size only, but how representative is the sample. It's useless if we survey a large number of people that's not representative of our target market. Depending on the population, your sample size can be as small as 100.
Q5: What are some ways to assess the quality of your panel?
Rudly: There are a number of things we can look at such as good response rate, good qualification rate, low drop out rates from survey and how often members are unsubscribing from the panel - otherwise known as attrition.
Q6: What are some ways to make sure your data and results are of good quality?
Rudly: I guess this should start at the very beginning (recruitment) to effective panel management practices. Survey data quality has a direct correlation to how well a panel is maintained. Does the provider weed out bad panelists or set participation controls or identify over aggressive panel members? Yes. Most members like to be incentivized, but are they leaving you VM inquiring about the frequency to which they can participate? If so, that's a problem and will undoubtedly show in your data file.
Q7: How much does a common market researcher spend on sample alone per project?
Rudly: It depends on how much business they generate. However, I can say that sample can account for 40% or more of the total research project cost.
Q8: Can you share some pitfalls or red flags for people to take note of when shopping around for sample?
Rudly: This is the business where you really don't want to go with the lowest bidder. Supply the provider with all the necessary project details when requesting a bid. If there are anomalies in pricing, there's probably a reason for it and you should inquire. Sample is typically priced based on the incidence of a population. Knowing this information or an accurate estimation may be helpful.