Search

The meaning of error bars is often misinterpreted, as is the statistical significance of their overlap.

Use box plots to illustrate the spread and differences of samples.

Appeal to intuition when making value judgments

Make navigational elements distinct and unobtrusive to maintain visual priority of data.

Choose distinct symbols that overlap without ambiguity and communicate relationships in data.

Limitations in print resolution and visual acuity impose limits on data density and detail.

Relate your data to the world around them using the age-old custom of telling a story.

Complete maps of DNA methylation in human cells provide insight into the epigenetic regulation of pluripotency.

Carefully designed subplots scaled to the data are often superior to a single complex overview plot.

Use alignment and consistency to untangle complex circuit diagrams.

Figure labels require the same consistency and alignment in their layout as text.

Translate the principles of effective writing to the process of figure design.

Visually organize complex data by mapping them onto familiar representations of biological systems.

Apply visual grouping principles to add clarity to information flow in pathway diagrams.

Constraining the magnitude of parameters of a model can control its complexity

Clustering finds patterns in data—whether they are there or not.

With four parameters I can fit an elephant and with five I can make him wiggle his trunk. —John von Neumann

Tabulating the number of objects in categories of interest dates back to the earliest records of commerce and population censuses.

A P value measures a sample's compatibility with a hypothesis, not the truth of the hypothesis.

It is important to understand both what a classification metric expresses and what it hides.

The P value reported by tests is a probabilistic significance, not a biological one.

When some factors are harder to vary than others, a split plot design can be efficient.

Good experimental designs mitigate experimental error and the impact of factors not under study.

The ability to detect experimental effects is undermined in studies that lack power.

Today's predictions are tomorrow's priors.

Incorporate new evidence to update prior information.

Quality is often more important than quantity.

Robustly comparing pairs of independent or related samples requires different approaches to the t-test.

When multiple variables are associated with a response, the interpretation of a prediction equation is seldom simple.

When a large number of tests are performed, P values must be interpreted differently.

Statistics does not tell us whether we are right. It tells us the chances of being wrong.

Good experimental designs limit the impact of variability and reduce sample-size requirements.

Nonparametric tests robustly compare skewed or ranked data.

Residual plots can be used to validate assumptions about the regression model.

For studies with hierarchical noise sources, use a nested analysis of variance approach.

When multiple factors can affect a system, allowing for interaction can increase sensitivity.

Regression can be used on categorical responses to estimate probabilities and to classify.

Some outliers influence the regression fit more than others.