158 • ii: MindFulnESS in EduCATing FOR SElF-REgulATiOn And EngAgEMEnT
for the teachers who are already time constrained and challenged with heavy workloads
who added research to their load; and to the students and their families who agreed to add
research to their commitments. To participate in research and help researchers conduct a
high-quality study is a generous act of civic philanthropy. It is a gift to future school dis-
tricts, schools, teachers, students, families, and generations. I am so grateful for each and
every school who has worked hard to support my team’s research and for all the schools
who help us better understand mindfulness and yoga in the schools.
There are other options that can yield results to help inform what we do in schools.
There are some RCTs that do not randomly assign individual students to the various groups;
rather they assign whole classes or schools randomly to the treatment group, active control
group, or school as usual control group. This process carries with it some complications as
it is difficult to sort out the effects of the classroom teacher, or the school environment, or
the cohort of students in a school or classroom and how they already function together in
a group. You probably remember being in a class or group that didn’t work well together
or a class that did and the teacher saying to you, “This is the best class I have ever had.”
Sometimes, these sorts of influences can affect study data.
Another option is a controlled trial. Controlled trials are like RCTs in that there is a com-
parison group. However, who is in the intervention group and who is in the other groups
are not randomly decided. In order to make sure that the groups are equal, researchers
sometimes match the control and treatment groups according to variables such as socio-
economic status, race and ethnicity, and age. This way, even though participants were not
randomly assigned, researchers can analyze data based on inferences that the groups were
comparable in many ways. This is called a matched controlled trial.
Sometimes, even with the best-laid plans, researchers are not able to secure control
groups or enough participants for an active control group (e.g., a group doing calisthenics
or something else physically active) or a school-as-usual control group (i.e., doing noth-
ing special for the study). In such situations, researchers create a simpler design with only
one control group, or do a study with the intervention-only group. These studies are less
compelling as we cannot be sure that the intervention yielded the results because there is
no group to compare the results against. It could be that the intervention group, and all of
the kids in the school, began working harder because of some other variable: Maybe the
school shifted the schedule around so all of the students were able to sleep in longer in the
morning. Maybe the school rolled out iPads for everyone that year. Or maybe there was a
dramatic event (e.g., a student killed in a car accident or the death of a beloved teacher) that
affected all of the students and the researchers’ intervention group got worse as did every-
one else. In the latter case, it looks as if the intervention had a negative effect, rather than the
school-wide or community-wide stressors.
When a field of research is new or evolving in an important way, researchers engage
in correlational studies to better understand how variables might be associated and affect
each other. This type of research is not considered causal. That is, we cannot infer that just
because two things tend to correlate together (e.g., mindfulness and fewer behavior prob-
lems) that one causes the other. An example is a child who is maturing neurodevelopmen-
tally who has been placed in a healthier, more nurturing foster home, and is now also better
able to sit for meditation class. Note, however, that there are some ways to analyze the data
to understand the relations among variables in a causal manner. For example, in a study
my research team recently conducted, we collected data across 14 different middle schools
all in one semester. We aggregated and cleaned the data and analyzed the variables to see