Skip to main content
. Author manuscript; available in PMC: 2021 Feb 25.
Published in final edited form as: Neuron Behav Data Anal Theory. 2020 Jul 21;3(5):https://nbdt.scholasticahq.com/article/13927-application-of-the-hierarchical-bootstrap-to-multi-level-data-in-neuroscience.

FIGURE 3.

FIGURE 3

Results from the simulation in which all data points were independent. a) A graphical representation of the experimental condition. As shown by the shades of group, subject and neuron means all being the same, the points are all independent in spite of an apparent hierarchical structure. b) Proportion of significant results when comparing the 2 groups with each statistical method at aof 0.05. As expected, the Traditional, Summarized and LMM methods give roughly 5% false positive results (black dashed line). However, the bootstrap gives a much smaller proportion of significant results suggesting a conservative bias. c) The size of SEMs computed using each of the methods. The bootstrap does give an error bar roughly 1.4 times that of the other metrics.