2 Sampling & Measuring Network Ties

Once we know what the aims are, and we’re ready to begin designing a new network study, there are a number of things that we have to address. In this chapter, I describe three such primary considerations—measurement, sampling and the “boundary specification problem.” This chapter will not be a recipe guide, providing precise directions on how your study should address each consideration. Instead I will describe the broad principles that govern approaches to those considerations, regardless of the type of study being conducted.

You can think of the task before us as if we were embarking on a series of cooking classes together. If we were going to learn how to bake, sauté, and grill a variety of dishes, we would not assume that each approach will follow the same procedures. But successfully engaging in either of these, each requires appropriately understanding how ingredients, temperature and time come together. How you combine these can alter the texture, taste, and appearance of each dish you make. And when all three principles are applied in harmony with any of these strategies, you can achieve a variety of desired results. In social network data collection, we will consider our “ingredients, temperature, and time” as:

  1. Tie Measurement - what will be measured about the relationship(s) among the studied population,

  2. Sampling Design - which of the relationships from which portion(s) of the population will be sampled from within the population, and

  3. Boundary Specification determining inclusion and exclusion criteria for the study, simultaneously as a dual consideration at the levels of nodes and ties.

These three principles must be addressed in all network studies, of any kind.

Just as baking, sautéing, or grilling would approach ingredients, time, and temperature in different ways, in the chapter that follows, I demonstrate how different data collection platforms and strategies can employ these principles in different ways. Because much of the terminology and many of the specifics of the strategies described here have been developed in the context of survey research (Marsden 1990, 2011), this chapter will lean heavily on survey-based examples. However, the key takeaway from this chapter should be that no matter which of the primary design approaches researchers choose to employ, (for a full discussion, see Chapter Chapter 3, if you want to optimize the utility of your study, questions of sampling, measurement, and boundary specification must be addressed in each of them. To illustrate this importance, I will draw examples from a number of studies, selected to include the range of design strategies represented later.