Why overheard conversation is harder to ignore than other office noise
Speech is not just noise. It is meaningful sound that the brain processes automatically, so even when you are trying to work, part of your mind keeps tracking the words. Researchers call this the irrelevant speech effect: background speech that has nothing to do with your task still pulls on mental resources and increases distraction, which in turn lowers cognitive performance [1]. In open-plan offices, irrelevant speech has consistently been reported as the most distracting sound, more so than general background noise [3].
The practical consequence is that the problem is not simply volume. A loud but steady hum can fade into the background, while a quiet conversation nearby can keep intruding because your brain keeps trying to follow it. That is why open-plan offices can feel distracting even when measured sound levels seem modest.
Which mental tasks suffer most, and by how much
The tasks that suffer most are those that require holding verbal information in mind while working on it. A common laboratory test is serial recall, where a person must remember a sequence of items in the correct order. Speech noise reduces performance on this kind of task, and the reduction depends on how intelligible the speech is [5].
Intelligibility is usually measured with the Speech Transmission Index, or STI, a number that describes how clearly speech comes through in a given setting. In a study of Chinese-language offices, serial recall performance dropped within an STI range of roughly 0.31 to 0.47 [5]. That range is narrower than what has been reported for non-Chinese language environments, and the average rate of change in performance was not smaller than in non-Chinese settings, even though the overall effect of speech noise on serial recall was lower in the Chinese environment [5]. In plain terms, the window where speech goes from harmless to harmful may be tighter for Chinese listeners, but within that window the drop can be just as steep.
Does it matter how many people are talking, and can masking help?
Real offices rarely have a single talker. Studies simulate open-plan settings with varying numbers of talkers and different levels of sound masking to see how performance and subjective distraction change [3]. The multi-talker character of the sound matters: these simulations are designed specifically to capture how a realistic office soundscape, rather than a single voice, affects cognitive performance [1][3].
Sound masking, which adds a controlled background sound to cover up speech, is one of the main tools for reducing the problem. Where the masking loudspeakers are placed appears not to change how well masking protects cognitive performance: local sound masking was found to be as efficient as conventional sound masking [2]. That is a useful practical result, because it means offices may have flexibility in where they install masking systems without losing the cognitive benefit.
Researchers have also built models to predict how much cognitive performance will drop under distracting background speech, and these models can account for multi-talker sound environments [4]. Such predictions are useful for designing offices, but they are models of expected performance, not a guarantee for any individual worker or task.
What this means if you work in an open-plan office
The evidence points to a simple rule of thumb: the more understandable the nearby conversation, and the more your task depends on holding words or numbers in mind, the more you should expect to be slowed down. Reducing intelligibility, whether through masking, distance, or layout, is the lever these studies support [2][5]. The effect is not imaginary or a matter of willpower; it reflects how the brain handles meaningful sound [1].
One caveat sits beside these conclusions. Much of this work uses simulated office environments and laboratory tasks such as serial recall, so the measured performance drops are real but occur under controlled conditions rather than in a live office over a full workday [3][5]. The direction of the effect is consistent across these studies, but the exact size of the hit to your own work will depend on the task and the acoustics around you.
Sources used in this answer
The irrelevant speech effect in multi-talker environments: Applications to open-plan offices
The irrelevant speech effect causes cognitive decline and increased acoustic distraction in multi-talker open-plan office simulations.
Effects of the location of sound masking loudspeakers on cognitive performance in open-plan offices: Local sound masking is as efficient as conventional sound …
Local sound masking was as efficient as conventional sound masking for protecting cognitive performance in open-plan offices.
Auditory distraction in open-plan office environments: The effect of multi-talker acoustics
In simulated open-plan offices with varying talker numbers and masking levels, multi-talker speech produced performance decrements and subjective distraction, with irrelevant speech reported as the most distracting sound.
A simple model to predict the cognitive performance in distracting background speech
A model was developed to predict cognitive performance under distracting background speech in open-plan offices, including multi-talker sound environments.
A laboratory study correlating serial recall performance and speech intelligibility of Chinese language in open-plan offices
In Chinese-language open-plan office conditions, serial recall performance decreased within an STI range of 0.31–0.47, a narrower range than in non-Chinese environments, with an average change rate not lower than in non-Chinese settings.