People do not reason logically, but decision making encompasses all human behaviour. Choices can be rational (internally consistent). People expect choices to show transitivity. If a relationship holds between item one and item two and between item two and item three then the relationship should also hold between item one and item three. Utility is the personal value we attach to outcomes rather than to their absolute monetary value.

Normative theories of decision making imply that some choices are better than other choices. The optimal choice in normative theories depends on the theory.

The expected value theory states that the optimal choice is the choice that offers the largest financial payoff. People also tend to go for choices with the maximum utility. People are not always consistent in their choices with expected value and utility, as this requires a lot of time and motivation. There are two principles in rational decision making:

1. Description invariance
   The description of the choice should not make any difference as long as basic structure of the choices is the same.
2. Procedure invariance
   The procedure of decision making should not make any difference in the decision that people make.

People are inconsistent with this. Psychic budgets refers to how we mentally categorize money that we have spent or are contemplating to spend. Sunk cost refers to an investment that is irretrievably spent and should not influence present decision making, but still does, as people want to get their investment out of it as much as they can. People also make decisions based on loss aversion, the unpleasantness of a loss is bigger than the pleasure of a similar gain. Therefore, people make decisions based on aversion of the unpleasantness of a loss. Satisficing refers to selecting the first choice that satisfies a certain demand (e.g: cost of a phone). People use satisficing to prevent them from having to compare everything with each other.

People tend to use heuristics to make decisions. There are several heuristics:

1. Representativeness heuristic
   An event is judged to be probable if it has properties that are representative of that category (e.g: we believe a person wearing a metal t-shirt is more likely to be part of a metal band than someone in a suit).
2. Availability heuristic
   An event is judged more probable if one is able to recall many examples of it (e.g: deadliness of plane crashes versus cardiovascular diseases).
3. Anchoring and adjustment heuristic
   The initial value of an event adjusts or estimate upwards or downwards on the basis of other information (e.g: if people first hear that someone wants an offer between 10 and 50 euros, they will have a lower offer than when people hear that someone wants an offer between 50 and 150 euros).

The odds of a conjunction of two events are always lower than the odds of a single event. The gambler’s fallacy refers to gamblers making decisions on what they believe should happen (e.g: after 6 times red in a row with roulette, they believe it should be black, as that is more random). People tend to ignore base rate and sample size when making decisions. Heuristics often provide sufficient decisions and are not necessarily maladaptive.

People are not flawless in deductive reasoning, deriving the answer using formal logic. Premises are statements of fact that are presumed to be true. Inductive reasoning shows that a conclusion is more or less likely to be true, but it does not allow us to conclude that the conclusion must be true.

Example deductive reasoning:

P1: If A, then B
P2: A
----------------------
C: B

Example inductive reasoning:
P1: If A, then B
P2: B
---------------------
A

Inductive reasoning is not necessarily is true, as A is not the valid answer here. Syllogisms make use of the if-then format. There are several forms of syllogisms:

1. Modus ponens (valid)
   If A, then B
   A
   -----------------
   B
2. Modus tollens (valid)
   If A, then B
   Not B
   ----------------
   Not A
3. Affirming the consequent (invalid)
   If A, then B
   B
   ----------------
   A
4. Denying the antecedent (invalid)
   If A, then B
   Not A
   ------------
   Not B

People tend to do the modus ponens and the modus tollens correct. They tend to make mistakes in affirming the consequent and denying the antecedent, as people believe they are valid. People are better at reasoning using formal logic if it is concrete (e.g. no abstract phrasings) and if the problem is familiar.

Pragmatic reasoning schemas are generalized sets of rules that are defined in relation to goals. They lead to inferences that are pragmatic, whereas formal reasoning might lead to valid conclusions, but not necessarily practical conclusions. People tend to be good at exposing and recognizing people that break social rules using reasoning.

People tend to make a conversion error, in which the person reverses terms that should not be reversed (e.g. all and some). Conversational implicature refers to people using terms in the conversational sense rather than the logical sense. People also tend to make mistakes because of the atmosphere of the premises (e.g. both premises are negative or both premises use ‘some’). People also make mistakes in syllogisms because of prior beliefs.