Using Residuals to Make Conclusions

#1
Hi,
I was recently given an assignment to analyse the residuals for 4 days of the following data. After plotting the residuals, what does this tell me about my data? I'm a bit confused with the context. Thanks.

1580660928197.png
1580660977836.png
1.3.jpg
I found the residuals by subtracting the observed value from the overall mean of the data (predicted value).
 

staassis

Active Member
#2
Do not understand your plot. Which software did you use? Please use some standard statistical package to produce a boxplot for the distribution of residuals in each group. Then see if the distribution is symmetric, fat-tailed, unimodal, etc. If you are performing data analysis using a statistical method requiring normality of residuals (e.g. ANOVA F-test), check whether the residuals are normally distributed (via Kolmogorov-Smirnov test, Shapiro-Wilk test, etc).
 
#3
Thanks for the reply. I used matlab for the residual plot and did it manually using number of bacteria - predicted number of bacteria (mean) for each day. This is a boxplot for the data. Boxplot.jpg
 

staassis

Active Member
#4
So analyze the plot for the distributional properties I mentioned (symmetry / asymmetry, tails, etc). To check unimodality, you may use histograms.