by Michael D. Anestis, M.S.
Having been trained in a scientifically oriented doctoral program, my understanding of depression is based, in large part, on cognitive theory (e.g., Beck, 2005). Developed by Aaron Beck, cognitive theory tells us that depression is, in part, a result of habitually inaccurate and negative thinking (keep in mind, however, this theory does not discount the contribution of other biological and environmental factors). In other words, depressed individuals come to think of themselves poorly, believe they have little control over their situation or hope for the future, and believe that any small negative aspect of their life or character generalizes on a broad scale. This theory, as you might guess, was the guiding force behind the development of cognitive behavioral therapy (CBT), which relies upon techniques such as cognitive restructuring, which we have detailed in earlier posts.
This theory makes a lot of sense to me and the data supporting the efficacy and effectiveness of CBT in the treatment of depression are impressive; however, another theory promoted by renowned researchers runs counter to it and, as such, leaves certain issues open to question. This theory, known as depressive realism (Alloy & Abramson, 1979), states that, while depressed individuals are certainly more negative than are nondepressed individuals, they are not less accurate. This theory was based upon a series of studies conducted by Alloy and Abramson (1979) in which college students with higher scores on a measure of depression demonstrated greater accuracy in their estimates of how much control they had over a task in which they had varying levels of control than were college students with lower depression scores. The authors concluded that people who are not depressed maintain a level of self-deception that enables them to have inflated views of themselves and that this self-deception breaks down when they become depressed and realize how weak they truly are and how little control they have in life (also see Alloy & Abramson, 1982; Benassi & Mahler, 1985; and Vasquez, 1987 for examples of studies on depressive realism).
Today, I would like to discuss a study by Richard Carson of Fisk University and Steve Hollon and Richard Shelton of Vanderbilt University soon to be published in Behaviour Research and Therapy. The authors noted all of the information I detailed above and also noted that, if depressive realism truly is an accurate description of depression, than it would be difficult to understand how CBT could truly operate through the mechanisms believed to make the therapy so effective (e.g., helping people to change inaccurate, negative thoughts to better reflect reality). Carson and colleagues (in press) mentioned, however, that past studies on depressive realism have been flawed in a number of important ways. First of all, rather than using samples of individuals who are actually diagnosed with depression, they have utilized samples of undergraduates with slightly elevated levels of depressive symptoms on self-report measures. Secondly, they have not investigated the manner in which "depressed" individuals have arrived at their conclusions regarding control. I'll explain this second part in more detail as I explain the procedures of the actual experiment, which was a bit complicated.
In their study, Carson, Hollon, and Shelton (in press) recruited 40 psychiatric outpatients who met criteria for depression as determined through a structured diagnostic interview. They also recruited 40 participants who did not meet criteria for any major DSM-IV-TR diagnosis. Doing this allowed them to compare truly depressed individuals to individuals with no diagnosable mental illness. Each participant then took part in a computer task. In the task, a solid square would appear on the screen signaling each time a trial had begun. The participant's job was to either press the space bar or not press the space bar, with the goal being to make a circle appear on the screen. They were told that the program was set up such that, sometimes they would be reinforced (the circle would appear) for pressing the space bar and sometimes they would be reinforced for not pressing the space bar. Half of the participants from each group were assigned to the "contingent" condition, in which they had a degree of control over whether or not a circle appeared on the screen. The other half were assigned to the "noncontingent" condition, in which whether or not they pressed the space bar had no bearing on whether or not the circle appeared. Within the contingent condition, half of the participants were reinforced 75% of the time they pressed the space bar and 25% of the time they did not press it and the other half were reinforced in the opposite manner. After the task, each participant was asked to estimate how much control they had over the outcome of the task and to estimate what proportion of the time they were reinforced for pressing and not pressing the space bar.
I realize that was a bit complex, but the experiment basically came down to this: half the participants had control over the outcome of the task and half did not. The authors wanted to see whether depressed or nondepressed individuals were more accurate in guessing how much control they had over the outcome. Additionally, the authors were curious how accurate the participants would be in remembering how often they were reinforced overall and how often they were reinforced for each particular type of response they gave. Their reasons for this are important. First of all, they were curious whether, as cognitive theory would predict, depressed individuals would underestimate how often they were rewarded for a particular behavior. After all, in CBT, we work with clients to test whether or not engaging in positive and productive activities will have any overt benefits. Second, they wanted to see whether depressed participants arrive at their estimates of control in a logical manner. In other words, if they are inaccurate about how often they are reinforced for particular responses, this has important implications for their judgments of control.
So...what did they actually find?
Depressed participants believed they had lower levels of control over outcomes than did nondepressed individuals, regardless of condition. Additionally, all participants believed they had more control in the contingent condition than in the noncontingent condition. This is important because, while they had a moderate amount of control (50%) in the contingent condition, they had none in the noncontingent condition, so any different answer would indicate that participants' answers were completely illogical.
More importantly, depressed individuals significantly underestimated the amount of control they had in the contingent condition (guess = 32%, actual = 50%, p < .01) whereas nondepressed individuals were rather accurate (guess = 54%, actual = 50%, p = .45). Both groups overestimated the amount of control in the noncontingent condition. Despite having no control over outcome in this condition, the depressed group estimated having 16% (p < >01) and the nondepressed group estimated having 34% (p < .01). In this condition, the nondepressed group was significantly more inaccurate (p = .02). So, in other words, when control was present, depressed individuals underestimated their impact on outcomes and, when no control was present, both groups overestimated their impact on outcomes, although nondepressed individuals did so to a greater extent.
This first set of findings is important for several reasons. First of all, it is not supportive of depressive realism to any legitimate extent. In the case of situations in which no control is available, depressed individuals were more accurate than their nondepressed counterparts, but neither group was particularly accurate. In situations in which control was available, however, depressed individuals were significantly more negative and inaccurate than their nondepressed counterparts, which is entirely consistent with cognitive theory. Remember, in CBT, therapists are not trying to make clients believe they have control when they do not actually have it. They are trying to help the client see the world accurately and challenge inaccurate beliefs. In this case, the impact of CBT would seem to be in helping depressed clients more accurately perceive when they can control the outcome of events. For instance, a therapists utilizing CBT for depression might help a client develop a greater sense of control over their mood, their productivity, their response to treatment, or any number of other tangible life events.
The study's findings did not stop here though. When asked to estimate how often they were reinforced for their responses, depressed participants reported receiving less reinforcement than did nondepressed individuals (p = .01) even though the two groups were reinforced equally. Additionally, depressed participants underestimated how much reinforcement they received (p < .01) whereas nondepressed participants were rather accurate (p = .42). In other words, depressed participants were less accurate in their recollections of how often they obtained the reinforcement response than were nondepressed individuals and, in fact, vastly underestimated the number of times this occurred. This again is highly consistent with cognitive theory, which states the depressed individuals do not accurately recall positive events and attend more closely to negative information that conforms to their negative self-views.
One final set of findings is worth noting, although there is more to this study and I encourage everyone to read the actual document. When asked to estimate how often they were reinforced for active (press the space bar) and passive (don't press the space bar) responses, nondepressed individuals were consistently accurate in their guesses whereas depressed individuals consistently underestimated the degree to which their action or inaction had an impact on whether or reinforcement was obtained. In other words, depressed participants perceived a diminished ability to remember the degree to which they impacted results in a positive manner. The bottom line with this result is that, even if the depressed individuals had been more accurate in their estimates of control, which they were not, their guesses would be based upon a faulty set of logical principles, as they were unable to accurately recall the impact of their actions on their environment.
In reading and considering a study like this, it is important to keep our conclusions reasonable. There are many studies on depressive realism and, while many are highly flawed, the results of this particular study are not representative of the results of all studies and, as such, more work similar to the comprehensive approach taken by Carson, Hollon, and Shelton (in press) is needed to enhance our confidence in these results. That being said, these results are a real blow to the idea of depressive realism. In this study, depressed individuals were, in fact, highly inaccurate in their estimates of control of their environment and in their memory of how often they attained reinforcement. Such results are very consistent with cognitive theory, which tells us that depressed individuals draw inaccurate, negative conclusions about their self, their world, and their future and apply these conclusions in a stable and global manner. CBT works to help depressed individuals more accurately perceive their environment and their impact on particular outcomes (including the outcome of therapy) by testing beliefs through experiments. If the results of this study are replicated by independent researchers, we will have further evidence that such techniques are, in fact, the reason why CBT is so effective for so many people in the treatment of depression.
************
If you would like to learn more about cognitive behavioral therapy for depression, we recommend the following items, each of which is available through our online store of scientifically-based psychological resources:
by Aaron Beck, John Rush, Brian Shaw, and Gary Emery
- Treatment Plans and Interventions for Depression and Anxiety Disorders
by Robert Leahy and Steve Hollon
- Feeling Good: The New Mood Therapy Revised and Updated
by David Burns
- The Feeling Good Handbook
by David Burns
Mike Anestis is a doctoral candidate in the clinical psychology department at Florida State University
Subscribe in a reader
