 |
Experimental design for laboratory biologists : maximising information and improving reproducibility / Stanley E. Lazic. -- Cambridge, United Kingdom : Cambridge University Press, c2016 .—(58.105 /L431) |
Contents
Preface
Abbreviations
1 Introduction
1.1 What is reproducibility?
1.2 The psychology of scientific discovery
1.2.1 Seeing patterns in randomness
1.2.2 Not wanting to miss anything
1.2.3 Psychological cliff at p = 0.05
1.2.4 Neglect of sampling variability
1.2.5 Independence bias
1.2.6 Confirmation bias
1.2.7 Expectancy effects
1.2.8 Hindsight bias
1.2.9 Herding effect
1.2.10 How the biases combine
1.3 Are most published results wrong?
1.3.1 What statisticians say
1.3.2 What scientists say
1.3.3 Empirical evidence I: questionable research practices
1.3.4 Empirical evidence II: quality of studies
1.3.5 Empirical evidence III: reproducibility of studies
1.3.6 Empirical evidence IV: publication bias
1.3.7 Scientific culture not conducive to 'truth-finding'
1.3.8 Low prior probability of true effects
1.3.9 Main statistical sources of bias in experimental biology
1.4 Frequentist statistical inference
1.5 Which statistics software to use?
Further reading
2 Key Ideas in Experimental Design
2.1 Learning versus confirming experiments
2.2 The fundamental experimental design equation
2.3 Randomisation
2.4 Blocking
2.5 Blinding
2.6 Effect type: fixed versus random
2.7 Factor arrangement: crossed versus nested
2.8 Interactions between variables
2.9 Sampling
2.10 Use &controls
2.11 Front-aligned versus end-aligned designs
2.12 Heterogeneity and confounding
2.12.1 Batches
2.12.2 Plates, arrays, chips, and gels
2.12.3 Cages, pens, and tanks
2.12.4 Subject/sample characteristics
2.12.5 Litters
2.12.6 Experimenter characteristics
2.12.7 Time effects
2.12.8 Spatial effects
2.12.9 Useful confounding
Further reading
3 Replication (what is 'N'?)
3.1 Biological units
3.2 Experimental units
3.3 Observational units
3.4 Relationship between units
3.4.1 Randomisation at the top of the hierarchy
3.4.2 Randomisation at the bottom of the hierarchy
3.4.3 Randomisation at multiple levels
3.5 How is the experimental unit defined in other disciplines?
4 Analysis of Common Designs
4.1 Preliminary concepts
4.1.1 Partitioning the sum of squares
4.1.2 Counting degrees &freedom
4.1.3 Multiple comparisons
4.2 Background to the designs
4.3 Completely randomised designs
4.3.1 One factor, two groups
4.3.2 One factor, multiple groups
4.3.3 Two factors, crossed
4.3.4 One factor with subsamples (pseudoreplication)
4.3.5 One factor with a covariate
4.4 Randomised block designs
4.4.1 With no replication
4.4.2 With genuine replication
4.4.3 With pseudoreplication
4.5 Split-unit designs
4.6 Repeated measures designs
Further reading
5 Planning for Success
5.1 Choosing a good outcome variable
5.1.1 Qualitative criteria
5.1.2 Statistical criteria
5.2 Power analysis and sample size calculations
5.2.1 Calculating the sample size
5.2.2 Calculating power
5.2.3 Calculating the minimum detectable effect
5.2.4 Power curves
5.2.5 Simulation-based power analysis
5.3 Optimal experimental designs (rules of thumb)
5.3.1 Use equal n with two groups
5.3.2 Use more controls when comparing multiple groups to the control
5.3.3 Use fewer factor levels
5.3.4 Increase the variance of predictor variables
5.3.5 Ensure predictor variables are uncorrelated
5.3.6 Space observations out temporally and spatially
5.3.7 Sample more intensively where change is faster
5.3.8 Make use of blocking and covariates
5.3.9 Crossed factors are better than nested
5.3.10 Add more samples instead of subsamples
5.3.11 Have 10 to 20 samples to estimate the error variance
5.4 When to stop collecting data?
5.5 Putting it all together
5.6 How to get lucky
5.7 The statistical analysis plan
5.7.1 Why bother?
5.7.2 What to include in the SAP
Further reading
6 Exploratory Data Analysis
6.1 Quality control checks
6.1.1 Data layout
6.1.2 Possible and plausible values
6.1.3 Uniqueness
6.1.4 Missing values
6.1.5 Factor arrangement
6.2 Preprocessing
6.2.1 Aggregating and summarising
6.2.2 Normalising and standardising
6.2.3 Correcting and adjusting
6.2.4 Transforming
6.2.5 Filtering
6.2.6 Combining
6.2.7 Pitfalls of preprocessing
6.3 Understanding the structure of the data
6.3.1 Shapes of distributions
6.3.2 Effects of interest
6.3.3 Spatial artefacts
6.3.4 Individual profiles
Further reading
Appendix A Introduction to R
A.1 Installing R
A.2 Writing and editing code
A.3 Basic commands
A.4 Obtaining help
A.5 Setting options
A.6 Loading and saving data
A.7 Objects, classes, and special values
A.8 Conditional evaluation
A.9 Creating functions
A.10 Subsetting and indexing
A.11 Looping and applying
A.12 Graphing data
A.13 Distributions
A.14 Fitting models
Appendix B Glossary
References
Index
