Author Archives: Martin Héroux

Top journals love their p-values

Researchers love p-values, especially when they are significant. It has previously been demonstrated that there is a disproportionate number of positive or significant p-values. In other words, many reported p-values must be false. That is to say, they are associated with false-positive findings. In a recent article, Cristae & Ioannidis (2018) investigated the characteristics of p-values reported in figures and

Read more

Research on research

Self reflection is important. As a human endeavour with the whole of society as its beneficiary, science should also have introspection; looking in on itself to identify things worth celebrating, as well as things that need improvement. In a recent article, Ioannidis explains why we need meta-research: research on research. Meta-research The goal of meta-research is to understand and improve

Read more

Multiple linear regression in R

In a previous blog, we applied simple linear regression to an interesting problem: how well does a measure of wine density account for alcohol content. This was considered simple linear regression because we had one outcome variable (alcohol content) and one predictor variable (wine density). We can extend this approach to have more than one predictor. Specifically, we can use

Read more
« Older Entries