1  Evidence, empirical considerations, organization

1.1 Structure of each ‘barriers’ section

In each ‘barriers’ section we consider…

  1. A category of barriers; this is a first-level heading, e.g. ‘Quantitative biases’ is section @ref(quant-biases).

Here we discuss this category in general, explaining its meaning and plausibility in terms of the underlying theories, models, and generalized observations.

  1. Each specific barrier in this category gets a second-level heading, e.g., “Proportional dominance effect” is section @ref(prop-dom).

  2. For each specific barrier we first present (third-level headings):

  • Definition of the barrier
  • Conceptual explanation of the barrier, relating this to theory and generalized observation
  • Relevance to effective giving (sometimes recapping what was stated above)

4. Next we consider the Evidence, with a third-level header preamble discussing the nature of the evidence.

  1. Within this, we consider the evidence for each specific empirical claim supporting the case for this barrier’s existence and importance.

The linked doc outlines how closely the presented structure is currently followed in the document

We hope to:

  • ‘operationalize’ each claim with a falsifiable set of predictions
    • justifying the connection to each prediction,
  • and carefully asses the strength of the evidence for each of these claims.

See, e.g., our work-in-progress on ‘CBA avoidance’ here.

This distinguishes the current project from broader reviews and literature surveys.

In considering the evidence for these claims, we generally consider one paper or project at a time (or a specific piece of independent evidence from this paper).1

1.2 Consideration of each paper

For each paper we focus on key elements, including:

  • Method (lab experiment involving real charitable asks, natural field experiment with charity partner, observational analysis asserting causality via a regression discontinuity approach, etc.)

  • The context and data (the population of donors or lab participants, the charity(s) involved, etc.)

  • The nature of the ‘treatment’ or independent variable (e.g., realistic information presented about impact per dollar)

  • The strength and credibility of the empirical results (considering, e.g., limitations and confounds in the design, experimenter-demand effects, estimates’ statistical strength and power, appropriateness of the statistical analysis, preregistration, signs of specification-fishing and multiple-testing without corrections, replications and citations, etc.)

  • Link where to find the paper itself as well as the materials and the data collected.

1.3 Evidence and meta-analysis: discussion

We consider2

  1. The nature of evidence for drivers of effective/ineffective giving, and how we might evaluate it

  2. Meta analysis: Existing work and plans for doing meta-analysis as part of this project

See innovationsinfundraising for some earlier discussion of this, as well as “What works to increase charitable donations? A meta-review with meta-meta-analysis” by Noetel et al

1.4 Airtable organisation; contributing content

These classifications of barriers are organised, outlined, and defined in our Airtable; see the view below.

The barriers themselves are organised as well, see the view below:

To suggest a new barrier, or a modification of an existing proposed barrier, please fill out the form below and linked here (or contact us otherwise, including via a hypothes.is comment):

You can also engage in this project directly, helping us build this content in R/markdown or in a Google docs integration. See the ‘tech for creating…’ appendix.

Did we leave out a paper or a source of evidence? Please fill out the form below or click the link HERE (or contact us otherwise, including via a hypothes.is comment):

All contributions will be credited and engaged with.

1.5 Other practical considerations, partnerships

This project is informally connected with the EA Market Testing team. However, this project is several years older, and the scope is not entirely overlapping. The present project focuses on giving, while EAMT has a broader agenda connected with the EA movement. The EAMT is mainly a practically-driven project, while the current synthesis also considers braoder and more theoretical issues.

Consider potential integration with…

Consider potential integration with (no-longer-maintained) PriorityWiki/Rethink Charity {-}

How have EA orgs been brought together …especially at universities?

  1. However, we are still considering the best narrative approach to presenting evidence from multiple papers/projects/analyses.↩︎

  2. Further discussion needed here↩︎