Data Analysis in Sociology
- The 1st year aims at beginners and serves to develop skills necessary to solve typical problems in analysing social data in R software environment.
- Student can formulate research goals, objectives and methods according to academic standards, can process the data from international sources; can assess the quality and analyze given samples of international research; can replicate the procedures from international studies; can carry out research projects in international teams
- Students can carry out statistical analyses of a data set, propose hypotheses and choose the methods needed to reach the goals, interpret the results and assess the quality of proposed solutions. Students provide reasons for their choice of techniques, interpret the outputs correctly, and assess the quality of their own and others’ models
- Student can set research goals, propose a research plan based on the results of previous research and social theory, carry out data analysis and report the results
- Students can apply a theoretical framework to define hypotheses and explain the results of a study; can apply appropriate statistical models and generalize the results
- Topic 1. Research hypotheses vs. statistical hypotheses. Variable typesThe cycle of research. Data analysis as part of the research process. Goals of data analysis. Theory of sampling recap. Survey data, behavior data. Posing and testing hypotheses. Research hypotheses vs. statistical hypotheses testing. Directed and non-directed hypotheses. Dependent and independent variables. Variable scales: nominal, ordinal, continuous (interval and ratio). Descriptive statistics of a variable depending on its type. Getting to know R and RStudio.
- Topic 2. Exploratory data analysisCentral tendency measures. Mean, median, mode. Standard normal distribution and its use. Z-scores. Moments of distributions. Distribution plots and reading them. Sources of bias in data. Interpretation of z-scores. Mean as a data model. Creating objects, types of objects, basic functions in R. Descriptive statistics in R. Tidy data. Plots for univariate and bivariate distributions. Histogram, bar plot, box plot, scatterplot, stacked bar plot.
- Topic 3. Chi-square and association measuresObserved and expected frequencies. Measures of association for categorical variables. Reading and interpreting chi-square tests. Assumptions of chi-square. Independence. Standardized residuals. Odds ratio. Chi-square and other association measures in R.
- Topic 4. Two means comparisonIndependent and paired samples. Assumptions behind the t-test. Two-sample t-tests. Testing for normality and homogeneity of variance. Nonparametric tests for two samples. Reading and interpreting means comparison. Confidence intervals. Weights in surveys. Post-stratification weights and design weights. Means comparison in R.
- Topic 5. One-way ANOVAOne-way analysis of variance (ANOVA). Difference from association measures and t-test. Assumptions and usage of ANOVA. Between-group and within-group variance, their ratio. Planned and non-planned comparisons; corrections. Post hoc comparisons for equal and unequal variances. Reading and interpreting ANOVA. Plots of post hoc tests. Non-parametric tests for multiple comparisons. One-way ANOVA in R. Presenting the results of ANOVA. Getting to know RMarkdown: reports and slide shows.
- Topic 6. Correlation and linear regressionIdea of correlations. Pearson’s product-moment correlation. Research problems for correlational analysis. Correlation coefficients for different types of data. Correlation matrices. ANOVA, correlation, and regression as linear models. Building a linear regression. Ordinary least squares. Fitting the regression line. Assumptions behind linear regression. Reading and interpreting regression coefficients. Goodness of fit measures. Presenting and interpreting a linear regression. Categorical predictors in a linear regression. Dummy-coding. Linear regression in R. Plotting linear regressions in R (case studies).
- Topic 7. Multiple linear regressionDifference between simple and multiple linear regression. The concept of interaction effects for categorical by categorical, categorical by continuous, and continuous by continuous variables. Multicollinearity. Centering. Model selection in multiple regression. Reading and interpreting interaction models in a linear regression. Testing for interactions in R. Graphs for interaction effects. Reporting and interpreting a linear regression with interactions in R.
- Projects 1-4 (0.1 * 4)Students create teams of 2-3 and work together on their project during the whole course, submitting and peer-reviewing them by each computer lab. Final projects are submitted in full and presented in the classroom. Each group selects one country from the European Social Survey, then picks the topic of interest within the scope of available survey questions (e.g. Health, Democracy, Religion, etc.) and performs all the tests covered in class on these data. One day before each computer lab, the due piece of work is to be submitted and blindly peerreviewed by two other groups in LMS. The instructors would assign reviewers, while students might not know who would be their reviewers next time. Final projects are presented in two steps. At the first stage, the group submits the code with interpretations. After this, they present the findings and procedures in class. Students are expected to choose and perform correctly the ways to analyse and interpret the data, as well as to demonstrate their knowledge and skills in presenting these results to the audience. Individual contribution of each student is graded. Projects themselves should be submitted as scripts or RMarkdown objects; in-class presentations should be adapted for the slide shows (e.g. Prezi, LibreOffice Impress, etc.). Project details are available in LMS
- TestAll students fill in a comprehensive paper-and-pencil test covering all previous topics.
- In-class activityIn-class activity during lectures and seminars. Students are expected to ask questions and participate in discussions, as well as help other students during practice sessions. Small regular tests held at seminars are also part of this grade.
- ExamThe exam is aimed at checking the skills students should have obtained during the course. Its structure is close to the structure of projects but covers all the topics: standard problems including descriptive statistics, measures of association, comparing two or more means, and linear regression.
- MOOCIf there are medical reasons for not completing the task, students should talk to the instructor before the deadline and arrange a later submission. Medical certificates should be presented no later than two weeks after the deadline, otherwise they are not taken into account.
- Interim assessment (4 module)0.2 * Exam + 0.2 * In-class activity + 0.1 * MOOC + 0.4 * Projects 1-4 (0.1 * 4) + 0.1 * Test
- Stowell, S. (2014). Using R for Statistics. Berkeley, CA: Apress. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=1174344
- Field, A. V. (DE-588)128714581, (DE-627)378310763, (DE-576)186310501, aut. (2012). Discovering statistics using R Andy Field, Jeremy Miles, Zoë Field. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edswao&AN=edswao.363067604