The Purpose and Function of Control Groups, Explained

UPI Study › Courses › Abnormal Psychology › Free lessons › Lesson 2.2 Free with a free account

The Purpose and Function of Control Groups

Abnormal Psychology · Chapter 2: Investigating Mental Health: Research FoundationsSection 2.2 · lesson 7 of 9716 min · free with a free account

← The Science Behind Clinical Research and Correlational MethodsNext lesson: Boundaries and Shortcomings of Clinical Studies →

Earn 3 college credits for Abnormal PsychologyRecommended for credit by NCCRS · Psychology 180 · self-paced · one payment, lifetime access
Enroll in Abnormal Psychology · $250 →Download the syllabus (PDF)

In this lesson

Why improvement without a comparison is uninterpretable, not merely uncertain.

  • Explain the counterfactual a control group stands in for
  • Separate random assignment from random selection
  • Choose a control type for the objection you intend to answer
  • Name the five artifacts that all bias claims the same way

Transcript

Everything said in the video, section by section (2,247 words). Read along, or search it.

Investigating Mental Health: Research Foundations

The Purpose and Function of Control Groups

A treatment is given. People get better. The treatment worked. That inference is wrong, and this section is about exactly why. We start with the question a control group actually answers. It is a question about a world that does not exist. Then the machinery. Random assignment, and what it uniquely achieves. Four kinds of control, each ruling out different things at different ethical cost. Blinding, and what you do when blinding is simply impossible. Then five artifacts that all push in the same direction. Every one of them flatters an ineffective treatment. And we finish by taking an uncontrolled claim apart, objection by objection.

Before You Begin

Every Trial Here Is a Composite

Two notes before we begin. This material is educational, and not a basis for diagnosing anyone. And every trial described here is a composite constructed for instruction. No real trial, treatment, or participant appears anywhere in this section. Here are three questions to bring to any treatment claim you meet. First, was there a comparison group at all? Second, what was the comparison? Beating a waitlist is not beating an established treatment. Both get reported in the same enthusiastic language. Third, how was allocation decided? If participants or clinicians chose the group, the comparison is observational.

The Counterfactual

The Question a Control Group Answers

Suppose sixty participants receive a described new treatment for eight weeks. At the end, their average symptom score has fallen substantially. It is natural to conclude that the treatment worked. The conclusion does not follow, and seeing exactly why is the foundation of everything else. Think about what the claim actually says. To say the treatment produced the improvement is to say that without it, it would not have happened. That is a statement about a world which does not exist. The same sixty people, over the same eight weeks, without the treatment. This is the counterfactual. The outcome that would have occurred had the intervention not been given. And it can never be observed directly. The same people cannot both receive and not receive a treatment. So a control group is a comparable group who do not receive the intervention. Included so that their outcome can stand in for that counterfactual. It is the only available answer to the question, what would have happened anyway. Without it, the improvement is uninterpretable. Not wrong. Uninterpretable.

Variables

Read the Definition Before the Result

A treatment study has two central variables. The independent variable is the thing the researcher controls and varies. Here, whether a participant receives the treatment or the comparison condition. The dependent variable is the outcome measured, to see whether it changed. The symptom score, the relapse rate, the number of days worked. The dependent variable depends on what the researcher did. Both need an operational definition. A statement of exactly how the variable is produced or measured in this study. For the treatment, that means enough detail for someone else to deliver it. How many sessions, of what length, delivered by whom, with what training. For the outcome, it means naming the instrument, the timing, and the threshold. Improvement can be defined in at least four ways that will not agree. A change in average score across the group. The proportion of participants crossing a stated threshold. The proportion no longer meeting diagnostic criteria. Or a change in functioning, such as a return to work. A treatment can look effective on one of these and unimpressive on another.

Two Different Randoms

Assignment Is About Comparing, Selection About Generalizing

Random assignment is allocation to conditions by chance alone. No characteristic of the participant influences which group they enter. And no judgment by the researcher influences it either. Because allocation is decided by chance, the two groups will on average be equivalent. Equivalent on everything. Age, severity, duration, motivation, income, social support. And every factor that nobody thought to measure. The third-variable problem could only be handled by measuring confounders one at a time. Which requires knowing what they are in the first place. Random assignment handles the unknown ones too, and no observational design can do that. The equivalence is probabilistic rather than guaranteed, so it holds better in larger samples. Which is why researchers report the characteristics of both groups at the start. Now a different procedure, answering a different question. Random selection. That is drawing participants from a defined population by chance. So that the sample resembles the population it came from. Assignment concerns comparing. Selection concerns generalizing. Most treatment trials use assignment without selection. Participants are self-selected volunteers, allocated by chance once enrolled. Such a trial supports a strong causal claim about its participants, and a weak one about anyone else.

Four Controls

The More It Rules Out, the More It Costs

Control group names a role, not a single design. Which comparison is appropriate depends on what needs ruling out. And on what can ethically be offered. No one type is universally correct. The untreated control receives nothing. It rules out the natural course of the condition, and the passage of time. It leaves open everything to do with expectation and attention. Participants know they received nothing, and the treated group know they received something. Its ethical cost is the highest of the four. The waitlist control receives the treatment after the trial ends. It rules out the same things, while ensuring nobody is denied treatment permanently. It still leaves expectation open, and it carries one subtler problem. Participants who know help is scheduled may postpone other efforts to help themselves. That would depress the control arm, and make the treatment look better than it is. The placebo control receives an inactive intervention indistinguishable from the real one. It rules out expectation, clinician attention, and the effect of being in a trial at all. Its ethical cost is serious, and it sits uneasily beside fully informed consent. The usual resolution is telling participants in advance that they may receive an inactive condition. The active comparison receives an existing treatment already known to work. It answers the question clinicians actually face. Is the new treatment better than what we do now? Its cost is interpretive. If the two perform equally, the finding is ambiguous. Both work equally well, or neither did anything in this sample. A third arm would settle it.

Blinding

What to Do When Blinding Is Impossible

Blinding is keeping people in a trial unaware of which condition a participant is in. In single blinding, the participants do not know. In double blinding, neither the participants nor those delivering and assessing know. Participants who know they received the real treatment report improvement partly because they expect it. Clinicians who know which group a participant is in behave differently in dozens of small ways. And assessors who know the allocation score ambiguous responses in the expected direction. That last one matters most, because outcome assessment usually involves judgment. Blinding is achievable in a medication trial. An inactive tablet can be made to look, taste, and be administered identically. It is not achievable in the same way in a psychotherapy trial, for an obvious reason. Someone receiving twelve sessions of structured conversation knows they are receiving a treatment. And the therapist certainly knows which one they are delivering. Assessor blinding first. Outcomes rated by someone who did not deliver the treatment. That is the single most important protection available in psychotherapy trials. Then an attention-matched comparison, giving the same contact and time without the active ingredient. Then fidelity checks, with sessions recorded and rated against a manual by independent raters. And finally, ask participants at the end which condition they think they were in.

The Dividing Line

Locate the Point of Assignment

It helps to picture a controlled trial as a line of stages. Recruitment first. Volunteers enroll. Then screening. Eligibility criteria are applied, and baseline characteristics recorded. Then the point that matters. Random assignment. Then delivery, which now happens separately in two arms. And finally assessment, using the same measures on the same schedule in both arms. Now notice where the dividing line falls. Recruitment and eligibility are shared by everyone in the trial. Every stage after assignment is done separately, in two arms. Anything happening before assignment cannot explain a difference between the groups. Both groups experienced it, so it cannot have produced a gap between them. Anything that differs after assignment is either the treatment, or an artifact. So when you evaluate a trial, locate the point of assignment. Then ask what differed afterwards. That is most of the work.

Five Artifacts

Every One of Them Flatters the Treatment

Five recur. Each would produce apparent improvement with no treatment effect whatever. Regression to the mean is the tendency for an extreme measurement to be followed by a milder one. Purely because extreme scores partly reflect temporary factors. People enter treatment when they are at their worst. Which is precisely when the next measurement is likely to be better, regardless. Spontaneous remission is improvement that occurs without any treatment at all. Many conditions improve on their own in a substantial proportion of people. A treatment given during that period will appear to have caused it. Expectancy is improvement produced by the belief that one is being helped. That is the placebo effect, operating inside the trial. Natural course is the ordinary trajectory of the condition, which is often episodic. Someone recruited during a bad episode will often improve as the episode ends. And repeated measurement effects are changes produced by the act of measuring. Participants become familiar with a questionnaire and answer it differently the fourth time. Being asked weekly about a behavior tends to change the behavior. Now notice something about all five. They push in the same direction. Every one makes an uncontrolled treatment look effective. None of them makes an effective treatment look useless. So uncontrolled claims are not merely uncertain. They are biased in a known direction.

Which Control Fixes Which

Only One Row Needs a Placebo

So which control removes which artifact? The mapping is not one to one. Regression to the mean is handled by any control recruited on the same criteria. The same regression occurs in both arms, so it cancels. Spontaneous remission needs an untreated or waitlist control. That shows you how many people recover anyway over the same period. Natural course is the same. Follow a comparable group over the same weeks. Repeated measurement is handled by any control measured on the same schedule. Now the exception, and it is the important one. Expectancy. Expectancy is eliminated only by a placebo control. A waitlist or untreated control does not touch it. So four of the five artifacts fall to any comparison group on the same schedule. The fifth needs a control in which participants cannot tell which condition they are in. That is why a waitlist-controlled trial is a real advance over no control at all. And why it still leaves the largest question in this literature open.

Two Validities

Why Efficacy and Effectiveness Are Different Jobs

Two kinds of validity now, and they pull against each other. Internal validity is the degree to which the design supports the causal conclusion. That the intervention, rather than something else, produced the observed difference. External validity is the degree to which the finding extends beyond the study. To other people, other settings, other conditions. Everything that raises internal validity makes the study less like ordinary practice. Narrow eligibility criteria give you a clean sample. They also exclude the participants clinicians most often see. A rigidly manualized treatment ensures the intervention was delivered as specified. It also removes the adaptation a clinician would ordinarily make. A study with weak internal validity establishes nothing, so it has nothing to generalize. A study with strong internal validity and no external validity describes a population nobody treats. The resolution is division of labor across studies, rather than compromise within one. Efficacy research asks whether a treatment works under ideal, tightly controlled conditions. Effectiveness research asks whether it works in ordinary practice. Treatments routinely perform less well in effectiveness studies, for specific reasons. Real caseloads include the comorbidity that trials exclude, which is the typical case. Ordinary clinicians have less training in the protocol, and less supervision. Efficacy trials answer whether this can work at all. That has to be settled first. The error is treating an efficacy result as though it were an effectiveness result.

Taken Apart

An Uncontrolled Claim, Objection by Objection

Now take an uncontrolled claim apart. This is a composite, describing no real intervention. A described six-week group program is offered to eighty adults reporting high anxiety. They complete an anxiety rating scale at enrollment, and again at the end. The average score falls from thirty-two to twenty-one. A drop of about a third. And forty-six of the eighty report feeling much better. The program is promoted as effective. Work through the objections. Regression to the mean. Participants enrolled because their anxiety was high at that moment. The average would be expected to fall at a second measurement even if the program did nothing. That alone could account for a meaningful part of an eleven-point drop. Spontaneous remission and natural course. Anxiety presentations fluctuate. A proportion of people improve without treatment over six weeks. Nothing in this design distinguishes those participants from responders. Expectancy. Every participant knew they were receiving a program described as helpful. The forty-six who felt much better were reporting on a treatment they knew they had received. Familiarity with the items changes how the second administration is answered. And attention. Participants met a group weekly for six weeks. Contact, structure, and the company of others are plausible sources of improvement. Sources that have nothing to do with the specific content of the program. So what survives? The eighty participants improved. That is all. It is a true statement, and it is not a claim about the program. Notice the shape of the argument. No step disputed the data. What was disputed is the attribution, which is a different thing entirely.

About Abnormal Psychology at UPI Study

Abnormal Psychology (Psychology 180) is a self-paced online course of 97 video lessons in 25 chapters, from the main theoretical perspectives and clinical assessment to anxiety, mood, psychotic and personality disorders, treatment and the law. It carries a NCCRS credit recommendation for 3 semester credits.

Every lesson transcript is free on pages like this one, and the videos are free with a free UPI Study account. Enrolling adds a practice quiz after each lesson, the graded work and final exam that earn the credit, lifetime access, and an official transcript on completion. UPI Study has helped over 130,000 students with college credits. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Enroll now → Course details Download the syllabus All Psychology and social science courses Every course with PRO →

Why take it here

Why students take Abnormal Psychology with UPI Study

Understand mental health disordersAbnormal Psychology covers how psychological disorders are defined, assessed, explained and treated. It is a core course for psychology majors and useful for nursing, social work, counseling and criminal justice.
Credit at a fraction of the priceThe same 3 credits can cost $750 or more at a US college. Here it is one payment, self-paced, with lifetime access.
Finish your degree soonerTransfer credits reduce the courses you still owe, so you can graduate earlier or lighten a heavy semester.
Strengthen an applicationA completed college-level course on an official transcript shows admissions and scholarship committees you can do the work.
Study on your own scheduleNo start dates, no deadlines. Watch the lessons, practice, and take the final when you are ready.
130,000+students helped with college creditsACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Where the credit goes

Universities that work with NCCRS

Abnormal Psychology carries a NCCRS credit recommendation. These universities work with NCCRS; some are UPI Study's signed partners. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Thomas Edison State UniversityThomas Edison State UniversitySigned partnerUniversity of the PeopleUniversity of the PeopleSigned partnerCharter Oak State CollegeCharter Oak State CollegeSigned partnerStrayer UniversityStrayer UniversitySigned partnerColumbia Southern UniversityColumbia Southern UniversitySigned partnerIndiana TechIndiana TechSigned partnerPenn StatePenn StateWorks with NCCRSSouthern New Hampshire UniversitySouthern New Hampshire UniversityWorks with NCCRSSUNYSUNYWorks with NCCRS +many more cooperating universitiesFind my college →

Find my college → Not listed? Ask Athena in the corner of this page, or book a 15-minute call.

FAQ

Questions students ask us

About The Purpose and Function of Control Groups

What is the purpose of a control group in an experiment?
A control group stands in for the counterfactual, meaning what would have happened to the same people without the treatment. Because the same participants cannot both receive and not receive an intervention, a comparable group that goes without it shows what improvement would have occurred anyway. Without a control group, a drop in symptoms cannot be attributed to the treatment, since time, chance, and expectation could all explain it.
What is the difference between random assignment and random selection?
Random assignment places participants into conditions by chance, which makes groups equivalent on average, including on factors nobody measured, so it supports causal comparison. Random selection draws participants from a population by chance so the sample resembles that population, which supports generalization. Most treatment trials use random assignment with self-selected volunteers, giving a strong causal claim about the participants but a weaker one about other people.
What is the difference between a waitlist control and a placebo control?
A waitlist control group receives the treatment after the trial ends, while a placebo control group receives an inactive intervention that looks like the real one. A waitlist rules out natural recovery and the passage of time but not expectation, since participants know they are not yet treated. A placebo control also rules out expectancy and attention, though it raises informed consent questions usually handled by warning participants in advance.
What is regression to the mean in clinical research?
Regression to the mean is the tendency for an extreme measurement to be followed by a less extreme one, because extreme scores partly reflect temporary factors. People often enter treatment at their worst, so their next score is likely to improve regardless of treatment. Like spontaneous remission and expectancy, it makes an ineffective treatment look effective, and a control group recruited on the same criteria cancels it out.

About the Abnormal Psychology course

How many college credits is Abnormal Psychology?
Abnormal Psychology (Psychology 180) carries a NCCRS credit recommendation for 3 semester credits. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.
How much does Abnormal Psychology cost and what is included?
$250, paid once, with lifetime access: all 97 video lessons, a practice quiz after each, the graded work and final exam, and an official transcript when you finish. Students who want several courses choose UPI Study PRO instead.
Which universities accept UPI Study credits?
UPI Study has signed partner agreements with universities such as Thomas Edison State University, University of the People and Charter Oak State College, and works with universities such as Penn State, SNHU and SUNY. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities. Use "Find my college" on this page to check yours.
How long does Abnormal Psychology take?
It is self-paced with no deadlines. The 97 lessons total about 24 hours of video; most students add reading and practice time and finish in a few weeks to a few months, on their own schedule.
How are the exams done?
Quizzes and assignments are online in the course. The final exam is taken online with proctoring; a person on the UPI Study team checks assignments, usually within two days.
How do my credits reach my college?
When the course shows 100% complete, request the official transcript from your dashboard. It is sent to the college you name, and you can send it to more than one. Check your college first with "Find my college" on this page, or ask Athena.

Ready for the full course? Earn 3 credits in Abnormal Psychology.

97 video lessons, a quiz after each one, and the graded work that earns the credit. Recommended for credit by NCCRS.

Enroll in Abnormal Psychology →Download the syllabusEvery course with PRO →
Abnormal Psychology
3 credits · NCCRSEnroll →