Week 4

How many clues do you need?

What did we have so far? We heard that when we have to make decisions, we either examine the most important cue that is available to us or we examine all available cues for the current situation. In case the decision maker examines all cues, he or she assigns certain values or weights to these cues and makes a decision. So far I wrote that the simpler matching heuristic is a better predictor of decision making, although, it only searches through a small subset of cues.

If you were offered £500 with a probability of .50 (500, .50) otherwise nothing or £2500 with a probability of .10 otherwise nothing (2500, .10), which gamble would you choose?

In a study with Austrian students 88% would choose the first gamble with the smaller probability of the minimum gain but the lower maximum gain (Branstätter et. al., 2006). The researchers developed the priority rule in which we consider reasons in the following order: minimum gain, probability of minimum gain, maximum gain.

However, when should one stop considering these rules, or should you always examine each of the criteria? To answer this question read the following gamble (it’s the last one, I promise!).

You were offered either £200 with a probability of .50 (200, .50) otherwise nothing or £100 for sure. If you would be offered £2000 with a probability of .50 (2000, .50) otherwise nothing or £100 for sure, you might select a gamble with a different outcome. The minimum gains differ by the same amount and the probabilities are the same. The maximum outcomes, however, are different (1. £100, 2. £1000). Therefore a stopping rule has been proposed: Decision makers stop examination when the minimum gains differ by 1/10 (or more) of the maximum probabilities.

These two rules together and the decision rule, which tell you to choose the gamble with the more attractive gain, define the priority heuristic. The priority heuristic is able to explain shortcomings in decision making, e.g. the Allais Paradox or the Certainty Effect.

Week 3

How much are YOU willing to risk?

Consider the following gamble: You are offered either a sum of money for certain or a lottery ticket that will give you a 50% chance of winning £10.000 and a 50% chance of winning nothing. What would be the certain sum that makes you indifferent whether to receive it or the lottery ticket?

Below you find two graphs that show my results with different sums and different probabilities.

These graphs show that I am a rather risk aversive person. Your graphs might look the same or you are a rather risk seeking individual, which would alter your curve from the bottom at lower sums and probabilities to the top with higher sums and probabilities.

Week 2

What do you get when you put two people in two different courts where they write down 342 decisions by 57 benches? 

Apart from a lot of paper work you might find out that the matching heuristic is a better predictor for judicial decisions than a more complex model. Dhami (2003) conducted the above-mentioned study. This study found out that the matching heuristic was a better predictor that franklins rule. 

This result might come as a surprise because franklins rule examines more cues in a decision and assigns them certain weights, as explained in my first week blog. The matching heuristic, however, searches only through a small subset of cues and bases a decision on one cue alone. 

Why, however, does this result occur? Judges are presented with a heavy caseload and are usually working under time pressure. Therefore many judges rely on the decision made by the police, previous benches, and prosecutors, and maybe unintentionally “passing the buck”. Judges also make decision as a bench and this involves shared responsibilities, which use fewer cues. 

In the end there is only one question remaining: How do they know that the right decision is the right decision?

Week 1

Have you ever wondered why judges in a court can make their decision accurately or whether they make the right decision at all? 

This week was the first lecture in Judgement and Decision Making and therefore it was a simple introduction lecture. The lecturer, Dr. David Hardman, sets as part of the assessment for the module to write a blog and contribute to a website. To do this we were supposed to set up a Google Mail account and a blog. I did these things before, and that’s why I mostly helped another student to get her mail account working and start her blog. 

The final part of the lecture, and to my mind the interesting part, was about decision making by using a multiple linear regression analysis. These models suggest that people consider certain information when making a decision. These different information are given a certain value and according to the final outcome of this analysis they make their choice. Not everyone, however, is able to use all the information that are available. Time pressure or exhaustion for example can be reasons why someone cannot use every clue. 

Can you remember one situation where you valuated every piece of information, assigned a certain value, and made the perfect decision? 

I was surprised when I read that decision making models outperformed human professionals in their areas. Libby (1976) proofed that statistical models were better with predicting the success of companies than bankers who used the same information. Maybe someone should have listened to Psychologists before borrowing money from government. One relief, however, is the fact that to get these models working one needs humans to choose the variables for the decision outcome. 

Although it only was the introduction lecture in this module, it presented some useful and interesting information about how people, especially professionals, make decisions.