Post No.: 0877
Furrywisepuppy says:
Phew! The fire in the main hold was swiftly snuffed out. Our oxygen supplies have taken a hit because one of the electrolysis machines got fried, but there’s still plenty to breathe for just a crew of two. The carbon dioxide levels are more of a concern though because some scrubbers have been damaged too. But we’ve got spares so I’ll get them replaced tonight.
Fluffystealthkitten says:
My poor Meowllennium Falcon! What caused the explosion?
Furrywisepuppy says:
I’m not sure. It could’ve been a lithium-ion battery in the cargo?
Fluffystealthkitten says:
We’ve got to be more careful with those in storage. I’ll check each item that has a battery tonight while you replace the scrubbers.
Furrywisepuppy says:
So that’s our plan for later. In the meantime, we do have a little moment to continue in our mission to explore biases, which was temporarily cur-tailed at the end of Post No.: 0863.
Fluffystealthkitten says:
Yes. We have now reached up to the verge of some frequent informational and statistics-related biases.
Furrywisepuppy says:
Audience bias – this is the propensity for individuals to consciously or unconsciously view information through their own biases, such as ingroup biases.
Fluffystealthkitten says:
Barnum or Forer effect – this propensity leads us to give high accuracy ratings to descriptions of our personality that are ostensibly tailored specifically for us but are in fact vague and broad enough to statistically apply to a wide range of people. This can help explain why many people believe in astrology and fortune telling.
Furrywisepuppy says:
Base rate fallacy or base rate neglect – the propensity to ignore base rate or general population information and to focus on information only pertaining to a specific case in question, as if it’s assumed to be ‘a special case’. So it’s about ignoring base rates when estimating probabilities.
Fluffystealthkitten says:
Berkson’s paradox – how we can easily misinterpret statistical experiments involving conditional probabilities. More specifically, this is when events that seem negatively correlated are actually not.
For instance, we may observe, from our own experiences, that restaurants in our town either serve good fish but bad chips, or bad fish but good chips. We may therefore conclude that fish quality and chips quality are negatively correlated. But that’s probably because we wouldn’t bother going to places that served both bad fish and bad chips, which means that we’ve failed to account for all these other restaurants, and the overall picture may be that fish quality and chips quality aren’t correlated at all or may even be positively correlated. It’s thus a kind of selection bias.
Furrywisepuppy says:
I now fancy some fish and chips before we get down to doing our other tasks tonight!
Fluffystealthkitten says:
Ooh I like that idea! I hope we’ve packed some in our cargo.
Furrywisepuppy says:
Clustering illusion – the tendency to over-read into small runs, streaks or clusters in large samples of random data; or seeing illusory patterns.
Fluffystealthkitten says:
Cognitive dissonance – this is where individuals discount the value or veracity of information, or actively avoid situations and information, that is likely to conflict with their preconceived beliefs, in order to try to maintain internal consistency in their beliefs. This is because if our beliefs and behaviours are dissonant with each other, we either have to change our beliefs or change our behaviours until they cognitively and logically accord with each other. For example, if we believe we are morally upstanding then we don’t want to believe that our behaviours are immoral, hence we’ll either have to stop believing that we’re morally upstanding or (more typically) deny that what we are doing is immoral.
Furrywisepuppy says:
Common source bias – the propensity to combine or compare research studies from the same source or from sources that use the same methodologies or data.
Fluffystealthkitten says:
Conjunction fallacy – the propensity to assume that specific conditions are more likely than general ones. So it’s thinking that ‘x and y’ has a greater probability of happening than just either ‘x’ or ‘y’, especially if we’re swayed by mental prototypes and stereotypes, like ‘naughty fluffy little kittens’ are more common than ‘little kittens’.
Furrywisepuppy says:
Domain neglect bias – the propensity to neglect relevant domain knowledge while solving interdisciplinary problems.
Fluffystealthkitten says:
Experimenter’s or expectation bias – how experimenters tend to believe, attest and publish data that agrees with their expectations, and vice-versa.
Furrywisepuppy says:
Gambler’s fallacy – the propensity to think that future probabilities are always altered by past events, when they could be independent to each other. This arises from a misunderstanding of the law of large numbers.
Fluffystealthkitten says:
Hot-hand fallacy – the belief that someone who has experienced success so far automatically has a greater chance of further success with further attempts, even if they’ve just been lucky. We can thus assume hot (or cold) streaks in areas like gambling or sports.
Furrywisepuppy says:
Illicit transference, fallacy of composition and fallacy of division – this occurs when a term in the distributive (referring to every member of a class) and collective (referring to the class itself as a whole) are treated as equivalent.
The fallacy of composition involves erroneously inferring that something that’s true of a part of something must be true of its whole.
The fallacy of division involves erroneously inferring that something that’s true of the whole of something must be true for some or every part of it.
Fluffystealthkitten says:
Illusion of validity – the belief that additionally-acquired information always generates additionally-relevant data for predictions even when it doesn’t because the new information could be already related to what was already known and thus is already accounted for.
Furrywisepuppy says:
Information bias – the propensity to search for information even when it cannot affect action. Or believing that the more information that can be acquired to make a decision the better, even if that extra information is irrelevant.
Fluffystealthkitten says:
Insensitivity to or ignorance of sample size – under-expecting variation or sampling errors in small samples. The variance from the true population average will likely increase the smaller the sample size we take.
Furrywisepuppy says:
Media bias – presenting stories in a way that primarily serves certain parochial interests, sensationalises or avoids offence, or only presenting stories that are dumbed-down or easy to explain. This sums up most of the news that people on social media consume nowadays!
Fluffystealthkitten says:
Misperception of randomness – genuine randomness can appear like what seems to be genuine patterns or trends in the short run.
Furrywisepuppy says:
Neglect of probability – how we can totally disregard probability when making a decision under fuzzy uncertainty.
Fluffystealthkitten says:
Observer and subject expectancy effects – this happens when a researcher expects a certain result and therefore unconsciously manipulates an experiment or misinterprets the data in order to find it, and when a subject unconsciously affects the outcome of an experiment or reports the expected result, respectively. These effects explain the need for double-blinding in experiments.
Furrywisepuppy says:
Parkinson’s law of triviality – how we can give disproportionate weight to trivial matters.
Fluffystealthkitten says:
Probability matching – the sub-optimal matching of the probability of choices with the probability of reward in a random context.
Furrywisepuppy says:
Proportionality bias – the propensity to assume that big events must have big causes. This may link to why people gravitate towards conspiracy theories that involve hidden yet somehow massive global organisations.
Fluffystealthkitten says:
Publication bias – publishing only what is desirable or likely to be published, amongst the total of what could be published, thus skewing the overall picture or complete truth.
For instance, a pharmaceutical corporation that fails to disclose any of its own studies that find negative or unfavourable results for any of its own products, all in the name of protecting the corporation’s profits.
Furrywisepuppy says:
Response, reporting or recall bias – a range of biases that influence the responses of respondents (like in a survey) away from an accurate or honest response, or the selective or incomplete reporting or suppression of information by respondents. This is the advantage of anonymising responses. Yet even this doesn’t always eliminate the problem of inaccurate or dishonest responses.
Fluffystealthkitten says:
Selection bias – where subjects for a study are self-selected or selected in a way that is not truly random.
Furrywisepuppy says:
Selective dissonance – subconsciously dismissing or distorting information if it counters one’s worldview.
Fluffystealthkitten says:
Selective exposure – similarly, this is the subconscious process of selecting only what one wants to see or hear.
Furrywisepuppy says:
Shared information bias – the propensity for group members, when seeking a mere consensus, to spend more time and effort discussing information that all members are already familiar with, and less time and effort discussing information that only some members are aware of. This can lead to sub-optimal group decisions.
Fluffystealthkitten says:
Simpson’s paradox – a positive/negative trend that appears for two separate groups can counter-intuitively appear as a negative/positive trend when these very same groups are combined.
As an illustration, in one quiz you answered 7 out of 8 questions correctly, while your friend answered 2 out of 2 correctly. Your friend therefore had a higher correct ratio than you that day. In a second quiz you answered 1 out of 2 questions correctly, while your friend answered 5 out of 8 correctly. Your friend therefore again had a higher correct ratio than you this day. Yet if you aggregate each of your scores for the two quizzes, you achieved 8 out of 10, while your friend achieved only 7 out of 10! You therefore overall had a higher correct ratio than your friend.
Furrywisepuppy says:
Subadditivity effect – the propensity to judge the probability of the whole, when judged as a whole, to be less than the sum of the probabilities of its parts, when each part is judged individually. Or the propensity to estimate that the likelihood of a remembered event is less than the sum of its mutually-exclusive multiple components.
Fluffystealthkitten says:
Survivorship bias – focusing on the people or things who or that survived some process, and inadvertently overlooking those that didn’t or simply never ever existed because of their lack of visibility.
Furrywisepuppy says:
Systematic bias – a bias resulting from the system, in contrast to random errors that cancel each other out.
My tummy is telling me I’m famished. Let’s see if we did pack some fish and chips on our voyage?
Fluffystealthkitten says:
…Alright, the good news is that we did pack some of those rations in our cargo.
The bad news is that they all got burnt to ashes in the blaze!
Furrywisepuppy says:
Nooooo!
Fluffystealthkitten says:
I’ve found some liquefied ham in a tube though.
Furrywisepuppy says:
Groan.
Woof.
Fluffystealthkitten says:
Sigh. By using the Twitter comment button just below, you can comment or expand on any bias presented in this post.
We’ll be back soon. Meow.
Comment on this post by replying to this tweet:
