Lecture 3: Decision Making Under Risk and Uncertainty

For our third lecture (week 4) we were asked to read The Priority Heuristic: Making Choices without Trade-Offs, by Brandstatter, Gigerenzer and Hertwig. Also to read Chapter 7 of Introduction to Judgment and decision making (JDMPP) to re familiarise ourselves with the theories of risky and uncertain decision making. Including the expected value theory, expected utility theory, and the Prospect theory. I found this reading useful as i had little memory of this subject!
Expected value theory states that a decision maker (DM) will assign a value to each attribute and they will choose that action with the highest expected value (i.e. highest monetary outcome). However this theory conflicts with peoples intuitions (see St Petersburg paradox). The expected utility theory takes peoples personal choices into account and looks at how a DM makes a choice between two or more risky or uncertain outcomes by comparing expected utility values (or subjective values) of the outcomes and multiplying them by their respective probabilities.
This theory has been highly criticized on three major points:
These theories argue that the DM is risk averse, in that they will avoid risky decisions because a loss looms harder than potential gain. However this risk aversion idea is not supported when outcomes are not sure things i.e. people will risk-seek when outcomes are both losses.
They state that a DM will always try to maximise their utility. However the Allais Paradox proves that this is not the case! It shows that when a problem is worded differently the DM will choose a route with a certain gain even if it is not the option with the largest outcome.
Lastly, the isolation effect: The theory assumes that people integrate possible outcomes with their current assets. Again this has been proven to be untrue, for example people will not always take into account money they get given before a task.
In the reading there were examples of these, and I even found that I would choose a certain outcome over an uncertain outcome, even if the uncertain outcome had a higher expected value! I think this is because if I chose the certain outcome I did not have to worry about the prospect of leaving the situation with nothing when there was a chance of winning some money.
Then came the prospect theory to combat these failures.. This theory accounts for the fact that a gain can be seen as a loss if you don’t gain as much as you expect to win. I can relate to this as when I received a payslip over the summer it was not as high as I expected it to be, the money I had earned therefore seemed like a bit of a letdown rather than a gain!
It also accounts for the impact of small probabilities. People overestimate small probabilities and underestimate large probabilities. For example lots of people -including myself- buy lottery tickets the expected value of a lottery ticket is very low as is its probability of winning however people still carry on buying them week after week.
I found the reading on priority heuristics a little hard going at times but I think I understood the overall gist of the article. The theory is similar to the take the best algorithm in that the DM will base their choice on 1 reason without looking at all the probabilities, just the outcomes. There are also 3 rules the DM must follow to make their choice. I do find this theory believable as it explains why we choose certainly over uncertainly because we don’t look at all the probabilities. Also in real life it is often impossible to work out all the probabilities of outcomes especially when we don’t have much time to make a choice!
During the lecture we were set a task to fill in two questionnaires. The were two different methods of measuring a person utility. The two graphs at the top of the page show my utility of certain monetary values for each method. Monetary values are on the x axis and utility values are on the y axis.

The first method was the certainty probability method. I was asked questions to test my utility, like if i was offered a lottery ticket with a 50% chance of getting £1000 and a 50% chance of getting £0 or getting a certain sum of money, how much would the certain sum of money have to be to make me indifferent between the two. to this particular question i answered £200 (my utility of 200 is therefore 0.5 because it was a 50% chance of getting all the money: £1000 or £0).

Then, the second set of questions was the probability method. This method should come up with the same graph as the first set because the questions were the same just framed differently . An example is: if i had the £200 for certain or a lottery of £1000 and £0, what probability of winning £1000 would make me indifferent between the two? I answered 80%.

As you can see the graphs are similar but not exactly the same! My utility for the £50 is a lot lower in the certainty equivalence method. This could be an example of the framing effect, that changing wording of a description of a decision can affect our preferences of decision. I did find it hard to give a totally honest answer because it was a made up scenario and not real life. I also found thinking in terms of probability was difficult. However they both curve showing that in both instances i was risk averse!