6
$\begingroup$

In logistic regression, I know that Odds and Probabilities are two different things, but still I can't distinguish between them.

I will be very grateful for your clarifications.

$\endgroup$
1
  • 1
    $\begingroup$ Given your question's tags, I will mention that odds are much more convenient (both mathematically, and for interpretation) than probabilities in the context of logistic regression. $\endgroup$ Feb 10, 2015 at 19:38

2 Answers 2

7
$\begingroup$

A probability lies between 0 and 1; I presume you're familiar enough with it that we don't need to define probability, but please clarify if you do require a basic definition. [So if I say something like "The probability of at least two heads in three tosses of a coin is 1/2" or "the probability that I miss the bus tomorrow is 1/10" I presume the sense of what that means is already clear.]

Odds are the ratio of a probability (p) to its complement (1-p) ($\text{odds}=\frac{p}{1-p}$. In this case the odds for missing the bus would be $\frac{1/10}{9/10}=1/9$.

[In elementary situations where probability is "number of cases where the event holds"/"total cases", odds are "number of cases where the event holds"/"number of cases where it doesn't".]

[In gambling, odds are usually expression not as a fraction but as a literal ratio ("1:9 for" rather than $\frac{_1}{^9}$), and most usually in terms of odds against (the reciprocal of the ratio, "9:1 against"), unless the odds against fall below 1:1, in which case they're again written the first way around, as odds for and referred to as "odds on"); this keeps the larger number expressed first. If there's no adjective (on, for, against), as in "the odds are 20:1", you would normally assume odds against in gambling situations. You'd flip those around and write as a fraction -- $\frac{1}{20}$ to get odds as a statistician writes them.]

To convert odds back to probability, divide the odds by (1+odds). So if the odds are 1/9, the probability is $\frac{1/9}{1+1/9}=\frac{1}{10}$

$\endgroup$
-1
$\begingroup$

Odds ratio is best understood in the gambling context. If you are throwing a dice, and the winning side is 6, then they say you have 5/1 chance to win, i.e. for every one winning chance there are 5 chances to lose. If the wining sides are 2 and 3, then your odds ratio is 2/1.

Now, you can see how the odds ratio is related to the probabilities: $r=\frac{1-p}{p}$. You can plug probability of losing, instead of winning $\pi=1-p$, then $r=\frac{\pi}{1-\pi}$, basically the same thing.

$\endgroup$
1
  • 1
    $\begingroup$ "Odds" and "odds ratios" are different things having different interpretations. $\endgroup$
    – whuber
    Feb 7, 2015 at 6:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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