MEDI362 Lecture Notes - Lecture 8: Scatter Plot, Biostatistics, Contingency Table
7 Aug 2018
School
Department
Course
Professor
Document Summary
Data cleaning: aims to minimise and correct errors or at least minimise their impact on study results. Limited by the software package: commences as a quality assurance check, double data entry. Which test to use: consider your research question, consider your data format. What is your hypothesis: must get research question down to a single question on which you base primary analysis, what is the most important question, this is what you can power the study on, will have secondary questions. Independent samples t test: one sample t test, one way anova (kruskal wallis, chi square, trend, correlation, regression, paired t test (2 measures per person, repeated measures anova (>2. Linear mixed model: mcnemar (change in category e. g. malnutrition score over time) What sort of data do you have: quantitative, discrete (number of children in household, continuous (weight, height, age, categorical may be, unordered (e. g. country of birth, ordered (e. g. malnutrition coding normal-moderate-severe)
