coefficient of determination R2 - value different to excel

[beipink]

Hi there

Libreoffice Calc 5.3.02 → R2= 0.886442
MS excel 2017 → R2 = 0.5669

to calculate R2 in Libreoffice I follow this Trend Lines - LibreOffice Help
(select the chart and insert - trend lines - linear)

data used


1996 | 20.67

1997 | 19.09

1998 | 12.72

1999 | 17.97

2000 | 28.5

2001 | 24.44

2002 | 25.02

2003 | 28.83

2004 | 38.27

2005 | 54.52

2006 | 65.14

2007 | 72.39

2008 | 97.26

2009 | 61.67

2010 | 79.5

2011 | 111.26

2012 | 111.67

2013 | 108.66

2014 | 98.95

2015 | 52.39

2016 | 43.73


[erAck]

I get R2=0.58602 with LibreOffice 5.3.0 and 5.3.2

[RGB-es]

I cannot reproduce your problem: I get R² = 0,564196437315994 with both, a linear trend line and a normal linear regression (which is reasonable, considering the data). SciDAVis gives exactly the same value. The only way I can obtain a R² of 0.8 is with a 6th degree polynomial fit.

One possibility: be sure your graph is “XY (dispersion)”, otherwise the first column will be considered as labels instead of numbers.

[beipink]

I went back and tested it again side by side. LO 5.3.02 and MS Office 2016.
Same set of data different results as shown below and in the picture here https://ibb.co/i6rw5F

LO: Linear, Force intercept, Show Equation, Show Coefficient of Determination (R2)
results: f(x) = 4.87451827242525x | R2 = 0.886442665062278

MS Office: Linear, Set Intercept, Display Equation on chart, Display R-squared value on chart.
results: Y = 4.8745x | R2 = 0.5669.

can you advise how did you get R² = 0,5641 with LO