3/15/26, 10:14 AM myMinerva
Academics > HC Resources > HC Handbook > #interventionalstudy
HC HANDBOOK
#interventionalstudy
Design and interpret experimental studies.
Please fill out this Google Form to submit suggestions for this page (i.e.
typos, study guide contributions, useful applications, etc.). We are
actively seeking feedback on anything big or small to make the HC
Handbook as useful as possible.
Defining the HC
Pitch
If you really want to connect cause and effect, sometimes you just have to do the
experiment.
Paragraph Description
A true experiment has one or more independent variables, which are
manipulated or selected independently, and one or more dependent variables,
which are measures of the effects of variations of the independent variable or
variables. An experiment relies on assigning samples to a treatment condition
(an intervention) and tries to minimize (or, ideally, eliminate) the influences of
extraneous variables. To determine that the treatment causes the outcome,
experimenters must control extraneous variables and other factors that may
impact the outcome. In contrast to observational studies, experiments always
involve interventions.
[Link] 1/9
3/15/26, 10:14 AM myMinerva
Categorization
Foundational Concept | Empirical Analysis | Thinking Creatively | Applying
Research Methods
Cornerstone Introduction
Class: Empirical Analysis | Semester One
Unit: Scientific Method
Big Question: How do we mitigate human-caused climate change?
General Example
You are a researcher studying bacterial resistance to antibiotics, and your job is
to find out the best way to administer antibiotics in order to prevent resistance to
antibiotics. Your hypothesis is that administering three different antibiotics
simultaneously is more effective than administering the same three antibiotics
cyclically (one after the other), which in turn is more effective than administering
only your strongest antibiotic. To address this hypothesis, you consider enlisting
the help of hospitals and having doctors prescribe one of the three methods you
are investigating to patients. You consider that there are two important issues
with this experiment design: first, since you already have good indication that
simultaneous antibiotic delivery is much better for preventing bacterial
resistance, it would be unethical to deny people the best treatment. Second, it
would be very hard to control relevant variables like type of bacteria, location of
infection, method of antibiotic application, and others. Whereas you recognize
that testing antibiotics on humans will have to be done eventually, for now you
decide to run lab-based experiments on bacterial cultures, using three different
antibiotics. You decide to have three intervention groups: one in which bacteria
are attacked only with your strongest antibiotic during 30 days, one in which
[Link] 2/9
3/15/26, 10:14 AM myMinerva
bacteria are attacked with three antibiotics for 10 days each, and one in which
bacteria are attacked throughout the 30 days with all 3 antibiotics
simultaneously. To account for the possibility that the injection of liquid into a
bacterial culture has an effect on bacterial survival, you add a “placebo” group to
the experiment, wherein a harmless liquid will be applied to bacteria.
Furthermore, you add a “natural history” group of bacteria who will suffer no
treatment to serve as a comparison, a baseline against the placebo and
intervention groups. Finally, you are sure to control for variables like availability
of sustenance, incidence of sunlight, temperature, moisture, and chemicals the
bacteria are exposed to. You keep your four bacteria groups inside a sterilized
lab, in petri dishes, to control for these variables. Because of the tightly
controlled design of your experiment, you recognize that there may be some
issues with generalizability of your results. You also consider that bacteria
inhabiting living beings may react very differently to treatment, and so you
suggest that, for a future experiment, you should try to kill bacteria infecting
mice rather than “infecting” petri dishes. However, you are confident that the
experiment can potentially reveal causal relationships.
Footnote: An interventional study design is proposed in order to test the
suggested hypothesis. Alternative study designs are discussed, and arguments
favoring the chosen experiment are presented. The study design is described,
including what the treatment groups will be and what treatments they will
receive. Controls are also mentioned. Finally, the constraints of this study
design are mentioned and a further study is suggested.
Is it this HC?
Is #interventionalstudy the right HC? If the answer to the following questions is
yes, #interventionalstudy could be useful:
[Link] 3/9
3/15/26, 10:14 AM myMinerva
1. Do you need to design a study that involves some kind of intervention
(treatment) to then observe outcomes?
2. Are you critiquing someone else’s interventional study design?
Depending on the work product, there might be other HCs to consider.
The following HC might apply if the
work:
#observationalstudy designs or critiques a research study
that does not involve researcher
intervention.
#casestudy focuses on a single sample (or few)
from the population. Case studies can
use an interventional design approach.
#hypothesisdevelopment generates a testable hypothesis that
explains the reasons or causal
mechanisms for underlying patterns or
observations. Hypothesis development
is a step prior to designing a study.
Depending on the specific research
question, one might then design an
interventional study.
#sampling provides detailed explanation of a
chosen sampling method and justifies
the choice. Note that sampling is a key
[Link] 4/9
3/15/26, 10:14 AM myMinerva
part of study design and
#interventionalstudy may be more
appropriate if you are referring to the
holistic study design.
#comparisongroups provides more elaborate justification
for the chosen comparison groups.
Simply listing the control and
alternative comparison groups as part
of a study design falls under
#interventionalstudy.
#confidenceintervals, Conducts statistical analyses on
#correlation, quantitative survey data. These HCs are
#descriptivestats, usually used after data collection.
#distributions, #regression,
#significance, #variables
Self Study
Try answering the questions on your own before checking the example answers.
👆 What is the difference between quasi-experimental and RCT
studies and how can you decide which one to use?
Answer: Quasi-experiments and RCT (Randomized Controlled Trials) are
both interventional studies. However, unlike RCT, the treatment assignment
[Link] 5/9
3/15/26, 10:14 AM myMinerva
is not random. While RCT is often the best at evaluating the outcome of an
intervention, randomization to treatment is sometimes not possible or
ethical. For example, when exploring the effect of hand-sanitizer usage in a
hospital, it is not practical or reasonable to ask some nurses to not use hand
sanitizer. Thus, a pre-post test design might be used instead. Note that there
are multiple types of quasi-experiments that you can learn about from the
quasi-experiment reading linked in the Key Resources.
👆 What are some measures you can take to minimize the effect
of bias and confounding variables when designing an
interventional study?
Answer: You can mitigate bias (e.g., confirmation bias) associated with the
researcher and participants by using double-blinding (all unaware of the
treatment assignment). In the case of non-human subjects (e.g., rocks), you
can still mitigate confirmation bias by blinding the researcher. Confounding
variables can be kept the same for all groups (e.g., by using the exact same
type of soil on study plots). You can also account for the confounding
variables using matching, which involves pairing participants with similar
characteristics from the treatment and control groups and comparing the
effects of the treatment between similar participants.
👆 How can you ensure that your study design tests the
hypothesis?
[Link] 6/9
3/15/26, 10:14 AM myMinerva
Answer: Before you start designing a study, you need to make sure that your
hypothesis is testable (review #testibility HC entry for more information) and
generate multiple predictions among which you need to choose one (or even
better, two) that would support/refute the hypothesis most convincingly.
Next, you need to select variables that would allow you to observe the
predictions. However, the research design process is not always linear and
you may need to go back and forth between the design and the hypothesis to
adjust it based on feasibility.
Applying the HC
Guided Reflection
Have you provided an in-depth explanation of why you chose an
interventional study?
Have you provided a detailed description of the study and backed up the
design choices with evidence (e.g., methods used in similar studies)?
Have you accounted for (or at least mentioned) extraneous variables and
cognitive biases?
Have you evaluated the study design by outlining its strengths and
limitations?
Common Pitfalls
The study type is inaccurately categorized.
The description of the study is superficial.
The study design would not effectively test the hypothesis.
The study design choices are not justified or explained clearly.
Applying the HC at Minerva
[Link] 7/9
3/15/26, 10:14 AM myMinerva
You will use this HC in NS assignments where you are asked to write a
research proposal or a critique of a primary paper.
You can also use #interventionalstudy in business assignments to test
different marketing strategies and products on the consumers.
Applying the HC Outside of Minerva
You can use this HC to critique research papers. For example, let’s say you
want to evaluate the effectiveness of a fiscal policy proposed by the local
government. You found a research paper about the implementation of the
policy as a pilot program, whereby two neighborhoods were compared. One
neighborhood received the policy implementation, while the other served
as a control group. The aim was to evaluate the impact and outcomes of the
policy within the context of the local community. Your background in study
design can help you assess whether the design choices were well justified
(e.g., are the neighborhoods selected sufficiently similar?).
You have observed that your toddler's attitude towards chicken noodle
soup varies, as there are days when they enjoy eating it, while on other
occasions, they reject it. You suspected that it is because of a spice you are
using in the recipe from time to time. In order to test your hypothesis, you
plan to conduct an interventional study where you will control for all
ingredients and add the spice as a treatment in your next batch of chicken
noodle soup. Then you will observe the toddler’s reaction.
Key Resources
Lesh, M. (2015). Observational vs. experimental study. [Video]. YouTube.
[Link]
This short video describes the key differences between observational
and interventional (experimental) studies.
[Link] 8/9
3/15/26, 10:14 AM myMinerva
Lacey, M. (1997). Experimental design. Statistics 101-103: Summaries of
topics. Yale University. [Link]
98/101/[Link]
This summary provides a succinct overview of experimental design
and introduces the major aspects of design that should be considered
when developing or critiquing an experiment.
Sibbald, B., & Roland, M. (1998). Understanding controlled trials: Why are
randomised controlled trials important? British Medical Journal,
316(7126), 201.
[Link]
This brief, one-page reading provides an overview of the key elements
of randomized controlled trials (RCTs), their appropriate use, and
primary constraints.
Harris, A. D., Bradham, D. D., Baumgarten, M., Zuckerman, I. H., Fink, J.
C., & Perencevich, E. N. (2004). The use and interpretation of quasi-
experimental studies in infectious diseases. Clinical Infectious Diseases,
38(11), 1586-1591.
[Link]
This review provides an overview of quasi-experimental designs in
clinical research.
Kosslyn, S. M., and Rosenberg, R. S. (2001). QALMRI instructions.
Adapted from Psychology: The Brain, The Person, The World. Boston:
Allyn & Bacon. [Link]
This reading suggests an approach to primary peer-reviewed
literature - QALMRI.
[Link] 9/9