Limitations

Tiptl;dr

The workflow aims to be general, rather than universal.

In this chapter, I consider constraints on the generality of the workflow, as well as some potential barriers to uptake.

Constraints on generality

The outlined workflow aims to be general enough for a majority of use cases that we come into contact with. But I’m sure there are exceptions where the workflow may not transfer so effectively, and which represent necessary constraints on the generality of the proposed workflow (Simons et al., 2017). For instance, maybe some research tools, such as MRI, EEG or pupillometry measures, generate specialised data formats and/or require specific software (FSL, SPM, EEGLAB) that may impose its own data processing pipelines. This is to be expected.

This is also where a workflow can be used as a lens to view a data science project through, rather than a toolchain with a handle that gets cranked. A one-size-fits-all workflow is unrealistic anyway, given the complex nature of scientific data and associated materials. Instead, a workflow can be seen as a toolkit, aspects of which can be deployed in a variety of contexts and across distinct projects. And one may not use any of the tools that I presented, but instead use another set of tools to satisfy the principle that it should be a minimum expectation and routine that scientific materials are reasonably easy to verify.

Barriers to uptake

As outlined in the Workflow overview section, the workflow that we use spans a number of interconnected tools. And depending on your familiarity with such tools, the sheer volume can be overwhelming. If you do decide to adopt aspects of this workflow, I suggest embracing principles of effective behaviour change (Poldrack, 2021). And watch this documentary on weight loss programme effectiveness, which is instructive and engaging when it comes to thinking about principles of effective behaviour change.

Some quick starting points would be:

  1. Don’t change everything at once, as it will be overwhelming and exhausting.
  2. Take a stepwise approach that builds in new behaviours incrementally.
  3. Wait for new behaviours to form habits, before adding new aspects to the workflow.
  4. Prioritise low-hanging fruit first to get some easy and enjoyable wins under your belt.