Cognitive Framing

Last week, in morning meeting, there was a presentation about a study of the framing effect. This effect describes a cognitive bias, where presenting the same options in different formats can alter people's decisions. This effect appears to have a greater effect on people who are inconsistent with their choices and whether the question is framed in loses/gains. This is very prevalent in every day life. Think about being presented with the choice to choose this route or that one, this type of meat or that one, should I try to buy this stock or that one, it goes on and on. It seems as though this effect is one of the best tools in the arsenals of lawyers, advertising companies and politicians. By framing an argument, a brand or a policy in certain ways, perhaps you are able to persuade people to think one way or another.

Here's an example...

A train is driving down a track, at a junction in the track there are two problems. On one track five people are tied down to the track and on the other track there is one person. You are there and next to a lever which you can throw to change or not change the course of the train (originally heading for the five people).

The two propositions are worded in the following ways...
1. You can throw the switch and save five people
2. You can throw the switch and kill one person

Although it is the same in both outcomes, people responded more positively to option 1.

The study that was proposed, would investigate framing in both normal and VMPC lesioned patients. The framing effects were broken into 6 different types of framing effects (e.g. Goal, Moral, Monetary). The researchers hypothesized that the VMPCs would differ from normals in there susceptibility to framing effects. With VMPCs, there is evidence of higher risk taking behavior and dysfunctional decision making behavior. It may be the case that the VMPCs may be more susceptible in certain framing situations and less in others.

They also proposed taking reaction times of the decisions. All though hard to measure, I think that it will be good to take into account both fast and slow reaction times. With fast reaction times, a situation is not being fully considered, while with a slow reaction time, the situation may be being thought out too much.

I can't remember the overall goal of the research, but I know that it is tied to moral reasoning. That's all for now

Moral Behavior

Today in morning meeting, Brad Thomas and Aaron Scherer presented a proposal for a study to investigate the moral behavior between VMPC patients and normals. The impetus for the study came from a social psych experiment in which participants where randomly assigned to three groups. In one group, the control, the members had to think about an average Tuesday. In the anger condition, the participants had to think about a time that they were angry. And in the last condition, the participants had to think about an instance of a moral violation. After the participants thought of their particular scenario, they were invited to participate in a 10 questions trivia test. There were two conditions, one with a reward of 20 cents per correct answer and the other with no reward. Be reminded that this study occurs on-line with participants participating in the safety of their own home. The way that participants were placed between the two conditions was through a flip of a coin, flipped by the participants themselves.

It is assumed that the with the flip of a coin you will land heads at a rate of 50% and tails at an equal rate. Deviation from chance (50%) would indicate some sort of outside influence. In the experiment there were the three conditions. In the control condition, 60% of the people ended up in the favorable condition, while in the anger condition a similar percentage of people were in the favorable condition. However, in the moral violation group over 75% of participants were in the favorable condition. A statistically significant and interesting result.

So what made these people cheat? That was the question and the supposed result of the study. These people are at home, receiving $4 dollars for initially participating in the study and then have the opportunity to make $2 more dollars if they end up in the favorable condition. Did the people in the moral violation group think that, "My cheating isn't as bad as that moral violation that I just thought of."

So the question is, how will the VMPCs perform in a similar situation. It has been shown before that these patients are impaired in decision making. It is the hypothesis of the researchers that there will be an increase in cheating behavior for developmental VMPC as compared to acquired VMPCs and normals.

The paradigm that they proposed was using a weighted die to decide what task patients would do. They proposed that 4 of the 6 side would be the unfavorable condition and 2 of the 6 favorable. However, with the weighted die, the favorable condition should only appear 2-4% of the time. A subject indicating a favorable toss would most likely be lying. However, a problem with studying these patients is the small sample size (n = ~10).

I proposed another way to investigate this question...

You have a task in which they are to complete some large number of difficult math problems, say 100.
It is assumed that these problems are difficult enough, yet small enough that they will get them wrong and also not take much time on them (they may even be multiple choice?).

The participants are then allowed to grade their answers with a flawed key. This is where the interesting part occurs. You could weight blocks of questions. Say you divide the test of 100 questions into 10 blocks of 10 questions and reward correct answers. However, with each block of 10 questions you increase the reward by a multiplier (perhaps also have a counterbalanced condition). In the first block the multiplier is small with incremental increases in the multiplier to the last block.

This design allows for investigation and interpretation from many facets. You can look at the raw number of instances of cheating, treating each problem as a moral situation (100 per subject!). You could also look at the blocks. Do people cheat at a higher rate in the blocks with higher reward as compared to the lower reward? Do the VMPC patients show consistently high levels of cheating across blocks? Similar to Chris's suggestion of the psychophysical measure, you could look to see if there is a certain break point in the multiplier which shows an increase in cheating. Does the break occur at 7 for normals as compared to 4 for VMPCs?

I don't know if this is actually showing morality or if people are just following the key...

I'm not sure, but it is something interesting to think about.

I think that this is an interesting but very difficult topic to study. Perhaps there will be some interesting follow ups and maybe some results.