Statistics III

Code:

P406

Acronym:

ESTIII

Keywords
Classification	Keyword
OFICIAL	Scientific Research Methodology

Instance: 2012/2013 - 1S

Active?	Yes
Responsible unit:	Psychology
Course/CS Responsible:	Integrated Master Psychology

Cycles of Study/Courses

Acronym	No. of Students	Study Plan	Curricular Years	Credits UCN	Credits ECTS	Contact hours	Total Time
MIPSI	139	Official Curricular Structure 2007/2008	2	-	6	54	162
		Official Curricular Structure	2	-	6	54	162
		Official Curricular Structure 2012	2	-	6	54	162

Teaching language

Portuguese

Objectives

1. GENERAL AIMS

- Elaborate the knowledge acquired in the two preceding CUs (Statistics I and II)

- Understand the relation between research methods and data analyses procedures

- Identify the conceptual the status of the variables within a plan of data analyses: levels of measurement, relation between variables, IVs and DVs, etc.

- Definition/identification of hypotheses, or “research questions”: taking into perspective the statistics procedures required to test them (alternative possibilities and respective limits of usage)

- Acquisition of basic knowledge regarding the conceptual statistical framework of every statistics test presented

- Training in the “practice” of data analysis with the software IBM SPSS Statistics 19

- Analysis, description and interpretation of results: presentation according to the APA rules

2. CONCEPTUAL ORIENTATION OF THE CU

- In each module we make explicit the extent to which the contents of Statistics I and II are articulated and necessary in this UC. However,

- In terms of conceptual orientation, in this CU we refer systematically to the “subsidiary” character of data analysis relatively to the research methods that sustain them. We therefore stress the relation with other UCs that focus on research methodologies.

- In the setting of the “statistical reasoning” of each of the tests presented, the tone is made on the principle of contrasting systematic variance vs. error variance in the data. This principle is evident in the ANOVA models or in PCA, but we demonstrate its presence in several models of increasing complexity since Statistics I (t tests, Simple and rm ANOVAs, r, MR, etc.)

- Strong emphasis and valuation of specific skills of interpretation and presentation of statics results, at all times supported by data analyses adequate to the research’s goals.

Program

Considering, at least, 13 classes (3½H)

1-2 – Procedures to prepare and manipulate data files: data entry errors, treatment of missing data, detection routines and dealing with outliers

3-4 – Logistic regression: model estimations and predictors analysis (Log-L, R e R2, Wald, Exp b)

4-6 – ANOVAs with more than 1 factor: factorial and mixed-design models (from the statistics rationale of each model to the analysis of main effects and the decomposition of up to 2nd order interactions)

6-7 – Multivariate ANOVA: conceptual contrast with multiple ANOVAs on the DVs, model’s rationale and the variance matrices, multivariate and univariate tests

7-8 – Covariance analysis in ANOVAs with more than 2 factors: ANCOVA and MANCOVA

9* - Cluster analysis with numeric variables – the K-Means method

10-11 – Analysis of reliability: Cronbach’s alpha, KR20, Split-Half procedures

11-13 – Principal Components Analysis: from the principle of Common vs. Unique Variance to the interpretation of factorial structures

Note: This module may be excluded if the progression on the CU's programmed contents justifies it.

Mandatory literature

Bryman Alan; Quantitative data analysis with SPSS for windows. ISBN: 0-415-14720-4
Hair Joseph F. 070; Análise multivariada de dados. ISBN: 85-363-0482-0
Howell David C.; statistical methods for psychology. ISBN: 0-534-51993-8
Reis Harry T. 340; handbook of research methods in social and personality psychology. ISBN: 0-521-55903-0
Tabachnick Barbara G.; Using multivariate statistics. ISBN: 0-321-05677-9

Complementary Bibliography

Judd Charles M.; Research methods in social relations. ISBN: 0-03031149-7
Kinnear Paul R.; SPSS for windows made simple. ISBN: 1-84169-118-6
Rosenthal, R., Rosnow, R. L., & Rubin, D. B.; Contrasts and Effect Sizes in Behavioral Research: A Correlational Approach, Cambridge University Press, 1999
Toothaker Larry E.; Multiple comparison procedures. ISBN: 0-8039-4177-3
Klockars, A. J., & Sax, G.; Multiple Comparisons: Structural Models for Qualitative Data, Sage Publications, Inc., 1986
Winer B. J.; Statistical principles in experimental design

Teaching methods and learning activities

1. STRUCTURE OF TEACHING METHODS

- Each class has a theory component relative to the statistics rationale of a test or group of tests. When considered fitting, we go back to the acquisition made on previous UCs, making it explicit how their increasingly complex learning is based on common “statistical concepts”.

- Modules’ contents are organized in such that every class has a practical component of data analysis with the SPSS. More than the “mechanics” of the process of executing an analysis, we stress the importance of knowing the statistical rationale that allows the necessary statistical decision making. Emphasis is made on the training of specific skills of analysis, description and results interpretation.

- Each module has support materials, namely one data file for the purpose of exercise. Students take part in an on-line study. Its objective is to be negotiated in the first class, and will aloe them to analyze their own data during the course.

2. CONCEPTUAL ORIENTATION OF TEACHING METHODS

- The organization of module’s contents in such a way that every class has a theory component and a practical one aims to serve a double objective: On the one hand, reinforcing the “applied” character of statistics learning with the immediate demonstration of “what does it serve for” the acquired theoretical knowledge; on the other, potentiate students orientation towards data analysis as a process that implies decision making supported by basic statistics and methodology knowledge.

- The statistical rationale of each statistic procedure is discussed with students and “broken down” in its more simple parts that go back, in most cases, to the basic principles acquired since Statistics I. Otherwise, complex analyses such as a PCA, or the test of an hypothesis that requires decomposing a 2nd order interaction, can become vacant exercises for 3rd semester students.

- Programmed contents are defined in such a manner that, when justified from the pedagogical point of view, parts of it can be excluded or some additional contents can be included. For instance, Kluster analysis is programed for a moment in which its exclusion does not affect the coherence of the syllabus. If, on the contrary, learning rhythms surpass what is outlined, additional contents will be included in the modules that justify it.

- All analyses performed in the class are conducted with data from psychological research either ongoing or already completed. By doing so, the use of any statistics test will always have the purpose of answering a question or testing a “real” hypothesis.
From the pedagogical standpoint, the use of data files from actual research is based on the premise of training data analysis skills with a structure provided by the professor.

- In different moments of each module students will have to use the data file from the study conducted in the UC. In these “free analysis” moments the only rule external to the class dynamics is the obligation of applying, at least once, the ongoing module’s contents. Students will be encouraged to perform analyses that make use of previous modules contents, as well as those from Statistics I and II.

Evaluation Type

Distributed evaluation with final exam

Assessment Components

Description	Type	Time (hours)	Weight (%)	End date
Attendance (estimated)	Participação presencial	45,50
	Total:	-	0,00

Eligibility for exams

Attendance in 3/4 of the classes is mandatory.