# Margin of Error

**URL:** <https://boards.straightdope.com/t/margin-of-error/170811>\
**Category:** Factual Questions\
**Created:** [April 24, 2003, 7:04am UTC](https://boards.straightdope.com/t/margin-of-error/170811 "2003-04-24T07:04:53Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![Jinx](https://avatars.discourse-cdn.com/v4/letter/j/c6cbf5/32.png) [@Jinx](https://boards.straightdope.com/u/Jinx)\
**Post date:** [April 24, 2003, 7:04am UTC](https://boards.straightdope.com/t/margin-of-error/170811/1 "2003-04-24T07:04:53Z")

</div>

If I said to you something has a 4% margin of error, is that:  
a) +4% and -4%

- OR -

b) +2% and -2%

I say it is the latter such that the spread is 4%, correct?

- Jinx

---

<div class="post-metadata">

**Author:** ![Dolomite21](https://avatars.discourse-cdn.com/v4/letter/d/e5b9ba/32.png) [@Dolomite21](https://boards.straightdope.com/u/Dolomite21)\
**Post date:** [April 24, 2003, 7:14am UTC](https://boards.straightdope.com/t/margin-of-error/170811/2 "2003-04-24T07:14:16Z")

</div>

I believe it’s -4% to +4%

---

<div class="post-metadata">

**Author:** ![Aankh](https://avatars.discourse-cdn.com/v4/letter/a/ebca7d/32.png) [@Aankh](https://boards.straightdope.com/u/Aankh)\
**Post date:** [April 24, 2003, 7:32am UTC](https://boards.straightdope.com/t/margin-of-error/170811/3 "2003-04-24T07:32:07Z")

</div>

it means (-4) to (+4).

google came up with [this](http://www.amstat.org/sections/srms/brochures/margin.pdf) pdf document. refer page 4.

---

<div class="post-metadata">

**Author:** ![desdinova](https://avatars.discourse-cdn.com/v4/letter/d/8baadc/32.png) [@desdinova](https://boards.straightdope.com/u/desdinova)\
**Post date:** [April 24, 2003, 1:46pm UTC](https://boards.straightdope.com/t/margin-of-error/170811/4 "2003-04-24T13:46:04Z")

</div>

It could be either depending on the intent of the person that wrote it. But whenever I’m using this method of describing confidence intervals, I mean -4 to 4, and everywhere else I’ve seen it used where I can mathematically derive their intent, the authors intend it that way as well.

It’s a rather inefficient way of describing confidence intervals, a less ambiguous way is to say “with 95% certainty the true value lies between X and Y.” But “4 percent margin of error” is shorter and easier to understand.

---

<div class="post-metadata">

**Author:** ![Popup](https://avatars.discourse-cdn.com/v4/letter/p/94ad74/32.png) [@Popup](https://boards.straightdope.com/u/Popup)\
**Post date:** [April 24, 2003, 1:54pm UTC](https://boards.straightdope.com/t/margin-of-error/170811/5 "2003-04-24T13:54:42Z")

</div>

It normally means that there is a 95% confidence interval of [x-4%, x+4%].

However, in certain scientific/engineering branches it can mean that the **sigma** is 4%, which leads to a 95% confidence interval of ±8% (if I’m not misstaken).