MEETING OF MAY 23TH 1984. ENTERING AND PROCESSING DATA
Epidemiological data analysis needs the application of a simple, albeit very general, method. Aims definition, hypothesis proposition, choice of a strategy, treatments propositions (variable transformations, statistics …), tries, and then the real analysis. This whole process must be drawn as soon as the plan of epidemiological data collection is thought, unless the data are already existing. In this last case, the critical analysis of data is essential. Existence of data processing programs running on microcomputers is important, but yet not as much as a good thinking of the treatment process.
With the help of examples, the author gives basic information’s about common statistical tests: frequency table’s analysis, variance analysis, linear regression model.
With the help of cooking recipes used as an example, the author gives basic way to read a multivariate analysis.
Madec (F) & Josse (J)
The diseases of multifactorial aetiology cause great economic losses in pig production. Experimental trials are difficult to carry on and they are hardly conclusive, while chemiotherapy is expensive and gives irregular results. In order to point out the environmental conditions regularly associated with the troubles, a detailed review of the herd management is required. So a great number of variables are involved and the statistical study becomes complex. The multidimensional data analyse methods are of good opportunity. Although, it is necessary to get an appropriate complementary software to prepare the data board and to help rendering the results obtained. Further, eco-pathological studies lead to follow a general proceeding of attentive observation, analysis and action in order to improve the failing parameters among the risk factors.
Eloit (M) & Koutchoukali (M-A)
The authors present the principles leading to the sampling rules to be used when looking for the infection rate of herds in case of an infection disease within an animal population. The number of herds to study, as well as the number of animals to test in each herd, are presented, following different situations.
A list of the main parameters which characterize the pig farming is presented. This list contains ten rubrics: “Environment”, “Animals”, “Farm-Building”, “Feed”, “Watering”, “Hygiene”, “Systematic interventions”, “Pathology”, “Performance”,” Previous analysis”.