R Tutorial: Surveys in marketing research

Want to learn more? Take the full course at https://learn.datacamp.com/courses/survey-and-measurement-development-in-r at your own pace. More than a video, you'll learn hands-on coding & quickly apply skills to your daily work. --- Survey says... welcome to the course. This course will walk you through each step of the survey development process with an emphasis on applications in R. In this first lesson, we'll sketch out the survey development process, along with introducing key terms that we'll use throughout the course. Let's get started! You've probably seen a survey like this before. This is a market research survey asking consumers to rate their attitude toward a set of statements. Looks pretty simple. There's actually a lot of cool statistics going on behind the scenes. Calculating those cool statistics in R is the focus of this course, but before we get there, let's cover some terms. The purpose of this survey is to measure brand reputation. The survey consists of nine items, which respondents rate on a "scale of one to five," but more formally this is known as a Likert scale. Likert scales don't have to be five-point, but that's the most common. A published Likert survey looks like it was developed pretty effortlessly, but there are a lot of steps involved in building a survey. Here's a helpful flowchart from a paper by Timothy Hinkin illustrating the entire survey development process. We will be referring to this flowchart throughout the course. First things first -- Step 1, item generation. Collecting data from busy consumers is difficult, so before we ask for their attention we want to assess the strength of our items. To do this, we'll enlist the opinions of subject matter experts. The first item generation diagnostic we'll look at is inter-rater reliability. There are several ways to do this. We could take a raw percentage agreement. Let's do that using the agree() function from package i-r-r. We have 50 percent raw agreement b