How do we reach conclusions from evidence? How do we decide what is truth and what isn’t based on the information we receive?

There are two types of reasoning that people theorized are used: deductive reasoning and inductive reasoning.

Deductive reasoning is reasoning from a set of known facts to a logical conclusion.

The most well-known of this type of reasoning is conditional reasoning, or reasoning about If-Then statements.

Researchers characterize this type of reasoning in terms of a set of deductively valid inferences (derived from formal logic) and examine where people fail to draw valid inferences.

Valid inferences:

Modus ponens                    p -> q; p, therefore q

Modus tollens                     p -> q; not q, therefore not p

Invalid inferences:

Denying the antecedent        p -> q; not p, therefore not q

Affirming the consequent     p -> q; q, therefore p
 
 

Wason found that most people use modus ponens easily, but not modus tollens, and in fact use denying the antecedent more frequently.

Once again, context matters.

If subjects are presented with a modified version of the Wason card task that makes it more concrete, they perform much better than with the abstract reasoning version.

Furthermore, the reason subjects are inclined to use denying the antecedent is because in many cases that works, as in a rebate for a purchase.

Research has found that people only use rules where appropriate, i.e. using modus tollens in the modified Wason task when checking whether a 16 year old is drinking beer vs. a 4 year old.

People have general reasoning rules related to particular goals or domains, such as permission schemas.  (see e-mail)

This is reasoning from two premises (major and minor) to a conclusion.

Another way to think about it is that this is reasoning about relationships among things. Each premise defines a relationship between two concepts, and at least one concept is shared between the two premises.

Linear syllogisms: The premises have a linear relationship, i.e. "Bill is taller than Bob." and "Bob is taller than Jim."

Categorical syllogisms: The premises describe the relationship of various sub-categories of a larger category.

Essentially, inductive reasoning is the process of drawing conclusions about broad classes of objects or events from a small number of examples.

Causal inferences: We make judgements all the time about causality of events, and we frequently use inductive reasoning to infer that these causalities will always exist.

Example: Confirmation bias: We more often see causation that confirms what we believe.

Correlation ? Causation

Categorical inferences: We also make judgements about category membership, and we extend these judgements to similar members of the category.

Example: Stereotypes.