preview

Manual 621 Assignment 4 : Clustered Data Essay

Decent Essays
Open Document

BIOS 621 Assignment 4: Clustered Data # I. Introduction Clustered data can be generated in a wide variety of settings. For instance, patients’ medical records collected throughout a city might be clustered at the clinic- and the health-practitioner-level.^1 In such cases, the individual observations cannot be considered independent and should instead be analyzed as clustered.^1 Unfortunately, many investigators using clustered data either don’t recognize it as such or aren’t aware of its implications for statistical analysis and fit incorrect models that don’t take clustering into account.^1 According to Erika L. Moen et al., this is especially prevalent in neuroscience, where experiments often produce data measured on the neuron level whose natural grouping at the mouse level is too often ignored.^2 Scientists with few available mice are especially prone to this mistake as they try to maximize their sample size by counting individual measures from the same mouse independently.^2 In their paper “Analyzing Clustered Data: Why and How to Account for Multiple Observations Nested within a Study Participant?” Moen et al. wrote that multiple measures taken from the same subject (in neuroscience subjects are most often mice) must be viewed as correlated, even when effects are tested at the level of individual neurons.^2 To illustrate this, Moe et al. analyzed the data from one set of experiments using multiple naïve (i.e. non-correlated) and correlated models and compared the

Get Access
Related
Get Access