Learning outcomes of the course unit
The spread of computer science has made statistical tests accessible to everyone and not only to researchers and specialists. This large growth in their use has not been matched by a similar increase in understanding of the methodology: what questions it is possible to answer, how data collection should be planned in nature and in the laboratory, what tests should be applied, what conditions should be complied with, so that they are considered universally valid by the scientific community.
From thesis work to international relations, every publication based on the interpretation of data requires correct statistical analysis, in order to be recognized as scientifically reliable and allow comparison with the results of other research.
Descriptive statistics topics and some models or theoretical distributions will be dealt with in the course; the largest part will be dedicated to inference, with classic parametric statistics tests.
Elements of mathematics as per secondary school syllabus.
Course contents summary
types of data, scales, tables
graphical representation of data
measurements of central trend, dispersion, shape (symmetry and kurtosis)
outline of combinatorial analysis: simple groupings (permutations, arrangements, combinations)
univariate inferential statistics
hypothesis and hypothesis testing
null hypothesis and alternative hypothesis
concepts of risk (or error) in statistics
tables of critical values
discrete and continuous statistical distributions: binomial, Poisson, normal (or Gaussian) and standardized normal
comparison between observed and expected distribution in small samples (Kolmogorov-Smirnov test)
comparison between two observed distributions: 2 x 2 tables, independence test
2 x n and m x n contingency tables
Fisher exact method
comparison of the means of two samples
Student's t test for two dependent samples (paired data)
Student's t test for two independent samples (non-paired data)
confidence interval of a mean with known and unknown standard deviation
analysis of variance (ANOVA) to a classification criterion (completely random sampling)
test for variance homogeneity
components of the ANOVA
variance to a classification criteria (randomized blocks)
simple linear regression
concept of cause-effect
stochastic and deterministic variable
observed Y values and estimated or expected Y values
intercept and angular coefficient
method of least squares
F test on the regression line
coefficient of determination
MANUALE DI STATISTICA PER LA RICERCA E LA PROFESSIONE NELLE DISCIPLINE AMBIENTALI E BIOLOGICHE, L. Soliani, Uni.nova, Parma
The abbreviated version of the text is sold in the form of lecture notes by the publisher UNI.NOVA of Parma, specialized in university texts, at a price that is below the cost of printing it with a computer: firstname.lastname@example.org
Camussi et al. - Analisi statistica per la sperimentazione biologica - Zanichelli Bologna
Over 2000 pages of free lecture notes can be found on the Internet:
Soliani Lamberto, con la collaborazione di Franco Sartore e Enzo Siri, (2004). Manuale di Statistica per la Ricerca e la Professione. Statistica univariata e bivariata, parametrica e non parametrica, nelle discipline ambientali e biologiche.
Classroom lectures, with short exercises during lecture hours.
Final assessment is by means of multiple choice questionnaire or oral exam.