GBonnaire / jspreadsheet-plugins-and-editors

Plugins and Editors for JSpreadsheet Pro 7, Pro 8 and Pro 9
https://repo.gbonnaire.fr/category/jspreadsheetpro
MIT License
13 stars 0 forks source link

STUDENT T-TEST PAIRED/ Discrepancies in calculation with MS Excel #30

Closed tp-development closed 1 year ago

tp-development commented 1 year ago

Plugin Name with version

3.2.2

JSpreadsheet version

v9.x

Describe the bug

See a spreadsheet with an example - there is still a discrepancy between Excel and stat package: https://docs.google.com/spreadsheets/d/1oBX-outkj2eX-kpwMW_bfOlwV_PCm6y3/edit?usp=sharing&ouid=105291088824732366674&rtpof=true&sd=true

Reproduction steps

1.
2.
3.
...
GBonnaire commented 1 year ago

Hi,

It's a difference because you compare paired result with unpaired result, on plugin, i have implemented only unpaired values. i search to implement paired

GBonnaire commented 1 year ago

I found a solution 😊 i will be available on 2.2.2

Warning : New syntax of formula TTEST =TTEST.TWOSAMPLES.T(A1:A100,B1:B100,FALSE)

TTEST.TWOSAMPLES.T(firstColumn → Range, secondColumn → Range, [pairedValue = FALSE])

tp-development commented 1 year ago

Do we need the additional parameter? There was already a parameter dependant which was to indicate that samples are paired(?), wasn't it?

GBonnaire commented 1 year ago

dependance is replace by pairedValue

You have now 3 parameters : firstColumn : Sample1 SecondColumn: Sample2 pairedValue : Indicate if values are dependant or not

tp-development commented 1 year ago

Ok, that makes sense. I was wondering if the name shouldn't be pairedSamples instead of pairedValue (because we have multiple values here, and the user might not be sure which value is meant here).

GBonnaire commented 1 year ago

Ok, i will change 👍 You prefere :

tp-development commented 1 year ago

my vote goes to: pairedSamples (true/false) :)

GBonnaire commented 1 year ago

It's on line on v2.2.3 😊

tp-development commented 1 year ago

After attaching 2.2.3 the changes are not visible (lack of the new parameter).