Jump to content

Talk:Variance

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia


Consistency of sample variance

[edit]

The article currently states: "If the conditions of the law of large numbers hold for the squared observations, s^2 is a consistent estimator of σ^2. One can see indeed that the variance of the estimator tends asymptotically to zero. An asymptotically equivalent formula was given in Kenney and Keeping (1951:164), Rose and Smith (2002:264), and Weisstein (n.d.).[15][16][17]"

I'm not certain this is completely accurate. As long as the population variance exists, then the sample variance should be a consistent estimator of it. The statement about the law of large numbers here seems to be confusing or misleading. If I am misunderstanding something, then the text should be edited to clarify what is meant. Which law of large numbers? Kenny and Keeping reference link is broken too. Jraudhi (talk) 21:16, 29 January 2020 (UTC)[reply]

Notation: several (possible) issues

[edit]

With respect to the section "Matrix notation for the variance of a linear combination":

  • a central dot appears, presumably for the dot product of two vector: why not stick to the transpose notation used before in the same section?
  • a primed 1 appears: what does it mean?

Further, throughout approximately the first half of the article, I observe a rather consistent use of:

  • square brackets around the argument of the expectation operator;
  • round parentheses around the argument of the variance operator.

That changes approximately halfway the article. Is there a good reason for this that I fail to know of? Or can I change the notation according to my observations for the the first half of the article?Redav (talk) 13:56, 18 May 2020 (UTC)[reply]

Minus symbol missing?

[edit]

On Windows Firefox everything looks good.

But on Windows Google Chrome: I just noticed, that there is a miunus missing in third line of the formulas to "The expression for the variance can be expanded as follows:". The minus symbol shows up in the source text, but somehow I don't see it when viewing the website. 212.201.73.246 (talk) 17:43, 26 January 2021 (UTC)b[reply]

"Definition" section, picture

[edit]

The picture in this section, its part 4., shows width: n, hight: 2. The hight shown should be: σ^2 ("sigma squared"). (The larger-sized picture shows n x σ^2 - correctly.) --Szozdakosvi (talk) 16:51, 28 January 2023 (UTC)[reply]

I second this, the image should be updated so it shows sigma^2 correctly as the height of that rectangle representing the variance. It's currently misleading 128.172.99.27 (talk) 20:02, 27 March 2024 (UTC)[reply]