0

Does it make sense to say that one has an "incorrect bias"? Or is this a tautology?

Can a bias be reffered to as correct or incorrect? Positive or negative?

Thanks!

closed as primarily opinion-based by Dan Bron, Cascabel, David, ab2, NVZ Jul 8 '17 at 16:06

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    I think you mean 'redundancy' not 'tautology'. Note that 'bias' has multiple meanings, some very technical. So answers should address the different meanings. There is engineering bias, there's bias like a pre-judgement, and there's bias like a systematic error in a particular direction. – Mitch Jul 5 '17 at 15:38
2

In engineering, a bias to parameters is used to translate a function into an operational domain. Transistors, for example, do not generally operate from zero volts to their maximum, but from 3/10 or 7/10 of a volt. So the correct bias to the signal will be the bias that allows the transistor to properly amplify the signal.

Similarly, we humans have biases. No matter how hard we try, we can't get rid of them. We can adjust them somewhat, but we are not even aware of most of them.

A correct bias is one that allows an individual to function in a particular context. An incorrect bias will be one that prevents functioning.

(afterthought)

I should give an example of a correct bias.

We have a bias against eating decayed matter. Relative to certain types of decayed matter, we have a very strong bias.

Without being too specific (human waste?), we can say this is a correct bias, can we not?

Some people have allergies specific to certain cheeses, and a bias against eating those cheeses would be correct for them, while such a bias might not be generally correct. So we can also see that the correctness of bias is sensitive to context.

I hope that's enough example.

(end afterthought)

I'll note that you are begging a question here. You are assuming that bias is inherently wrong, and that assumption is not proven.

It can, in fact, be demonstrated that completely suppressing all biases tends to leave a person without motivation to act in either good or bad ways.

1

Does it make sense to say that one has an "incorrect bias"? Or is this a tautology?

It is not a tautology, as it is not repeating the same meaning.


Can a bias be reffered to as correct or incorrect?

Not really. The meanings do not match.

He is an incorrect student.

While this sentence is grammatically valid, it makes no sense at all. What is the definition of a correct student? How can a student be incorrect?
The same applies to bias. What is the definition of a correct bias? What would constitute an incorrect bias?

Let's look at the OED's definitions of a bias:

Inclination or prejudice for or against one person or group, especially in a way considered to be unfair.

Can something that's unfair be correct? I would say that it inherently cannot be. Because if it's correct, then it would be fair.

Something that is correct/incorrect, is objectively measured to be true/false. For something to be (in)correct, it either fits (or conflicts) with the definition of what is correct.

Thomas is a boy.
Thomas thinks red is the prettiest color.

These statements can be correct/incorrect, as they have a measure of objective truth to them. E.g. Thomas has XY chromosomes, or Thomas has already stated that he thinks red is the prettiest color.

Red is the most popular color.
Red is the prettiest color.

The first statement is still objective. Assuming there is a measure for popularity (e.g. it is the most frequently chosen favorite color), it is possible to objectively measure popularity.

The second statement, however, is incapable of being correct/incorrect, since it is inherently subjective. There is no objective measure of "pretty" that applies globally (not just on a personal level), and therefore this statement cannot be measured objectively.

A bias is inherently a subjective idea; which means it cannot be objectively measured (note: you can objectively measure the existence of a bias, but you cannot describe the bias as being correct or incorrect).
Similar to prettiness, correctness (in regard to a bias) cannot be defined in a way that it applies globally (rather than on a personal level).

Can a preference be correct? I don't think it can be. A preference is subjective. It has no global definition of correctness because that's what "subjective" (functionally) means.

However, due to human speech, you might get confused with seemingly objective statements that actually communicate subjective preferences:

A Red is the prettiest color.
B That is correct.

B is not saying that A is objectively correct, but rather that he (subjectively) agrees with A (because he shares the same opinion).
This is somewhat confusing; because these people seem to be talking about objective truth, while in reality they are both communicating subjective preferences.


Positive or negative?

Yes. A good example of positive and negative biases can be found in stereotypes.

All Asians are inferior drivers => Negative bias
All Asians are superior mathematicians => Positive bias

For example, I would say that a person who thinks that their race is superior to the average human, is positively biased towards his own race.
But I would say that a person who thinks that one specific race of people is inferior to the average human to be negatively biased towards that race.

Please note that when I say "positive bias", I mean "an assumption of superiority" and not "a bias that is good to have".

  • Thank you for the detailed reply. Interested on your take on the reply posted above by Joel Rees, with regards to incorrect vs correct bias. – Rich Ford Jul 6 '17 at 4:06
  • @RichFord: I had typed a lengthy comment and then decided I should not ;) In short, I agree with Joel's definition of a bias in physics. But everything else has a hidden component that Joel did not address: a pre-existing definition of what is good and what is bad. He considers a correct bias one that delivers a good (better) result. But he then does not define what "good" means. That still means that a bias is not objectively correct, but rather correct according to the subjective definition of what is good and what is bad, which fits with my example of A and B's conversation. – Flater Jul 6 '17 at 7:41
1

Coming from the field of statistics and mathematical modeling, bias has a very specific meaning, and is one of the sources of error in a model. Error can come from bias, which is systemic and results from incorrect assumptions in choice of model or learning algorithm. The other source of error is variance, which is related to the sensitivity of the method to the particular training data.

In this field, there is no such thing as "correct bias", as it always refers to a model's departure from the true, correct model. A perfect model would have zero bias, not correct bias. In that sense, "incorrect bias" is redundant, because we know that bias results in error. Do note that this is well-defined terminology for a specific domain, and this answer is not intended to assess broader meanings of the term when used in psychology, sociology, or engineering.

Not the answer you're looking for? Browse other questions tagged or ask your own question.