The (Un)disputed Champion of Psychotherapy – Clinical psychologists and their theoretical orientations

Clinical Psychologists Theoretical Orientations 1960–2010

A recent study by Norcross & Karpiak (2012) looked at survey data from members of APA Division 12 (Society of Clinical Psychology). In their study they included data that has been collected and published regularly since 1960, via self-administered questionnaires. In this new study the authors randomly sent a self-administered questionnaire to 1,285 members of APA Division 12 residing in the United States. 588 were returned giving a response rate of 45,8%, and, according to the authors, the sample was geographically and demographically representative of the population. However, they do add the reservation that the sample might not say anything about psychologists who are not affiliated with APA Division 12.

The (Un)disputed champion of Clinical Psychology

Looking at the data two clear trends emerge, first cognitive therapy is the most prevalent theoretical orientation in 2010s, and secondly psychodynamic and the other major theoretical orientations are on a steady decline.

Cognitive and Behavioral Therapies are the primary theoretical orientations reported by 46% of the respondents, 22% see them selves as eclectic/integrative and 18 % as psychodynamic-orientated. This means that every other psychologist in the US is identifying with a theoretical orientation, with at best modest research support and in many cases with none research support at al.

To be the best, you have to beat the best

I wanted to investigate the topographical research landscape of psychotherapy orientations. Using R statistical software I downloaded data about the number of publications each year for PDT and CBT. I extracted the number of relative publications each year by querying PubMed from R via PubMed’s E-utilities and extracted hits with R’s XML package. The plots were then made with ggplot2.

The queries used:
PDT = “psychodynamic therap*[tiab] OR psychodynamic psychotherap*[tiab]”
CBT= “cognitive behav* psychotherap*[tiab] OR cognitive behav* therap*[tiab]”

Update: Neuroskeptic pointed out, in a comment, that I could narrow down my search by using PubMed’s affiliation field. So I did that by adding (United States of America[ad] OR United States[ad] OR US[ad] OR USA[ad]) to the queries. Since, PubMed started adding affiliation fields 1988 the US-data is more zoomed in compared to the global chart.

Comparing clinical psychologists theoretical orientation to PubMed publications using R statistical software and ggplot
Figure 1. Comparing US clinical psychologists theoretical orientation to PubMed publications using R statistical software and ggplot

As you can see in Figure 1. psychodynamic proponents are absolutely not putting in the hours in the research labs, but nonetheless they still have a disproportionately large amount of practitioners. Moreover, practitioners who identify themselves as eclectic/integrative should really consider the data and heed the tremendous research effort put up by cognitive and behavioral therapists.

Flogging a dead dodo?

As I’m a psychologist myself, I’m well aware of the jargon concerning evidence based treatments from its dissidents. I don’t think these data will do anything to sway their minds. However, I strongly urge policy makers to look at the available research data before making decisions affecting the mental health services. One can’t simply disregard the sheer amount of evidence and research regarding cognitive behavioral therapies.

Notes about the R code

I will release the complete code to automatically download PubMed counts in a later post. I still have some bugs to sort out before it’s ready. Update: I’ve posted my script in another post: An R Script to Automatically look at PubMed Citation Counts By Year of Publication
Norcross, J., & Karpiak, C. (2012). Clinical Psychologists in the 2010s: 50 Years of the APA Division of Clinical Psychology Clinical Psychology: Science and Practice, 19 (1), 1-12 DOI: 10.1111/j.1468-2850.2012.01269.x

Kristoffer Magnusson

I'm a PhD-student and a clinical psychologist from Sweden with a passion for research and statistics. This is my personal blog about psychological research and statistical programming with R.

Comments (14) Write a comment

  1. Good post. but it looks like the PubMed hits are for research from anywhere, whereas the therapist orientation data is only from the USA? So they’re not strictly comparable.

    There’s a MEDLINE field for author institution so if you added a criteria there with “USA” you’d be able to narrow it down…


    • Thank you, you’re absolutely correct about the data. That’s a great idea to look at author’s affiliation. I was looking at the [PL] field earlier but abandoned it since it only contains publication country. I’ve now updated the charts with US-only data, and I would say it looks pretty much the same.


  2. Excelen datat! Although I dont have any data of my own at hand, my impression is that down here in Chile the situation is quite similar. Moreover, during undergraduated psychology courses, we saw CBT as the recommended treatment for various conditions; however we recieved little to none training in the administration of this therapeutic modality. Most of my colleagues integrate various techniques, drawing heavily from the Psychodinamic tradition, and do not pay much attention to the available evidence.


    • I would say that the situation is pretty much the same here in Sweden. During my studies we had to have one client in PDT and CBT for one year each. Of the 9 universities in Sweden were you can study to become a psychologist I believe most of them teach booth CBT and PDT. I don’t really understand our professions distaste for evidence based treatments.


  3. Hi Kristoffer,
    nice work! Would you mind sharing the code?
    I’d like to use it in an R course I give for psychologists.
    With a note about the auhtor of course :)


    • Thank you! I’m doing some changes to the code, but I believe I can share the code in a day or two.


  4. Thanks for sharing. I’ve just been doing something similar (i.e. similar technically but not similar in the search content). Used the “RISmed” R package to query Pubmed via an intermediate service (Hubmed). Unfortunately, the online connections are slow and unstable and one has to incorporate a lot of try: catch error handling, and even then some connection drops occur, so that some data is missing for some years.

    Nonetheless, very nice to do in R so as to get all the data in a useful data frame and then visualise nicely as you’ve done via ggplot. So will be very interested to learn from your code when you publish it.


    • Thanks for commenting! I began using RISmed but I also ran into connection drops and other errors, so I abandoned it and decided to write my own code. I will put up the code in a day or two.


  5. popularity contest? Probably more relevant is if CBT is more efficacious than other treatments in ameliorating psychological (or otherwise) symptomology. Unfortunately, the evidence is not terribly convincing (see Wampold et al., 1997, etc). However, I did find the post to be interesting, particularly in the use of R. Thanks for sharing!


  6. There is a clear mistake in the presentation, confounding the notions of treatment popularity and treatment efficacy. Clearly, CBT is the most researched orientation. That’s interesting. To take that interesting nugget regarding popularity as evidence for the claim that practicing from a psychodynamic orientation is somehow irresponsible and passé is grossly unfounded and scientifically unsound. Number of studies does not equal efficacy. In fact, recent work (see Shedler 2010) suggests that the evidence base for the efficacy (and, especially among personality-disordered patients, the superiority) has reached a critical mass. And this continues to grow, anecdotally speaking.

    Good science prefers multiple perspectives in dialogue. Be careful with sweeping generalizations based on mere popularity – there are likely many extra-Scientific factors at play (NIH funding, for one). I concede that psychodynamic research has “loafed” in the kind of research you are reviewing (and there are other kinds of research), but i cannot sit idly while you measure the height of stacks of paper and make your conclusions in the name of Science.

    Great methodology and thanks for the code. I will hopefully be able to provide more evidence for the nuance I suggest above, building on your work.


  7. AC & Mike Finn,

    In a sense it’s a popularity contest, but one effect of the greater amount of studies done on CBT is that we do know a lot more about it’s efficacy. And I do believe that CBTs increased popularity is due to its massive research corpus. Though, I’m glad that more PDT practitioners are changing their attitudes towards quantitative research and that we see more high constraint studies done on the efficacy of PDT. My take on evidence based treatments in regard to PDT is that we don’t know enough about its efficacy to recommended it as the first treatment choice for different mental disorders. I really hope that will change so that patients will have more treatment options, similar to the evolution of IPT.

    Regarding Wampold et al (1997) I do not find their evidence terribly convincing. Crits-Christoph’s (1997) rebuttal sums up the main issues with Wampold’s et al study pretty good. I know there’s also several articles by Leichsenring and Lambert on the same subject, but do not believe there to be enough high quality studies done for PDT and other psychotherapies to draw the conclusions they are presenting.

    Some Cochrane reviews have tried to compare different psychotherapies for various disorders, and they have come to the conclusion that more studies are needed for PDT before any robust conclusions can be drawn (see and Which really brings us back to my data about the issue of research loafing. But I think we’re far from putting the last nail in coffin for the dodo bird.

    And in addition, looking at APA’s Div 12 (Society for Clinical Psychology) list of empirically supported treatments one can see that they only mention PDT one time and that they conclude there’s only modest research support for Short-term PDT for depression. In Sweden, the National Board of Health and Welfare have presented similar conclusions in their treatment guidelines aimed at licensed practitioners. This is not evidence of the superiority of CBT; it merely reflects that more high quality studies were done on the efficacy of CBT at that time. I do really feel that it’s unjust to the clients, in the context of publicly funded health care, to provide treatments that are not the recommend treatment of choice according to it’s governing institution (referring to Sweden). My personal preferences for psychological treatments is actually not CBT but ACT and mindfulness-based treatments, but until more research is presented for the efficacy of those treatments, I feel it would be unwarranted of me to present it as the first treatment choice to a client who’s never tried CBT.

    However, I’m glad you took the time to comment so that other perspectives are presented on my site and it’s great that my conclusions don’t stand unchallenged.


  8. Great post! Full disclosure–I’m a cognitive-behavioral psychologist, and I constantly beat my colleagues up about evidence-based treatment.

    One thing to be aware of is that not all psychology journals are indexed on Medline. Most, if not all, are indexed on PsychInfo, which is the American Psychological Association’s service, which is not free like Medline. More info is at:

    The website is confusing, but if you pick individual journals, you can see where they are indexed. So, for example, a journal like Neuropsychology is indexed in Medline & PsychInfo, but, a journal like Psychology of Men and Masculinity is not indexed in Medline.

    Still, I don’t think that your conclusions would be wrong. My understanding is that historically, psychodynamically oriented psychologists have not published research with control groups.

    Also, there is a large amount of research into “common factors,” factors that might be shared by various orientations, and their contribution to treatment success, but that is another topic.


    • Thanks for commenting! You are correct that PsychInfo and PubMed differ in their indexing, however not all psychotherapy articles get published in psychology journals. I recently performed a meta-analysis of CBT’s effect for social phobia, where I did quite a comprehensive literature search. My feeling was that PubMed and PsychInfo had about an equal share of unique citations, though unfortunately I did not keep count. But I should add that I only looked at RCTs, which I believe are more likely to get picked up by PubMed than other types of psychological research designs.

      I’m actually working on comparing citations for PDT and CBT in PubMed and PsychInfo, my hypothesis is that PsychInfo will have a larger proportion of PDT articles than PubMed.


Leave a Reply