Have you ever wondered why or how researchers categorize people into different groups? Of course, there are specific types of groups you might first think of, such as differentiating people by race, gender, income level, marital status, and education. However, there are a variety of other types of groups that researchers might find interesting that you may not be aware of. For instance, some researchers are interested in studying people based on patterns of their activities over a span of time. An example of this would be if a researcher was interested in studying the differences between types of offenders. For example, each time they were: 1) expelled or suspended from school, 2) arrested, 3) convicted, and 4) incarcerated. There are other variables a researcher might also find to be important: periods of 1) homelessness, 2) unemployment, 3) graduation from high school and 4) marriage.

This might seem like a very daunting task: to categorize people based on all the possible sequences of events that happened in their lives. And it is. However, researchers have tried to find more efficient methods of categorizing people into groups. One relatively newer method of determining groups uses a computer program to examine how many different sequences of events across time exist in a sample of subjects. This method is called sequence analysis. Sequence analysis for social science research had its origins in the study of evolutionary patterns and modes of descent of DNA molecules, initially (and still) carried out by biologists. However, social scientists have co-opted this method for social sequence analysis, in which they study patterns of events. The use of this term in this blog post, therefore, refers specifically to identifying patterns of events based on their classification into different categories and the specific ordering of the events. [1]

So how does this technique work? Very simply, in fact. There are two main strategies. In the first one, a researcher can ask the program to find all possible sequences of events and to report how many subjects fall into each “group.” While this may be very interesting, the program could come up with a lot of sequences of varying importance to the study. For example, if a researcher is interested in studying self-care habits and the timing of a person brushing their teeth: If the researcher assumes people will most likely brush their teeth either 1) before they eat breakfast or 2) after they eat breakfast. Maybe the program identifies a sequence that is not very enlightening, such as one person out of 100 subjects doesn’t eat breakfast at all. Overall, this is useful if the researcher knows little or nothing about the population of people he/she is studying. By categorizing subjects into a number of groups allows the researcher a starting place, to find out how and why groups are different.

The second strategy, and what tends to be more useful but requires some previous research on the type of people the researcher is interested in, is to identify sequences of events that govern human behavior so that the program can categorize people into different groups based on those sequences. An example of how to use this technique used in the field of criminal justice would entail identifying patterns of criminal offending. Let’s say there are two types of offenders: those that start committing crime early and those who wait until later in life to become criminally active. The researcher could tell the program to look for criminological events (i.e., expulsion/suspension, arrest, and incarceration), that occur either before or after a particular time in a subject’s life (e.g., high school graduation). The program would then categorize the subjects into these two different groups. From here, the researcher could look at other interesting categorical differences among the groups. For example, are subjects who commit crimes earlier in life more likely to be male or female? Or do those who commit crimes later in life do so as a result of an experience such as unemployment or homelessness?

Pathway 1 (Early-onset Offending):

Unemployment → Homeless → Criminal justice involvement → High School Graduation

Pathway 2 (Late-onset Offending):

High School Graduation → Drug use →  Criminal justice involvement  

In the aforementioned example, the various sequences of events – or type of offenders – could help researchers to understand why some people commit crime young and others commit crime when they are older. These findings could have a very significant impact on the criminal justice system. For example, stakeholders such as police, probation and parole officers, judges, prison officials, and many more criminal justice actors could use this information to allocate financial resources. Categorizing subjects into different groups allows the researcher to study how they differ and then perhaps explain different types of offending patterns, in turn, this could even help predict criminal behavior and allow targeted action to reduce said criminal behavior.

Edited by Fabio Mendes, Guillaume J. Dury, and Briana K. Whitaker

[1]  Cornwell, B. (2015). Social sequence analysis: Methods and applications (structural sequence analysis in the social sciences). New York, NY: Cambridge University Press.