Episode 12.7: Benefits, Challenges, and Getting Started with Open Science

Last updated on 2026-10-04 | Edit this page

Overview

Questions

  • What does open science actually gain a researcher, beyond goodwill?
  • What are the real, practical obstacles to working openly?
  • If a team wants to start, where should they actually begin?

Objectives

Learners will be able to:

  • Identify at least two strengths and two limitations of open science as a whole.
  • Explain the most common practical barriers teams face when adopting open practices.
  • Match a research team’s situation to a realistic, small first step toward openness.

A Quick Recap

Across the last five episodes, we’ve covered eight practices, each opening up a different part of research:

Practice Opens Up
Preregistration & Registered Reports The hypothesis and analysis plan
Open Data (FAIR & CARE) The raw dataset
Open Materials & Code Instruments and analysis steps
Open Access & Preprints The written output
Open Educational Resources Teaching and training materials
Open Evaluation The peer review process
Team Science Who participates, and how credit is shared
Open Source Software The tools used to do the research

None of these require adopting all the others. Let’s look honestly at what each category of practice actually gains you, and what it costs.

Strengths of Open Science

  • Stronger, more trustworthy findings: Preregistration and open data make it far harder for a result to look stronger than it really is, because the analysis plan and the raw numbers are both on record.
  • Faster progress: Preprints and open data let other researchers build on new findings within a shorter timeframe. No waiting for a full review cycle before anyone else can act on the work.
  • Wider access to knowledge: Open access and OER mean people outside well-funded institutions (students, practitioners, researchers in lower-resource settings) can actually read and use the work.
  • Errors get caught sooner: Open code and open source tools mean mistakes in an analysis, or in the software behind it, can be found and fixed by anyone, not just the original author.
  • Broader participation: Team science and open collaboration create a real way in for contributors who aren’t already part of an established network.

Limitations of Open Science

  • Time and effort: Documenting a protocol, cleaning a dataset for public release, and writing a clear codebook all take real time, time that competes with everything else a research team is doing.
  • Not everything can be fully open: Data involving vulnerable populations, sensitive health information, or proprietary partnerships may need to stay restricted, in part or entirely. This is exactly where CARE-style governance questions matter.
  • Uneven incentives. Many institutions still evaluate researchers mainly on publication count and journal prestige, which doesn’t always reward the extra work openness requires.
  • Cost can shift, not disappear: Some open-access journals charge authors a publishing fee, which can be a real barrier for teams with limited funding. Openness solves a reader-access problem but can create a different on the author-side.
  • Openness isn’t the same as quality: A study can be fully preregistered, fully open, and still be poorly designed. Open practices make it easier to see whether research is sound; they don’t automatically make it sound.

Real-World Open Science Applications in Practice?

  • Preregistration & Registered Reports: clinical trials, psychology experiments, and policy evaluations, where the risk of chasing a “publishable” result is highest.
  • Open Data (FAIR & CARE): public health surveillance, genomics, and any study built on community-level data, where reuse and community governance both matter.
  • Open Access & Preprints: fast-moving fields (like public health during an outbreak) where getting findings out quickly can matter as much as getting them peer-reviewed.
  • Open Educational Resources: under-resourced schools and universities building courses without the budget for commercial textbooks.
  • Open Evaluation: fields grappling with reproducibility concerns, where visible review builds back reader trust.
  • Team Science: large-scale, multi-country studies where no single team has the reach to answer the question alone.
  • Open Source Software: any field that depends on shared analysis tools, where a single tool error could otherwise ripple silently through thousands of studies.

Cross-Field Differences

Different fields lean on different open practices, largely based on what they can most easily control: | Field | Practices Most Commonly Used | Why | |——–|——————————|————————-| | Public Health | Open Data, Open Access | Findings need to reach practitioners and policymakers fast. | | Psychology & Social Science | Preregistration, Registered Reports | The field has a well-documented history of the “found it, then predicted it” problem. | | Software & Computing | Open Source Software, Open Materials | The tools are the research output. | | Education | Open Educational Resources | The materials themselves are what gets reused. | | Genomics & Large-Scale Biology | Open Data, Team Science | No single lab can generate enough data alone. |

Getting Started: Where a Team Should Actually Begin

Trying to adopt all eight practices on a first attempt is a good way to adopt none of them well. A more realistic path:

  • Pick one practice that fits your next project, not your whole research programme.
  • Preregistering a single upcoming study is a smaller, more concrete commitment than trying to overhaul how your whole team works.
  • Start documentation early, not at the end. A codebook written while you’re still collecting data takes an afternoon. Reconstructed a year later from memory, it can take a week, if it’s possible at all.
  • Check what your institution or funder already expects. Many funders now require a data-sharing plan. If yours does, that requirement is a ready-made starting point, not an extra task.
  • Use existing infrastructure rather than building your own. Established repositories and registries already handle the technical side of sharing data, materials, and preregistrations; there’s rarely a need to build something from scratch.
  • Treat the first attempt as a learning project. A first preregistration or first shared dataset doesn’t need to be perfect. It needs to exist, and to get a little more thorough the next time.

Test Your Knowledge!

Challenge

Challenge 1:

Which of the following is a genuine limitation of open science, rather than just a myth about it?

A. Open practices make research take zero extra time.
B. Documenting protocols and preparing data for sharing takes real time and effort.
C. Open research is always lower quality than closed research.
D. Open access journals are always free for both readers and authors.

B. The time and effort required is a genuine, practical limitation.
A, C, and D are all false statements about open science.

Challenge

Challenge 2:

A small research team wants to start working more openly but has no dedicated budget or extra staff time. What’s the most realistic first step, based on this episode?

A. Restructure the entire research programme around all eight open practices immediately.
B. Wait until funding is available to do anything at all.
C. Pick one practice (like preregistering their next study) and start documentation early rather than at the end.
D. Only adopt open source software, since it’s the only free option.

C. Starting small, with one practice applied to one upcoming project, is realistic and sustainable. Trying to do everything at once, or waiting indefinitely, are both less workable in practice.

Discussion

Challenge 3:

True or False: A study that is fully preregistered and fully open is automatically well-designed.