that is used for comparison. Participants in the two groups are treated the same, except for the manipulation of the independent variable—in this case, the treatment. At the end of the study, researchers expect to see better outcomes among participants who received treatment than controls.
Let’s look at an example of a randomized controlled trial designed to investigate the efficacy of neurofeedback, a popular and relatively new treatment for ADHD.
Neurofeedback (sometimes called “neurotherapy” or “brain training”) rests on the premise that children with ADHD show atypical brain activity. This atypical brain activity, in turn, leads to problems with attention, concentration, and impulse control. In neurofeedback, children learn to regulate their brain activity to decrease ADHD symptoms (Thibault, Lifshitz, & Raz, 2017).
In a typical neurofeedback session, a clinician uses a device called an electroencephalogram (EEG) to measure the child’s brain activity in real time. Electrodes placed on the child’s scalp measure the wavelength of electrical activity in various brain regions, especially areas responsible for attention and concentration. Two wavelengths are especially important: low-frequency theta waves that reflect a state of inactivity and mental drowsiness and high-frequency beta waves that reflect a state of attention and alertness. Children with ADHD tend to show an abundance of low-frequency theta waves and relatively fewer high-frequency beta waves than children without the disorder. This atypical brain activity is believed to underlie their attention problems (Loo & Makeig, 2013).
Information from the EEG is delivered to a computer that converts it to a visual image on a screen. The child can learn to control the image by altering his or her brain activity. For example, the child might see a dolphin on the computer screen. When the EEG shows low-frequency theta waves, the dolphin stops swimming and floats to the surface. However, when the EEG shows high-frequency beta waves, the dolphin swims faster, dives to the ocean floor, and collects treasure. Children learn to control the dolphin’s actions and earn treasure by regulating their brain activity (Image 3.4).
Image courtesy of Pixabay Creative Commons
Children can learn to decrease theta and increase beta activity over time. Some clinicians believe that as children’s capacity to control their brain activity increases, so too will their attention and concentration. Training can take several months with multiple sessions each week (Janssen et al., 2017).
To evaluate neurofeedback, researchers recruited a sample of children with ADHD and randomly assigned them to either a treatment or control group. Children in the treatment group received neurofeedback, while children in the control group did not. To evaluate outcomes, the researchers asked parents to rate children’s symptoms before and after the study. As expected, the parents of children who received neurofeedback reported significant improvement, whereas the parents of children in the control group did not (Steiner, Frenette, Rene, Brennan, & Perrin, 2014).
Types of Control Groups
A critical decision in experimental research is what to do with participants in the control group. Some researchers compare the treatment group to a no-treatment control group. In this type of study, participants in the control group get no treatment whatsoever. The main problem with using a no-treatment control group is that it is unethical to deny treatment to children with mental health problems. By withholding treatment from children with ADHD, we might be contributing to their attention and behavior problems at home and school.
A potential solution is to delay treatment for a short period of time instead of denying it altogether. In this case, researchers compare participants in the treatment group to participants assigned to a waitlist control group. Treatment is delayed to children assigned to the waitlist, but not altogether withheld. At the end of the study, children on the waitlist are also allowed to participate in treatment. Waitlist control groups can be an ethical way to evaluate the efficacy of therapy (Table 3.2).
Table 3.2
Note: Based on Sorger and colleagues (2020).
For example, in another study, researchers randomly assigned children with ADHD to either a neurofeedback treatment group or to a waitlist control group. As before, researchers assessed children’s outcomes by asking their parents to rate attention problems before and after the study. As expected, the parents of children in the treatment group reported a significant reduction in symptoms, whereas the parents of children in the control group did not. At the end of the study, children on the waitlist were also provided neurofeedback (Cortese et al., 2017).
Both no-treatment control groups and waitlist control groups suffer from another shortcoming: participants know that they are either receiving treatment or serving as controls. This knowledge can affect their behavior during the study and their perceptions of improvement. The placebo effect refers to people’s tendency to improve simply because they know that they are receiving treatment and they expect the treatment to work. Researchers must control for the placebo effect; they need to be able to attribute symptom reduction to the treatment itself, not to people’s awareness that they are receiving treatment or their belief that they will get better (Thibault et al., 2017).
To reduce the placebo effect, researchers use an attention-placebo control group. In an attention-placebo controlled study, participants assigned to the control group receive a theoretically inert form of “treatment” that resembles the therapy received by participants in the actual treatment group. In medical research, participants might take a pill that looks like the actual medication but has no active ingredients. In psychological research, participants might talk with a therapist about day-to-day events, but not receive any form of real therapy. Participants in the attention-placebo control group would likely believe that they were receiving actual treatment, thereby controlling for the placebo effect.
For example, researchers compared children with ADHD who participated in neurofeedback to another group of children randomly assigned to receive “sham” neurofeedback. Children in the sham condition were connected to an EEG and they played neurofeedback games identical to children in the treatment group. However, the feedback that they received was random and not associated with their actual brain activity. Because these children and their parents believed that they were receiving actual neurofeedback, the researchers were able to control for the placebo effect. Surprisingly, the parents of children who received actual neurofeedback reported only slightly better outcomes than the parents of children who received sham neurofeedback. These findings tell us that the benefits of neurofeedback were largely explained by placebo, rather than by neurofeedback itself (Arnold et al., 2013).
A final option is to compare participants who receive a new treatment to treatment as usual (TAU). Participants assigned to a TAU control group are referred to clinicians in the community and receive whatever care these clinicians provide. Use of a TAU control group is the most stringent test of a new form of therapy. The new treatment must show that its benefits match or exceed those offered by existing therapies.
For example, researchers randomly assigned children with ADHD to two conditions. Youths in the first condition received stimulant medication (TAU). Youths in the second condition received neurofeedback. Parents reported significantly greater improvement among children prescribed medication than children receiving neurofeedback. Moreover, teachers who did not know that children were receiving treatment reported significant improvement in children taking medication but no improvement in children receiving neurofeedback. The researchers concluded that medication is superior to neurofeedback and that many of the effects of neurofeedback are explained by parents’ expectations (Gelade, Janssen, Bink, Maras, & Oosterlaan, 2016).
Randomized