Clinical Versus Statistical Prediction: A Theoretical Analysis & a Review of the Evidence

by Paul E. Meehl

This volume explores clinical issues, such can we rely on clinical expertise in making decisions about people's lives; when should statistical data be used; and what kind of treatment is best for a particular patient?

Genres: Psychology, Nonfiction, Science, Mathematics, Philosophy, Economics, Reference

149 pages, Hardcover

First published January 1, 1954

Reviews

[Steve · Rating 5 out of 5]

Paul Meehl was a clinical psychologist who had a poster of Freud in his office. When he died on February 14, 2003, I cried. He kept his poster because he believed that human behavior was complex, and simplistic explanations could not possibly capture the complexity of human behavior. People are surprised to find that Meehl had such a poster, because he exemplifies "Dust Bowl Empiricism," particularly in this book. Meehl was careful to point out how predictions and diagnoses based simply on intuition are far worse than those that are systematically using statistical decision rules. For example, because the base rate of some disorder is low in the population, it is not likely that one will encounter clients with this disorder; while this may seem common-sense, there are psychological clinics in which I have worked that seemed to see tons of clients with the same, extremely rare diagnosis. Many psychologists wrongly believe that there is something special about their "clinical skill" or "intuition" in some way protects them from making errors in thinking about human behavior. Meehl offers suggestions for ameliorating the problems that humans systematically have. In fact, it is at this point that Meehl provides inspiration for Nobel Laureate Danny Kahneman who explicated several of the systematic biases that are involved in our decision-making processes. Meehl was a welcome read for a psychologist who was dissatisfied with the murky, intuitive decision making and diagnosis that some psychologists use. He gave me faith that there are better ways to make decisions and told me what they were.

Additionally, Meehl illuminates important issues in philosophy of science and provides references for readers. In particular, my interest in Wesley Salmon's worked was piqued by reading Meehl, and this has led me to interesting thoughts on psychology and science in general. Hence, Meehl's book is wonderful not just for its own content, but because of directions that it gives its audience for further reading.

[Arshad Pooloo · Rating 2 out of 5]

I got the gist of it but in no mood to read something this detailed. It is a good book just not meant for me. At chapter 5 I realized that it was taking lots of effort to even pick up from where I stopped... clearly I was not enjoying it. Maybe i will attempt to read it later on when I find a use for it or I am in the mood. For the moment, I feel like I'd rather read a good fiction and lose myself into its world and that is exactly what I am going to do.

[Santiago Aguirre · Rating 5 out of 5]

Muy buen libro. Es difícil de leer por el lenguaje tan técnico, incluso así fue entretenido y rico de información. Me pareció muy valioso el debate y creo que refleja bien la manera de practicar la psicología, que es una ciencia relativamente nueva en la cual seguimos buscando la mejor manera de combinar el arte con la ciencia.

Considero especialmente valioso el enfoque a la validez clínica y es que es necesario que como clínicos, aprendamos a notar si nuestras predicciones son mejores que un lanzamiento de moneda. Para esto es necesaria la estadística, de la cual huimos por miedo a los números.

Este libro me invitó a promover una psicología apegada a la ciencia, entendiendo que la matemática y la estadística son elementales para encontrar la validez en nuestros procesos. Esto se ve especialmente en latinoamerica, en donde la práctica del "ojo clínico" sigue dominando y en la que los terapeutas se rehusan a explicar sus procesos y únicamente argumentan que su experiencia los blinda contra la mala práctica basada en deducción y no en métodos válidos.

En conclusión, este libro es un buen recorrido del debate entre la metodología estadística y clínica. Ambos extremos fallan y la respuesta está en el medio, específicamente en utilizar métodos válidos acompañados de la experiencia. En esto Kahneman es más preciso y es que él afirma que la intuición o el "gut feeling" es útil en escenarios consistentes en los que el practicante tiene mucha experiencia. De esta manera la predicción no se basa en especulación, sino en métodos válidos.

[Hadiana Sliwa · Rating 5 out of 5]

Test test

[Madhur Bhargava · Rating 3 out of 5]

The author basically argues that, for the purpose of behavioural research/analysis, statistical(mechanical) method outperforms the clinical(interviews/questioning) method as the clinical method has a human component to it which is prone to bias. The length of this book is ~150 pages and maybe 2 pages are dedicated to the conclusion above and the rest ~148 are used by the author to defend the theory against various naysayers. TLDR - A lot of Daniel Kahneman's work was inspired by the author and has more(& better) substance to it. A personal recommendation to the readers would be to focus their personal time on that work.

[Andrew Dalby · Rating 2 out of 5]

I fail to see why anyone considers this a master work. It is confusing , poorly written, contradictory, unnecessarily subjective and at times ad hominem. What it wants to say could have been said in two pages and much more effectively. The only useful part is Chapter 8 which is the actual comparison of actuarial and clinical predictions and as all of those studies are from 80 years ago. This is hardly of any relevance other than historical interest. The issue of reproducibility in science has rendered many of the arguments irrelevant and the work of Kahneman and particularly Tversky with his work on measures has gone far beyond Meehl's contribution.

[Aleksandr]

I read chapters 1 to 5 only, and it is still worth it if you're interested in psychological assessment (not only clinical).
Chapter 6 seemed (unnecessary) over-complicated and too "philosophical" to me, but that might be due to my insufficient intellectual abilities.

Also check Meehl's 1989 lectures on Philosophical Psychology, which are available here or on YouTube.

[Dan Slimmon]

I can't rate this book. It was too far outside my wheelhouse and I couldn't make it through.

[John J Milton · Rating 3 out of 5]

At its time it would be considered revolutionary but nowadays his points are pretty obvious. Important work.

[Adam Driedger · Rating 1 out of 5]

Over my head in what I could understand but the parts I did were interesting.