HiI am testing that data are normally distributed by using

Click For Summary

Discussion Overview

The discussion revolves around testing the normality of a dataset using the chi-square test. Participants explore the implications of a high chi-square coefficient and consider alternative tests for normality, such as the Kolmogorov-Smirnov test and Anderson-Darling test. The conversation includes considerations of degrees of freedom and the impact of data aggregation on test results.

Discussion Character

  • Exploratory, Technical explanation, Debate/contested

Main Points Raised

  • One participant reports a high chi-square coefficient of 173 while testing for normality, despite observing that the data fits a 3-sigma test and that the median and mode are equal.
  • Another participant notes that the central chi-square distribution approaches normality with a large number of degrees of freedom, suggesting that this could be relevant to the discussion.
  • A participant expresses concern about having insufficient degrees of freedom and questions whether the number of data points (310) is affecting the results.
  • Some participants mention that alternative tests like the Kolmogorov-Smirnov and Anderson-Darling tests yield satisfactory results, implying they may be more reliable in this context.
  • There is a suggestion that data aggregation might be influencing the interpretation of the results, which could be a factor in the high chi-square value.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the reliability of the chi-square test results, and there are competing views regarding the appropriateness of different tests for normality. The discussion remains unresolved regarding the implications of the high chi-square coefficient and the role of degrees of freedom.

Contextual Notes

Participants mention potential issues with degrees of freedom and data aggregation, but these factors remain unresolved in the context of the discussion.

Mark J.
Messages
81
Reaction score
0
Hi
I am testing that data are normally distributed by using chi-square test.
I get a very high chi coefficient about 173.
This is very strange because data fits at 3sigma test and median mode are equal so I am almost sure that is normal distribution.
What I am doing wrong?
Which test do you advice for a better approach?
Regards
 
Physics news on Phys.org


Mark J. said:
Hi
I am testing that data are normally distributed by using chi-square test.
I get a very high chi coefficient about 173.
This is very strange because data fits at 3sigma test and median mode are equal so I am almost sure that is normal distribution.
What I am doing wrong?
Which test do you advice for a better approach?
Regards

The Central Chi Square will approach the normal distribution as the number of degrees of freedom gets large. For over 50 df, the distribution is effectively normal.
 


The problem is that I don't have so much degrees of freedom to get this value.
The other tests Kolmogorov or AD work just fine.
Maybe I am miscalculating degrees of freedom or 310 data are an issue?
Please help
 


Mark J. said:
The problem is that I don't have so much degrees of freedom to get this value.
The other tests Kolmogorov or AD work just fine.
Maybe I am miscalculating degrees of freedom or 310 data are an issue?
Please help

Are you saying you have 310 data points? If they are independent observations that would mean you have 309 df. In this case, you certainly should be able to assume a normal distribution with the central chi square. When you say the other tests of normality work fine, do you mean they are consistent with normality? If so, you should be fine. However, I think you might be aggregating data in some way, and how you do this could be important in interpreting the data..
 

Similar threads

  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 11 ·
Replies
11
Views
3K
  • · Replies 6 ·
Replies
6
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 30 ·
2
Replies
30
Views
5K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 5 ·
Replies
5
Views
4K
Replies
1
Views
4K