🔭 astrophysics

An update to Allocating Time on Scientific Platforms in Outer Space using five cycles of JWST General Observer programs

This research note updates a 2025 study on James Webb Space Telescope time allocation by analyzing five cycles of General Observer programs, confirming earlier findings on team size and data sharing while revealing emerging biases favoring US-based Principal Investigators and those with prior JWST experience.

Original authors: Christopher Williams

Published 2026-06-04

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Christopher Williams

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the James Webb Space Telescope (JWST) as the most exclusive, high-performance sports car in the universe. It's incredibly expensive, runs on limited fuel (time), and thousands of scientists are lining up to borrow it for a spin. Because there isn't enough time for everyone, a committee has to decide who gets to drive it and for how long.

This paper is an update to a previous study that checked if the committee was playing favorites. The original study looked at the first three "seasons" (Cycles) of borrowing. This new study adds the fourth and fifth seasons to see if the rules of the game changed as the competition got fiercer.

Here is what the study found, explained simply:

1. The "Team Size" and "Sharing" Rule

The biggest factors in getting a long drive with the telescope aren't who you know or where you are from; they are how big your team is and whether you promise to share the photos you take.

The Analogy: Think of it like a group project. If you bring a large team of helpers and promise to let everyone else use the final report (not keeping it secret), you are much more likely to get the car keys. The study confirms that bigger teams and "open sharing" policies are the strongest predictors of success.

2. The Gender Question: Did the Pendulum Swing Back?

In the first study, there was a worry that while things looked fair at the start, the committee might slowly start favoring male leaders (Principal Investigators) again as time went on, like a pendulum swinging back to old habits.

The Update: The new data says no. The pendulum hasn't swung back. Overall, female leaders are getting just as much time as male leaders in the new seasons.
The Catch: However, the study found a small "shadow" in specific areas. When looking only at the most difficult, deep-space missions (looking at galaxies far, far away), there is a slight trend where male leaders are getting a bit more time. It's not a total bias, but it's a subtle difference that needs watching.

3. The "American" Factor and Politics

The study wondered if a change in the US government (specifically a new administration focused on national priorities) would make the committee favor American scientists over international ones.

The Update: The results are mixed. The committee didn't suddenly start picking Americans just because of politics. However, American scientists are getting slightly more time in the new seasons compared to before. The study suggests this might be a slow trend worth watching, but it's not a sudden, dramatic shift caused by political pressure.

4. The "Veteran" Advantage

A brand new finding in this update is the power of experience.

The Analogy: Imagine a video game. If you've already beaten the level before, you know the controls, the glitches, and the best strategies.
The Finding: Scientists who have successfully used the telescope before are now getting a significant advantage. They are winning more time in future rounds. The study suggests that once you know how to navigate the system and write a winning proposal for this specific telescope, it becomes easier to win again. It's a "rich get richer" effect based on past success.

5. The "Fine Print" Check

Finally, the researchers checked if the way the telescope's time was calculated changed (like if the car's engine became less efficient, requiring more time for the same distance).

The Result: They checked the "completed" projects and found that these technical adjustments didn't change the main results. The rules of the game regarding who wins remain consistent, regardless of minor operational tweaks.

The Bottom Line

The committee running the James Webb Space Telescope seems to be doing a good job of keeping the playing field level. The biggest winners are those with big teams who share their data. While there are tiny cracks where old biases (like favoring men in deep-space projects) or new trends (favoring experienced players and slightly more Americans) are appearing, the overall system remains fairer than it has been in the past. The "pendulum" of bias hasn't swung back yet.

Technical Summary: An Update to 'Allocating time on scientific platforms in outer space' using five cycles of JWST General Observer programs

Problem Statement
This research note updates a previous study (Williams, 2025) regarding the allocation of observing time on the James Webb Space Telescope (JWST). The original study, based on the first three General Observer (GO) cycles (2022–2024), utilized multi-variable regression to analyze potential biases related to Principal Investigator (PI) gender, institution, and country. While the initial analysis suggested progress toward reducing bias, it left three critical issues unresolved:

Gender Bias Trends: Whether the initial support for female PIs in Cycle 1 would revert to long-standing male-bias norms as competition intensified in Cycles 2 and 3.
Political Influence: Whether the transition to a second Trump administration (affecting Diversity, Equity, and Inclusion policies and federal funding priorities) would introduce bias favoring US-based PIs.
Policy Robustness: Whether changes in telescope allocation committee (TAC) policies and overhead adjustments between cycles would alter the regression findings.

Additionally, the study sought to verify if operational adjustments to "Completed" program statuses materially affected the interpretation of time allocation data.

Methodology
The study employs the same methodological framework as Williams (2025), utilizing bi-variate correlations and multi-variable regression models with the natural log of allocated time as the dependent variable.

Data Expansion: The sample size was expanded from $n=723$ (Cycles 1–3) to $n=1,208$ programs, incorporating data from Cycles 4 and 5.
Variable Updates:
- Scientific Categories: A revised mapping was created to align new Cycle 4–5 scientific category labels with the existing "Distance cat" control variable (Solar System, Galactic, Extra-galactic).
- Status Flags: The "Completed" status variable was updated as of May 21, 2026 (compared to January 1, 2025 in the prior study), increasing the number of completed programs from 502 to 687 to test for robustness against overhead changes.
- New Determinants: The analysis explicitly tests for "JWST experience" (prior successful proposals) and "USA PI" status.
Analytical Approach: The study utilizes full sample models, partitioned models by Cycle, and partitioned models by scientific distance category and PI gender. Margins plots are used to visualize interactions between gender, country, and cycles.

Key Results
The expanded analysis confirms the robustness of the original findings while revealing new determinants of time allocation:

Gender Bias:
- Overall: There is no evidence of a systemic swing back to male-bias in the full dataset across Cycles 4 and 5. The coefficient for PI gender remains insignificant in the full model.
- Partitions: However, significant differences persist in specific partitions. In extra-galactic programs (Distance cat = 3), there is a statistically significant positive coefficient for male PIs ( $p<0.05$ ), suggesting a growing bias in this specific high-stakes category. Conversely, in smaller time-allocation partitions, female PIs show a positive trend, though significance is weaker than in the full model.
USA PI and Political Context:
- The "USA PI" variable, which was insignificant in the Cycles 1–3 full model, becomes positive and statistically significant ( $p<0.05$ ) in the Cycles 1–5 full model and specifically in Cycle 5.
- The proportion of US-led programs increased slightly from 57.6% (Cycles 1–3) to 60.2% (Cycles 4–5). While the authors note this is not a sudden shift, the emergence of a significant coefficient suggests a potential trend toward national prioritization that warrants monitoring.
JWST Experience:
- A major new finding is the significance of "JWST experience." In the Cycles 1–3 model, this variable was insignificant ( $p=0.11$ ); in the Cycles 1–5 model, it is highly significant ( $p<0.01$ ).
- This variable is a strong positive predictor for time allocation, particularly for extra-galactic programs and male PIs. This suggests that prior success on the platform confers a competitive advantage, potentially through improved proposal quality or risk appetite.
Robustness of Operational Data:
- Re-running the Cycles 1–3 models with the updated "Completed" status flag yielded nearly identical coefficients and significance levels ( $R^2$ moved from 0.24 to 0.25). This confirms that operational overhead adjustments do not materially alter the interpretation of allocation bias.
Consistent Predictors:
- Team Size and Non-exclusive data sharing remain the two strongest positive predictors of allocated time across all models and cycles.

Significance and Claims
The paper claims to provide tentative answers to the unresolved issues regarding the stability of bias reduction on mega-science platforms.

On Gender: The study refutes the hypothesis that gender bias inevitably returns as platforms mature, attributing the overall stability to sincere institutional interventions. However, it highlights that bias may persist or evolve in specific high-competition niches (extra-galactic astronomy).
On Politics: The analysis suggests that the broader political narrative (specifically regarding DEI and national priorities) has not drastically disrupted the allocation process, though the emerging significance of the "USA PI" variable indicates a subtle shift that requires future tracking.
On Experience: The study contributes to the literature on scientific productivity by identifying "platform-specific experience" as a critical determinant of success, suggesting that the "learning of the reward system" is a tangible factor in competitive time allocation.
On Policy: The findings support the robustness of using accepted program data even before completion, as operational adjustments do not skew bias detection.

The authors conclude that while the institutional framework for JWST has successfully mitigated broad gender bias, the competition is increasingly driven by team size, data openness, and prior platform experience, with subtle emerging indicators of national bias in the most competitive scientific categories.