I've had to ask this question multiple times, to which a few had either been copied from the previous answer or from the internet so please, if you can produce a direct and concise answer it would be much appreciated. I have recently produced a linear regression model in R, to which everything seemed fine with the code and model, but my standard residual plots appear to be non-linear. Why is it that my qq plot and initial plot seem linear but the residual contradicts? Furthermore what does this mean in context to the topic? The topic is on the dependency of strength on body weight. Also, why is it that my r-squared value is high and my Shapiro test fails to reject the null hypothesis but my residual plot is non-linear The data I used was of a linear trend as well, as shown by the initial scatter plot. Context in the answer as well as why this has happened is vital in my understanding so sincerely please do your best I can’t attach more than 2 images so I will attach my dataset and results. Here’s the R script: Residuals: Min -10.531 -4.989 1Q Median 30 Max 1.869 5.267 7.950 Coefficients: Estimate Std. Error t value Pr (>|t|) (Intercept) 42.62755 6.86517 6.209 0.000157 *** 'Body Weight (kg)' 0.62692 0.07461 8.402 1.49e-05 *** Signif. codes: 0 ***** 0.001 **** 0.01 *** 0.05 '.' 0.1 ' ' 1 Residual standard error: 7.077 on 9 degrees of freedom Multiple R-squared: 0.8869, Adjusted R-squared: 0.8744 F-statistic: 70.6 on 1 and 9 DF, p-value: 1.492e-05 > abline (Regression.model, col=4, 1wd=3) > anova (Regression.model) Analysis of Variance Table Response: Bench press standards per class of body weight (kg) - Average Df Sum Sq Mean Sq F value Pr (>F) 'Body Weight (kg) 1 3536.0 3536.0 70.601 1.492e-05 *** Residuals 9 450.8 50.1 Signif. codes: 0 ***** 0.001 **** 0.01 *** 0.05 ' ' 0.1 ' ' 1 > stdres‹- rstandard (Regression.model) > print (stdres) 1 2 3 4 5 6 7 -1.4864868 -1. 0645996 -0.5043759 0.2843252 1.1786018 0. 6513863 1.1083291 11 0.9248828 0.5985440 -0.3245869 -2.0230776 > plot ('Body Weight (kg)', stdres, main="Std residuals versus explanatory variable") > fits‹- fitted (Regression.model) > plot (fits, stdres, main="Std residuals versus fits") > qqnorm (stdres, main="e-l Plot") > galine (stdres) > shapiro. test (stdres) Shapiro-Wilk normality test data: stares W = 0.91857, p-value = 0.3069 Sorry for the mess of it, as I’ve said I can’t attach more than 2 images.