Hypotheses and predictions

In the deductive approach, experimental psychologist use a theory to predict data. (We'll do this with the integrative buffer theory ). Predictions (or hypotheses) are used to test the theory. Empirical observation (i.e., data) either confirm or disconfirm a hypotheses; in other words, by specifying hypotheses, the theory will either predict the data from an experiment or it will not.

According to Popper (1961), a philosopher of science, experimental data that confirm a hypothesis is not to be taken as confirmation of the truth of a theory. This is because many theories could account for the data—not just the one currently being tested. The best that science can do, argues Popper, is disconfirm or disprove theories. In short, this position claims that we can never really know how some aspect of the world works—we can only know how it doesn’t work.

To demonstrate the deductive approach, we are going to first come up with some hypotheses, design a quick experiment, run the experiment, and then see if the data confirm the hypotheses. If the data do support the hypotheses, then we know that an integrative visual buffer could be one way the our brains encode and represent complex visual information—but other theories (which we haven’t discussed) could also account for how we see the world. So—as you can see—support for a theory isn’t all that helpful. What is decisive is a disconfirmation! If a hypothesis derived from the integrative visual buffer theory is not supported by the data, then we know that the integrative buffer theory is wrong. We won’t know how the brain really constructs a smooth and continuous perception of the visual world as we move our eyes, but at least we’ll know it isn’t using an integrative visual buffer. (Note that a disconfirmation usually causes scientists to shift into an inductive mode of reasoning. This means using data to construct a new theory. )