These are the notes I made whilst watching the video recording of Paul Meehl’s philosophy of science lectures. This is the ninth episode (a list of all the videos can he found here). Please note that these posts are not designed to replace or be used instead of the actual videos (I highly recommend you watch them). They are to be read alongside to help you understand what was said. I also do not include everything that he said (just the main/most complex points).
Probability 2 quantifies the proportion or relative frequency of attributes/properties (e.g. being schizophrenic, dying within the next 4 years) of a class or objects (e.g. coin flips)
p=1 doesn’t prove that mathematically something is certain.
3 ways of interpreting probability 1: 1) the way everyone is initially taught (linked with Reichenbach). 2) formal or axiomatic way (linked with Kolmogorov). 3) Karl Popper’s method of taking the formal way and coupling it with this idea of propensity or disposition (in which a given physical situation will likely give rise to a certain outcome).
Probability 1 (logical probability, the logical relationship between evidence and the conclusion from the evidence) is related to weighing up the evidence in relation to the hypothesis and stating how likely the hypothesis is because of this evidence.
The facts that we have to support a scientific theory may themselves be of a relative frequency but it doesn’t follow that the connection between those facts and the hypothesis can be stated as a statistical frequency. The probability (p) of the theory (T) given the facts (f) can’t be made into a relative frequency (what would be the class?).
Subjective Bayesian statistics involves stating how much you would be willing to bet on something. But the way those odds are assigned doesn’t mean it is similar to the way we calculate the odds by probability 2 (using an algorithm). Probability 1 cannot use an algorithm to create a number.
Probability `1 is about relationships between compositions/between beliefs. If you want to stick to a hypothesis regardless of the evidence, you can do it.
The odds you place on the theory (causal or historical theories, rather than statements of parameters) being true isn’t derived from an algorithm.
Theory (connections to and from) Hypothesis (connections to and from) Sample. Have algorithms for the connections between hypothesis and the sample, we don’t have algorithms for the connections between hypothesis and theory.
Attempts to merge probability 1 and 2 have failed (Reichenbach’s identity conception, Carnap’s universal language).
Both probability `1 and 2 lead to the definition of a fair bet, which is why they share the same name.
Probability 1 and probability 2 have a correlation relationship (and in the long-run very highly).
Offering treatments/trying to make a change in someone’s life always involves prediction (will be probabilistic).
Most decisions are made informally; by mulling over the information whilst thinking about the finite outcomes.
Must distinguish between the kind of data (e.g. psychometric, non-psychometric) and the mode of combining the data (informal, formal) for predictive purposes. Controversy surrounds how you combine the data.
Yonce, J. L., 2016. Philosophical Psychology Seminar (1989) Videos & Audio, [online] (Last updated 05/25/2016) Available at: http://meehl.umn.edu/video [Accessed on: 04/02/2018]