Statistics Course: MESA students learn how to test a hypothesis and approach data with multiple variables

Statistics helps us make sense of large amounts of information by summarizing and drawing conclusions from data. It involves techniques for gathering data, organizing it into meaningful patterns, and using mathematical tools to make well-informed decisions.

In the online Measurement, Evaluation, Statistics and Assessment programs at UIC, students are introduced to a multivariate statistical analysis technique in the EPSY 587: Structural Equation Modeling course. Students become well versed in current data analytic techniques and kept abreast of emergent techniques by being aware of contemporary methodological literature.

Dr. Ting Dai, an Assistant Professor in the UIC Department of Educational Psychology, admits the EPSY 587 course can be challenging, but insists the techniques are beneficial to a MESA student’s overall educational journey. She’s never had any students say that the course “wasn’t worth their time.”

Learn more about the EPSY 587 course, and the importance of learning statistics.

Why is it important for MESA students to learn about statistics? 

Statistics is essential in quantitative research method. A lot of people think it’s just about numbers or graphs, but statistics is so much more than that. It goes beyond the numeric data participants provide and the statistical results we obtain from the analyses—it plays an important role maybe even at the beginning of when a research question occurs, because it determines how you approach your research questions. For example, I want to know whether there’s a relationship between people’s heights and people’s shoe sizes. Although it’s a silly question, I can think of multiple ways to approach it. One of the approaches would be involving statistics where you go out, get a large sample, and ask for individuals’ shoe sizes and heights. You’ll use methods and models to see if those two variables are correlated and what is the strength of the relationship. If you choose a quantitative pathway, then statistics will determine your scientific inquiry process very early on.

From a non-researcher standpoint, if we lack statistical knowledge, it is hard to understand information that is presented to us and we begin to ask, ‘what do I trust, what do I believe, and what do I even need to know?’ Even with our simple example of the correlation between shoe sizes and heights, not having the basic statistic knowledge makes us wonder what it means what I see a specific positive or negative correaltion coefficient, e.g., r = 0.15.

Can you explain why your EPSY 587: Structural Equation Modeling course is beneficial for MESA students? 

In the EPSY 587 course, we aim to understand the relationships between different multiple variables. Structural equation modeling (SEM) is one of the few multivariate modeling approaches where you simultaneously look at multiple variables rather than just one target variable, or what we call the dependent variable. When we have a lot of dependent variables and a lot of predictors that we conceptualize, we want to model a web of relations that we can assess simultaneously.

We’re often dealing with multiple variables that interplay with one another. For example, you want to understand your students’ GPA growth. You’re looking at GPAs from the first semester to the second semester plus the following school year, so we have 4 GPAs at hand. How do you depict the changes that are happening? Simultaneously, you’re thinking about all of your student’s attendance, engagement, and behavior variables that could have impacted their GPAs. You may be curious about the relationships between those variables and the GPA growth. That’s where you specify SEM models to look at those relationships.

What is structural equation models, and why is it important for students to learn this analysis technique? 

Structure equation models (SEM) are a way to look at multiple variables and the interrelationships among variables as conceptualized based on prior research and theory. Creating a SEM model helps you understand whether a conceptualization fits the reality of what your study data is telling you.

To me, structural equation modeling is not just a statistical method, but also is a way of thinking that opens up your mind in terms of how to approach educational research and psychological research. Students will start to learn how things are connected in real-world situations. Problems that we are faced with in the real-world are not in the vacuum of a lab. We can’t control variables or possible triggers for a variable as much as we want to. We need to look at a lot of variables that are intricate and are interconnected and SEMs help us to do that.

Are there any real-world examples or projects students can look forward to in this class? 

Yes, I believe in nothing else but student-motivated problems for helping them connect with what they learn in class. At the very beginning of the class, I tell students that their final project is going to be a project where they will utilize structural equation modeling, but the substantive research topic will be chosen by themselves and derive from work or research interest. The data can be from a student’s research, work, life, or any other public data that is available to them. I want them to do something that they’re interested in and actively seek answers to. Once their final project is complete, I encourage them to submit it as a proposal to research conferences such as the annual meetings of American Educational Research Association, or as a work report to their boss.

Why do you recommend students take EPSY 587? 

The EPSY 587 course requires a very high level of engagement and has a very steep learning curve, but I make sure, from the first week of the semester and onward, that I provide enough support beyond the lecture and instructional materials, such as hands-on worksheets, demonstration of problem solving, timely and individualized feedback on assignments, one-on-one consultations, group office hours, and twice-a-week check-ins. I also encourage students to find study partner to learn SEM collaboratively. I also acknowledge the difficulty of the materials, share my own SEM learning journey, and provide plenty of homework revision opportunities to emphasize an achievement goal structure of “mastery” rather than “performance.” I’ve never had students say to me that the course wasn’t worth their time. You’ll start to think about statistics from a different angle and from a big picture perspective. I tell them that what they will learn can be difficult but will be very cool. The ability to say that you know how to do SEMs on your CV is enough of an incentive.