Nowadays quality programs have been receiving great attention, so methods and techniques for control and identify variations and outliers are being needed more and more. This paper shows the use of boxplots, very simple, to identify outliers in measures of protein and ash on doubles of 30 wheat flours. The same approach in this article may be used in any other material, analysis and number of repetitions. Were employed Kjeldahl (half micro) for protein and incineration for ashes. The results of protein (60 observations) showed lower and upper limits 12.89 and 15.08%, and the limits for ashes were 1.35 and 1.99%. There were no outliers for protein and ashes. Nevertheless, refering to the absolute value of the differences between the doubles in the 30 flours, the lower limit for protein was 0.010 and the upper was 0.295%, while those limits for ashes were 0.01 and 0.19%. Two outliers in the differences for protein were identified, corresponding to flours 2 and 8 with values 0.40 and 0.31, while three outliers in the differences for ashes were identified, wich corresponded to flours 9, 15 and 16, with values 0.21, 0.26 and 0.39%. Those flours whose differences were found being outliers must be analysed again and no value, outlier or not, must be dismissed.
outliers; boxplot; quality for analysis laboratories; reproducibility; communication of data image
