User:Tiger1112/sandbox

= Omission bias =

Omission bias is the phenomenon in which people prefer omission (inaction) over commission (action) and people tend to judge harm as a result of commission more negatively than harm as a result of omission. People who prefer commission (action) over omission (inaction) are said to have an action bias. It can occur due to a number of processes, including psychological inertia, the perception of transaction costs,the perception that commissions are more casual than omissions and a bias towards indirect causation which omission brings. In social political terms the Universal Declaration of Human Rights establishes how basic human rights are to be assessed in article 2, as "without distinction of any kind, such as race, colour, sex, language, religion, political or other opinion, national or social origin, property, birth or other status." criteria that are often subject to one or another form of omission bias. It is controversial as to whether omission bias is a cognitive bias or is often rational. The bias is often showcased through the trolley problem and has also been described as an explanation for the endowment effect and status quo bias. The Bystander effect may well be a form of omission bias with both showing a preference for inaction, however, the very important presence of people in the Bystander effect separates these 2 cognitive biases apart.

Baron and Ritov: Vaccination experiment
The first vaccination experiment by Baron and Ritov was one of the earliest studies conducted on Omission bias in 1990. Later renditions like Baron and Ritov’s new study on Omission bias in 2004 as well as other altered similar studies conducted by other psychologists make this one of the most well-known experiments in the field.

The basis surrounding the original 1990 study was that participants had to choose whether or not to vaccinate a group of children from a disease which would kill X children in every Y children (X and Y were used here as 5 total experiments were conducted and odds of death and children population were changed across some). The catch was that the vaccination also had risks of death to the children, hence providing the participants with 2 options which both lead to potential death (The risk of death also varied across the 5 experiments). The results across all 5 experiments showed an omission bias, a tendency for inaction even if it meant a worse outcome in some cases. Within a qualitative report conducted among the participants, those who answered claimed the sense of responsibility was their reason for omission bias. Many participants didn’t want to feel responsible for something as serious as a death of a child because of their actions. If a child had died due to catching the disease, they would feel arguably less responsible for their deaths.

Within the new 2004 report, Baron and Ritov theorised cases where action bias may be favoured. One major reason stated for omission bias was the sense of responsibility. In the 2004 report, the condition involving putting the participant in a position of responsibility could perhaps cause commission to be favoured over omission. The theory behind this was that a position of responsibility had an equal responsibility to prevent harm through inaction and action e.g., a healthcare advisor making the decision on the vaccination programme in the 1900 experiment would feel equally responsible for deaths caused by vaccination or not vaccinating and catching the disease. However, none of these cases was found. A case of action bias was found in a different study conducted in 2002. The study found that in the premise that the current situation is a failing one e.g., a company failing to make money, commission was preferred over omission, as omission guaranteed continued failure whereas commission had a chance for improvement.

All issues discussed previously suffer from a low sample size. No single experiment stated above had a sample larger than 100, even though Baron and Ritov had in total over 100 participants, splitting them across 5 independent studies meant their sample size was relatively underpowered. However, with Baron and Ritov’s 1990 study being a pioneer in the field, it became the basis for many future studies which in turn caused it to be repeated many times. Its high repeatability made it very reliable in analysing omission bias in one specific scenario. The other studies such as the 2002 study by Zeelanberg and the 2004 study by Baron and Ritov all have the issue of the small sample but don’t have the repeatability to back it.

Power of Knowledge
The original 1990 study by Baron and Ritov although the most widely known one within the field fails to add breadth to omission bias, restricting omission bias towards the singular case of vaccination. Baron and Ritov conducted 5 separate experiments within that study and unintentionally altered the knowledge available to the participants between different experiments. Due to this Baron and Ritov conducted another study in 1995, this time focused on the power of outcome knowledge as well as many different scenarios rather than vaccination alone. Across 4 experiments, each having many different scenarios within them, participants showed varying degrees of omission bias in a wide range of scenarios. Here are all the outcomes:


 * 1) When given no knowledge of the outcome participants favoured action, this could be due to the feeling of responsibility hence guilt was not existent in this case.
 * 2) When knowledge of the outcome is given in the form of both omission and commission will lead to a positive effect, participants still favoured action, perhaps for the same reasons as above.
 * 3) When given knowledge of the outcome that one of the 2 choices is more favourable but not which one, participants favoured omission.
 * 4) When given knowledge of the outcome that omission was safer, but action held greater rewards but had a risk, participants favoured omission. This is perhaps due to regret being associated more with commission than omission. Risk aversion could also affect the outcome here.
 * 5) When given knowledge of the outcome that fatalities could be involved, the omission was preferred but at varying degrees of how much more information was given. Omission bias was least prevalent but still prevalent in cases where the least information was given and was highest in cases where all information was given e.g., statistical chances of death

Impact of Death
Numerous studies conducted on omission bias involved dilemmas resulting in high stakes, most of the time fatalities. Decisions involving Fatalities seem to always be strongly enforced by Omission bias. A study done by Baron alone in 1992 showed that when participants are faced with an option to shoot a person to save another person’s life, omission was favoured. The 2004 study by Baron and Ritov tackled the train dilemma, omission was always favoured except for conditions that both resulted in the same person dying e.g., trying to stop the train and failing vs not trying at all to stop the train. The high value of human life usually resulted in people not wanting to deal with its responsibility hence, favouring omission bias. In the 1995 study by Baron and Ritov, commission was favoured when the stakes were considerably lower e.g., when both choices resulted in positive outcomes.

discussed in the experimental history

Examples and applications
Spranca, Minsk and Baron extended the omission bias to judgments of morality of choices. In one scenario, John, a tennis player, would be facing a tough opponent the next day in a decisive match. John knows his opponent is allergic to a food substance. Subjects were presented with two conditions: John recommends the food containing the allergen to hurt his opponent's performance, or the opponent himself orders the allergenic food, and John says nothing. A majority of people judged that John's action of recommending the allergenic food as being more immoral than John's inaction of not informing the opponent of the allergenic substance.

The effect has also held in real world athletic arenas: NBA statistics showcased referees called 50 percent fewer fouls in the final moments of close games.

An additional real-world example discussed in the experimental history is when parents decide not to vaccinate their children because of the potential chance of death—even when the probability the vaccination will cause death is much less likely than death from the disease prevented.