RESEARCH_FILE
The Science of Imposter Phenomenon in Software Engineering
'Everyone here is smarter than me and it's only a matter of time before they notice.'
SEE THE PRACTICE
Turn a thought this research explains into one clear move.
THE THOUGHT
“I spent a decade learning what just got automated”
YOUR RECORDED RESPONSE
“The decade taught me what to trust and what to throw out. The tool only speeds up the throwing-out part.”
ONE PRIVATE MOVE
Open a blank note and write down one real decision from the last year where your experience changed the outcome, not the typing. Read it back once, then close the note.
That thought — or something close to it — is so common among software engineers that a 2024 study presented at ICSE found measurable imposter phenomenon in a majority of surveyed practitioners. It isn't a quirk of personality. Psychological research dating to 1978 has documented it as a recognizable pattern driven by competence-assessment errors, code-review anxiety, and team environments that either magnify or dampen it. The old script — 'I just need more confidence' — misses the structural half of the story.
How the science changed
- 1978
Pauline Clance and Suzanne Imes publish the first formal description of 'impostor phenomenon' — a pattern of high-achieving individuals who, despite objective evidence of success, believe they are frauds and fear being exposed. Their clinical observations focused on high-achieving women but noted the pattern appeared across gender. ↗
- 1999
Kruger and Dunning publish their landmark study showing that low performers systematically overestimate their competence (the Dunning-Kruger effect), while high performers often underestimate theirs — a contrasting asymmetry that helps explain why the most skilled engineers are frequently the most likely to feel like impostors. ↗
- 2016
Google publishes Project Aristotle findings from a study of 180 teams: psychological safety — the belief that one can take interpersonal risks without punishment — was the single strongest predictor of team effectiveness, outweighing talent composition, seniority mix, and co-location. Teams where engineers feared judgment performed worse across every metric. ↗
- 2017
Santos et al. publish a meta-analysis of imposter phenomenon across populations, finding prevalence in approximately 70% of the general population at some point in their lives and noting that the experience was particularly pronounced in high-stakes evaluation contexts — directly applicable to code review, on-call incidents, and senior-engineer promotions. ↗
- 2021
Byskov and colleagues examine imposter phenomenon specifically in STEM and technical fields, finding that the combination of objective performance feedback (test failures, bug counts, benchmark scores) and social comparison within high-performing peer groups creates a compounding effect that elevates imposter feelings beyond what individual self-doubt alone would produce. ↗
- 2024
Guenes et al. present at ICSE-SEIS 2024 the first large-scale empirical study of imposter phenomenon in software engineering: measurable imposter scores in a majority of surveyed engineers, with seniority level, team size, and hiring source (self-taught vs. CS degree) all predicting variance — reframing imposter phenomenon from a personal quirk to a structurally explainable occupational pattern. ↗
- 2024
Lee et al. publish in Empirical Software Engineering a quantitative study of code review anxiety: they find it is measurable, distinct from general social anxiety, significantly associated with imposter feelings, and partially mitigable through structured review practices and explicit norms around the purpose of code review. ↗
What people believe vs. what the data shows
The belief“Feeling like an imposter means you actually lack the skills for your role.”
The dataKruger and Dunning's 1999 research found the opposite asymmetry: those with the least skill most consistently overestimate their competence, while those with the highest skill systematically underestimate theirs. Imposter phenomenon is a poor proxy for actual skill level. ↗
The belief“Imposter phenomenon in tech is mainly a problem for junior engineers and goes away with experience.”
The dataThe ICSE 2024 study by Guenes et al. found imposter scores varied with seniority, but senior engineers were not exempt. The research showed that as responsibilities grow — leading teams, making architectural calls, public speaking — new imposter triggers can emerge even for experienced practitioners. ↗
The belief“Code review anxiety is just normal nervousness and doesn't have measurable effects on engineering work.”
The dataLee et al. (2024) found code review anxiety is a measurable, distinct construct associated with avoidance behavior — delaying submitting code, under-commenting, and not requesting reviews from certain reviewers. It predicted real workflow changes, not just discomfort, and was significantly correlated with imposter phenomenon scores. ↗
The belief“If your team is talented enough, psychological safety doesn't matter — results speak for themselves.”
The dataGoogle's Project Aristotle analysis of 180 teams found psychological safety outranked every other factor — including individual talent, seniority mix, and team structure — as a predictor of team effectiveness. High-talent teams with low psychological safety consistently underperformed lower-talent teams with high psychological safety. ↗
The belief“Imposter phenomenon only affects self-taught developers or those without computer science degrees.”
The dataGuenes et al.'s 2024 ICSE study measured imposter scores across both CS-degree holders and self-taught engineers and found it was prevalent in both groups. While the triggers differed — credentials anxiety vs. vocabulary anxiety — imposter phenomenon was not specific to any one educational background. ↗
TEST_YOURSELF · How well do you know this science?
01 Pauline Clance and Suzanne Imes coined 'impostor phenomenon' in 1978 to describe what?
Clance and Imes described a pattern in high-achieving individuals who, despite objective evidence of competence, internally attribute their success to luck or deception rather than ability — and persistently fear being exposed as frauds. source ↗
02 What does the Dunning-Kruger effect tell us about the relationship between skill level and self-assessment of competence?
Kruger and Dunning found a systematic asymmetry: low performers overestimate their competence (they lack the metacognitive skills to recognize their own gaps), while high performers underestimate theirs — making imposter-like self-doubt more common, not less, at higher skill levels. source ↗
03 What did Google's Project Aristotle research identify as the single strongest predictor of engineering team effectiveness?
Project Aristotle analyzed 180 Google teams and found psychological safety — whether team members felt safe to take risks and be vulnerable in front of each other — outperformed every other variable, including talent level, as a predictor of team output and effectiveness. source ↗
04 What did Lee et al.'s 2024 study find about code review anxiety in software engineers?
Lee et al. found code review anxiety is distinct from general social anxiety, measurable through survey instruments, predicts avoidance behaviors like delaying code submission, and correlates significantly with imposter phenomenon scores — making it a specific occupational stressor, not just general nervousness. source ↗
05 What did the ICSE-SEIS 2024 study by Guenes et al. establish about imposter phenomenon in software engineering?
Guenes et al.'s 2024 study at ICSE-SEIS was the first large-scale empirical measurement of imposter phenomenon in software engineers specifically, finding it present across seniority, team size, and educational background — structurally rooting what had often been treated as individual self-doubt. source ↗