This tutorial outlines the basics concepts in experimental design and statistics.
Learning outcomes
- Compare different types of research design
- Identify various research variable types
- Implement different types of descriptive statistics
- Identify appropriate sample types
- Explain the statistical assumptions and how to test each one
- Utilise inferential statistics flowcharts to select the appropriate test for your experiment
- Implement basic tests (T-test and ANOVA) in RStudio
Prerequisites
It is recommended that you have followed the Concepts in Computer Programming and Introduction to R tutorials before starting.
Approximate time to finish tutorial
- Lecture: 2.5 hours
- Tutorials: 1 hour
- Pre/post surveys: 10 minutes
Order of tutorial
Please do the pre-learning quiz, then watch the presentation.
During the presentation there are points to stop and do exercises, which are linked below. The answers to the questions in the exercises are linked within each one.
Towards the bottom of the page, there are worksheets for doing statistics in R or Excel.
Once finished the tutorial, take the post-learing quiz.
Statistics Pre-tutorial Survey
Presentation
Tasks from slides with sample answers
What type of research?
Qualitative vs quantitative
A survey of mask wearing habits in a student population - Qualitative
Assessing the impact of low GI diets on insulin production - Quantitative
Cross sectional vs longitudinal
The number of new cases of coronavirus globally on a single day? - Cross sectional
The change in vaccination rates between January and December 2021 - Longitudinal
####Experiment vs correlation
The relationship between the amount of glucose and level of insulin in the local population - Correlation
The amount of insulin produced by mice based on a defined range of glucose intake - Experimental
The impact of a plasmid presence/absence on tetracycline resistance in E. coli - Experimental (but badly worded - vague)
The presence of a plasmid in bacterial strains sensitive or resistant to tetracycline - Correlation (but badly worded - vague)
What type of data?
Variable types
Number of white blood cells in a sample - Discrete
Presence of inflammation - Binary
Sodium consumption split into low/medium/high - Ordinal
Gram stain result - Binary (or nominal if also counting inconclusive results)
Independent, dependent or control
The number of cells left after disinfecting - Dependent
The constant temperature throughout the experiment - Control
The amount of glucose added to the petri dish - Independent
Independent or paired
2 different growth conditions on the same set of bacterial strains - Paired
Mice in litter A given drug 1 and mice in litter B given drug 2 - Independent
Difference in glucose levels in the morning and in the evening in a patient - Paired
qPCR expression data is compared between clinical samples and a negative control sample - Independent (control refers to sample)
5 patient samples treated separately with a negative control and a drug - Paired (control refers to drug)
Which parametric statistical test?
Do 3 different drugs each have the same or different effects on insulin levels in mice? (different mice tested for each drug) - One-way ANOVA
Is there a difference in cell size at the start and end of a 12-hour period of treatment? - Paired t-test
Worksheets
Statistics in R
Descriptive statistics using R
Independent and Paired T-test using R
Normality and homogeny of variance testing using R
One Way ANOVA using R
Statistics test in Excel
NOTE It is not recommended to do inferential statsitics in Excel. You cannot test for the assumptions required for choosing between parametric and non-parametric tests. However, if you wish to understand the T-test and ANOVA, these worksheets will outline how to do them in excel.
Descriptive statistics in Excel are fine and can be completed following the worksheet
Descriptive statistics using Excel
Independent and Paired T-test using Excel
One Way ANOVA using Excel