Hello again. Okay. You just beat out a competitor and got a huge project that requires interviewing parents of foster care children to find out how the process of becoming a foster parent works. This project requires 600 completed interviews in six weeks. That is a short timeline. Now what do you do? Well, the short answer is, you will need to make sure you are selecting the optimal sample source for your survey. So after this lesson, you will be able to identify appropriate sample sources and be able to plan your strategy for completing these interviews within that timeline. You will also be able to figure out how to collaborate with other companies in finding ideal samples. Now, before you can do all that, let's clarify some terminology. In quantitative market research, a sample source is how you go about finding the ideal sample. For example, there are companies that have lists of individuals broken down by demographics or other categories that you can buy as your sample. Here's a list of potential sample sources. There is convenient sampling. It's a type of non-probability sampling where you get participants wherever you can find them. There are no inclusion criteria. This would be good for stopping random people on the street to see if they would be willing to be interviewed, as better known as passerby surveys. This type of sampling worked beautifully one time when I had to interview 65-year-olds and older about their opinions on advanced directives. I could tell visually if they were old and if they seemed open to being interviewed. I went to locations where this age group frequents, such as the grocery store and the mall. You may also want a random sample list for your sample source. You can either generate random digit dialing or random census addresses by zip code, or you can buy lists that have been put together by formal survey houses. Online survey houses, such as Qualtrics and Survey Monkey, can provide lists made to whatever specification you desire. For a fee, certain associations or large research firms such as Pew Research may provide lists of their members. Basically, anything you can join, including Facebook, Amazon, TV news services, various computer and touchscreen applications, can sell your name and specific demographics to use for a sample. These are steps I use for selecting a sample. One, discuss and finalized target market with the client. Two, see if the client can provide a list of their customers. If so, you can do a random selection from that list. Three, if your client is not supplying the list, there are different methods to apply depending on the type of survey. There is phone; random digit dialing, mail, random census tract by zip code, physician, association lists, email or online, online survey tools such as Qualtrics, Survey Monkey or BrainJuicer. Four, make sure the list you receive is accurate and complete. Random digit dialing, RDD, has the advantage that includes unlisted numbers that would be missed if the numbers were selected from a phone directory. It is also used for election polling or experimental survey research. Step one in the process in establishing a random phone survey by random digit dialing would be, one, begin with all of the area codes in the US and then identify all the exchanges within each area code. Two, use a computer that is programmed to dial a four-digit random sequence of numbers which is added to each combination of area code and telephone exchange. Three, the actual number of respondents in any code is determined by the actual number of telephone numbers assigned in the geographic region for which the area code is used. Four, it will usually take at least five calls to get one working number that is not disconnected or a business. Five, to reduce unusable phone numbers, you may want to use what is called the Walksberg Method, which takes advantage of the fact that telephone companies tend to hand out telephone numbers in sequence rather than randomly. As a result, once a number is found to be a wrong number, nearby numbers are likely to be wrong numbers too. Step two in this process is to mail random census tract by zip code. You can either buy a list online through a variety of services or you can do a Google search. To do this, you would go into the web and Google Maps. The entire US map will come up. Tap on the city you wish to get addresses. Bring the map in for closer view. Hit each dot for a viable address. Step three in this process is to ask associations for a list of members for harder-to-reach people such as surgeons. To do this, you would call the association, ask for the marketing director and ask if you could purchase a list. Step four relates to the process of doing an online interview. To do this, you would call some of the most respected online data research houses, like Qualtrics, Survey Monkey and BrainJuicer. Ask them how much of a cost to purchase their mailing list versus having them conduct the survey. Decide which alternative best fits your need. Either purchase their list or let them do most of the work for you. Now that you have identified appropriate sample sources, you need to choose the optimal one. Here is how you make that choice with confidence. Make sure you understand which target market you want and the specifics. So, in the case of the foster care parents, we already know the target market, but how do we get the list? What I did was suggest one-on-one interviews with both state and county foster care agency caseload workers who then supplied a random list of parents they worked with, who had foster children. This ended up adding another segment of interviews to an already large project, which turned out enhancing the relationship the client had with these agencies. Another example of where I had to decide on an appropriate sample source to conduct a random survey about whether YouTube videos about smokeless tobacco products influenced the purchase behavior of young adults. My client wanted a random survey but each person had to meet these difficult screening criteria. At first, I thought I could look on YouTube under smokeless tobacco videos and merely put up questions via a YouTube video with an intriguing title that might draw smokeless tobacco users in. When I tested out the idea on YouTube, I found that unless there is an interesting video to accompany the questionnaire, people didn't really pay attention. My team instead consulted Qualtrics and found out that they have a panel of consumers who fell within the specific age range, were active on social media, and were smokers. In other words, the panel list that Qualtrics met had many of the criteria of the YouTube audience I was looking to find. The only aspect I didn't know from the start was whether the chosen panel were also smokeless tobacco users. This turned out to be a good screening question which led to other answers about product advertising for cigarettes and smokeless tobacco. So you can see why it is important to identify the appropriate sample sources and plan your strategies to complete your project within the given timeframe while collaborating with other companies to find out the ideal sampling. Until next time.
