Abstract:

It’s becoming increasingly obvious that statistics is an important component of many disciplines, in education and research. Students need to have a good training in statistics given how important it is in many scientific disciplines or otherwise. But it isn’t just the statistics that is important, its statistical software. Teachers should support to the students in the lab experimental work with tools that they will find later in life because most of the times, quite a high number of students don’t appreciate that they have to take data analysis or research studies as part of their chosen course.

Following we describe the experimental lab work proposed to students in subjects of Science or Engineering related with the Environment, where they have to apply statistical software. In this experimental lab course, students have to take at least 15 hours of compulsory classes. Students has to optimized the concentration of activated carbon, the pH and time by means of an incomplete factorial design [1], in order to study the influence of the entrapped activated carbon (X1), pH (X2) and time (X3) on the reduction of colored compounds of waste water from winery industry. A quadratic function is obtained for the three variables. Experimental data will be analyzed by the students, using the Response Surface method (Statistica 7.0 software).

The standardized (coded) dimensionless independent variables employed, with variation limits (-1, 1), will be defined as x1 (coded entrapped activated carbon), x2 (pH) and x3 (temperature). Students has to establish the correspondence between coded and uncoded variables by linear equations deduced from their respective variation limits.

Coded variables are then assigned values of -1, 0 and +1, corresponding to the lowest, central and maximum variation limits for each variable. Thus, the response surface obtained from the coded variables is not influenced by the magnitude of each variable, allowing the combination of factors into a dimensionless scale.

Consequently, students will carry out an incomplete factorial design of 15 experiments, to optimize the following independent variables: concentration of entrapped activated carbon (1 – 5 %), pH (4 - 7) and time (30 - 120 min); and the dependent variable studied will consist on the reduction of colored compounds. The utilization of statistical software, like Statistical 7.0 software, during laboratory courses in Science and Engineering, will encourage at the students later in life when they have to handle with real data or situations.

References:
[1] Box, G.E.P., Hunter, W.G., Hunter, J.S., 1978. Statistics for Experimenters: An Introduction to Design Data Analysis and Model Building, John Wiley, New York, 125–175.
[2] Bezerra, M.A., Santelli, R.E., Oliveira, E.P., Villar, L.S., Escaleira, L.A., 2008. Talanta 76, 965–977.