Description: It marks an error in the regression analysis of four variables, specifically when one of the values of the variable is equal to 2, any other value in this data does not cause an error. I mark it with an asterisk. Data: Y X2 X3 X4 4 *2 4 1 7 4 7 2 8 6 5 3 5 8 8 4 9 10 4 5 6 12 8 6 5 14 5 7 Steps to Reproduce: 1. Open Calc 2. Enter the data that I have shared with you in the comment, except do not include the asterisk. 3. go to Data/Statistics/Regression with the following options: Select all independent variables (X) and their names. select the dependent variable (E) and its name. check the checkbox indicating that variables have labels. select the cell where the results will be returned select the check box grouped by columns select the Linear Regression check box the confidence level of 0.95 do not check the Calculate Residuals box. Nor force intercept to zero Agree! Actual Results: results will appear with #¡VALOR! Regresión Modelo de regresión Lineal Salida bruta de ESTIMACION.LINEAL #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! Estadísticas de regresión R² #¡VALOR! Error estándar #¡VALOR! Recuento de variables X 3 Observaciones 7 R² ajustado #¡VALOR! Análisis de varianza (ANOVA) df SS MS F Precisión F Regresión 3 #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! Residuo #¡VALOR! #¡VALOR! #¡VALOR! Total #¡VALOR! #¡VALOR! Nivel de confianza 0.95 Coeficientes Error estándar Estadística t Valor P Inferior 95% Superior 95% Intersección #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! X2 #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! X3 #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! X4 #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! #¡VALOR! Expected Results: would expect values to appear Reproducible: Always User Profile Reset: No Additional Info: [Information automatically included from LibreOffice] Locale: es Module: SpreadsheetDocument [Information guessed from browser] OS: Windows (All) OS is 64bit: no
Created attachment 184913 [details] example error regression
Thank you Luis! I can see the same in: Version: 7.4.4.2 / LibreOffice Community Build ID: 85569322deea74ec9134968a29af2df5663baa21 CPU threads: 8; OS: Linux 5.15; UI render: default; VCL: gtk3 Locale: en-AU (en_AU.UTF-8); UI: en-US Calc: threaded and: Version: 7.6.0.0.alpha0+ (X86_64) / LibreOffice Community Build ID: e125e6623fa1c0f39d927bb37547ca6d1e299cb1 CPU threads: 8; OS: Linux 5.15; UI render: default; VCL: gtk3 Locale: en-AU (en_AU.UTF-8); UI: en-US Calc: threaded Same thing in: OpenOffice.org 3.3.0 OOO330m20 (Build:9567) Trying to find some logic in it, I think it has to do with having two series with a skewness of 0. Changing any number in X2 or X4 will change the skewness and give results in the LINEST array. MS Excel does *not* have the same issue. Microsoft® Excel® for Microsoft 365 MSO (Version 2212 Build 16.0.15928.20196) 64-bit
UI Regression dialog ask for independent X variables. As X2 = 2 * X4, your variables are not independent. X3 has no effect on the bug.