Data: Dealing with Real Data
Content
The data that one encounters in empirical research are usually less ‘manicured’ than those presented in illustrations of data-analytical methods. One set of problems involves ‘missing data’: the absence of cases that ideally should have been observed but were not due to drop-out or attrition and cases where not all variables were measured. Another kind of problem is that measurement instruments may behave differently for different groups of cases (non-uniform calibration). Still other problems involve undesirable distributional properties of the data at hand such as heteroscedasticity or non-linear relationships between variables.
Solutions for these problems include, weighting, missing data imputation, synthetic data matching, modelling idiosyncratic calibration and data transformation. These methods are not always easy to implement, nor are they panaceas for the problems mentioned. This clinic introduces each of them, and discusses criteria to be used in deciding whether or not to use them.
The clinic also deals with the problem how to make complex or high-dimensional data structures amenable to analysis methods that require a more simple structure such as the 2-dimensional data matrix that is common in many forms of data analysis.
This clinic will be supplemented with online learning materials which can be accessed after the event.
Prerequisites
Understanding and familiarity with specific statistical models, e.g. multiple regression, analysis of variance, (ANOVA) or factor analysis. If you are attending a clinic as part of our Researcher Development Initiative then you automatically fulfil the prerequisites .