topic Re: Why does JMP Not calculate STDEV correct in Discussions

Why does JMP Not calculate STDEV correct

MikeDereviankin — Fri, 09 Jun 2023 14:42:20 GMT

JMP is over estimating the STDEV. In R/Excel this number is less than the mean. Why is JMP calculating these over estimations?

Re: Why does JMP Not calculate STDEV correct

Dan_Obermiller — Fri, 08 Apr 2022 16:15:42 GMT

I think you will need to provide more details. JMP does indeed correctly calculate a standard deviation. So, a few questions:

* What version of JMP are you using?

* Can you share some of your data, even if it is anonymized? You may have your data formatted in an unusual way.

* How much data do you have for each of these categories?

* Have you verified these exact same calculations with this exact same data in R or Excel?

Remember that there is no requirement that a standard deviation be less than a mean. For skewed distributions, the standard deviation could certainly be larger than the mean.

Re: Why does JMP Not calculate STDEV correct

MikeDereviankin — Sat, 09 Apr 2022 15:24:55 GMT

Using JMP 16 Pro.

I've attached some data where you can calculate the standard deviation per the 3 factors above (Gender, Age, Ethnicity) for each Congener. You'll notice that this calculation will indeed deviate if you run it with R. Even when analyzing these distributions or doing a normalization technique (log normalization), you'll note that the standard deviations are still completely off.

Attached is the data and a summary table I generated in R. JMP cannot replicate these calculations.

Re: Why does JMP Not calculate STDEV correct

Craige_Hales — Sat, 09 Apr 2022 17:21:03 GMT

Are you using a script like this?

Data Table( "NHANES 2003-2004" ) << Summary(
    Group( :Gender, :Age, :Ethnicity, :Congener ),
    Mean( :Lipid Weight ),
    Std Dev( :Lipid Weight ),
    Freq( "None" ),
    Weight( "None" )
)

The largest difference seems to agree to the resolution of the CSV numbers.

Re: Why does JMP Not calculate STDEV correct

Georg — Sun, 10 Apr 2022 09:56:36 GMT

Dear @MikeDereviankin ,

I cannot reproduce the problem.

When importing your data into JMP (JMP Pro 16.1, Win10), I get exactly the same numbers (see Tabulate left) than you posted in your Summary (Summary Table right).

Perhaps there is an issue with importing the data?

Re: Why does JMP Not calculate STDEV correct

Dan_Obermiller — Sun, 10 Apr 2022 15:09:20 GMT

The CSV data is helpful, but the issue could be how you have your JMP table formatted. Could you attach your JMP file??

Re: Why does JMP Not calculate STDEV correct

Byron_JMP — Tue, 19 Apr 2022 14:31:10 GMT

Change Gender Age and Ethnicity to Nominal (not continuous)

I ran tabulate just like you did in your first post, and made a data table from it.

Then I used tables>compare on my table and your summary.csv

The tables are identical

coding data correctly is very important