This is a replication study by Yagmur Deniz Kısa and colleagues that looks at an insight that has been held for decades: people will speak more fluently when they use gestures. Or more concrete:

According to an influential version of this proposal, the Lexical Retrieval Hypothesis (LRH), gestures facilitate speech production by helping speakers find the right words; however, gestures are only posited to affect words of a particular kind. Krauss and colleagues (1996; 1998) noted that people gesture far more frequently during phrases with spatial content than during phrases without it, and hypothesized that gesturing helps speakers find the right spatial words.

This means:

If gesturing helps speakers find the right words, then preventing speakers from gesturing should make their speech production more disfluent, compared to when they are free to gesture.

And that is exactly what was checked in the replication. And no…

Does gesturing facilitate speech production by helping people find the right spatial words? We found no evidence that speakers speak less fluently when they are prevented from gesturing, compared to when they are allowed to gesture freely. We failed to find an effect of preventing gesture during speech with metaphorical spatial content, a finding that would have expanded the scope of LRH to encompass speech about abstract concepts. More fundamentally, we also found no significant effect of preventing gesture during speech with literal spatial content. We thus failed to find support for Rauscher and colleagues’ (1996) influential claim that preventing gesture increases disfluencies for spatial language. More broadly, our data provide no support for the idea that gesturing facilitates speech production by helping people find the right words.

While the study was with a small group of participants, the authors of this paper did something more. They re-evaluated the original study:

Rauscher and colleagues’ (1996) study has been widely cited as evidence for the idea that gesturing helps speakers find the right words and, more broadly, as some of the first evidence that gesturing serves a cognitive function for speakers (e.g., Krauss, 1998). However, a careful reexamination of Rauscher et al. (1996)’s disfluency rate results shows that they did not find statistically significant results supporting the idea that gesturing helps speakers find the right spatial words.

Abstract of the replication study (free version here):

Why do people gesture when they speak? According to one influential proposal, the Lexical Retrieval Hypothesis (LRH), gestures facilitate speech production by helping people find the right spatial words. Do gestures also help speakers find the right words when they talk about abstract concepts that are spatialized metaphorically? If so, gesture prevention should increase disfluencies during speech about both literal and metaphorical space. We sought to conceptually replicate the finding that gesture prevention increases disfluencies in speech about literal space, which has been interpreted as evidence for the LRH, and to extend this pattern to speech about metaphorical space. Our large dataset provided no evidence that gestures facilitate speech production, even for speech about literal space. Upon reexamining past research, we conclude that there is, in fact, no reliable evidence that preventing gestures makes speech more disfluent. These findings challenge long-held beliefs about why people gesture when they speak.