Action observation (AO), the act of learning by watching others, is a well-documented method for skill acquisition. A recent study investigated the impact of action observation, and action observation combined with motor imagery (MI), on an individual’s confidence in their skill ability. The research specifically aimed to understand whether incorporating motor imagery could moderate the often-observed overconfidence that arises from action observation alone. This exploration delves into the Learning Consequence Of An Action By Watching Others, particularly how it affects self-perception of competence.
Consistent with previous research, the study confirmed that repeated action observation of a skill leads to increased confidence in one’s ability to perform that skill. Interestingly, while there were no significant overall group differences, the interactions between the groups and the trials suggested that confidence levels were more tempered in participants who combined action observation with motor imagery. This group appeared to have a more realistic assessment of their abilities, aligning more closely with their actual skill level at the study’s conclusion compared to those who only engaged in action observation. Despite this, both groups exhibited overconfidence when their perceived ability after observation was compared to their actual demonstrated skill.
These findings contribute to a growing body of evidence suggesting that action observation and motor imagery are distinct cognitive processes. Participants in the study seemed capable of integrating the simulated sensory experience from motor imagery with the visual information gained from action observation. This integration influenced their self-evaluation of skill. While both AO and MI are covert forms of action, AO is generally considered more passive. This passivity might lead to a more abstract understanding of the skill, less connected to the sensory experiences of physically performing it (Vogt et al., 2013). The subtle differences in confidence ratings across trials between the groups imply that action simulation isn’t an automatic component of action observation, or at least not consistently so. It’s also plausible that any spontaneous motor imagery occurring during action observation is less potent than deliberate, focused motor imagery.
Further supporting this idea is the increasing evidence that motor imagery facilitates an internal simulation of the sensory outcomes of a movement. This simulation is thought to occur through a predictive mechanism, often referred to as a forward model (Dahm & Rieger, 2019; Kilteni et al., 2018). For instance, research using dart throwing tasks showed that individuals using motor imagery could predict outcomes with similar accuracy to those physically performing the task, even without visual feedback (Dahm & Rieger, 2019). This ability to predict movement success through motor imagery suggests that it provides access to diagnostic information about one’s capabilities.
One significant implication of motor imagery, linked to its ability to simulate sensory outcomes, is that it may also enhance the capacity to accurately imagine failure. This imagining of failure is less likely to occur with action observation alone, especially when individuals are observing a proficient model (Lirgg & Feltz, 1991). The finding that individuals can predict errors during motor imagery (Dahm & Rieger, 2019) reinforces the idea that MI offers a realistic sense of one’s abilities, including potential imperfections in execution. While both AO and MI may lead to overconfidence, particularly with repeated exposure, the combination of MI with AO appears to mitigate this overconfidence. The current study’s data supported this expectation, although perhaps not as robustly as previous literature might suggest. Future studies should include physical practice conditions to compare motor imagery with actual performance, determining if motor imagery still leads to overestimation of ability, as hinted at by exit-rating data and prior research (Dahm & Rieger, 2019). Such comparisons would also help position AO and MI on the action continuum relative to physical execution (Vogt et al., 2013).
Research on how movement representations develop after observational practice also indicates that action observation, without physical practice, results in relatively abstract mental representations of actions (Kim et al., 2010, 2017). AO provides information about the organization and visual aspects of movement, making it easier to use in the early stages of learning compared to motor imagery (Kim et al., 2017). The visual information from AO may be stored in memory as learners understand and learn movement organization and perceptual elements. However, it might bypass the need for kinesthetic sensory input in facilitating learning (Frank et al., 2016).
Considering the data on overconfidence following AO and its reduction with MI, it’s relevant to note other studies where perceived fluency increased after video observation. Kardas and O’Brien (2018) and Jordan et al. (2022) have demonstrated overconfidence from brief observation of diverse skills, from magic tricks to landing planes. Jordan et al. propose that video observation boosts confidence by helping individuals create more detailed mental images (aligning with source monitoring error theories; Lindsay, 2014). However, the current study suggests a need to refine this mechanism. It’s not merely about enhanced imagination from video observation, but rather that these initial imaginations may not effectively engage the simulation processes necessary to accurately assess one’s actual capabilities. Explicitly instructing individuals to use motor imagery seems to provide more accurate self-perceptions of ability compared to spontaneous, unprompted imagery during observation.
These findings raise crucial questions for future research. How these effects translate into skill acquisition and how they can be used to enhance learning remains unclear, especially regarding the use of demonstrations in teaching. The two-ball juggling task was chosen based on prior work showing confidence increases with repeated observation, moderated by physical practice (Hodges & Coppola, 2015). Using a task that could be mastered allowed for ability assessment, relating high confidence ratings (~70%) to lower self-reported juggling ability (~50%). More complex or unattainable tasks, like flying a plane (Jordan et al., 2022), might show a greater reduction in confidence when combining MI with AO compared to AO alone. As tasks become more challenging or novel, imagining the sensory experience may become harder, potentially leading to less confidence from simply watching.
The study did not include groups with repeated physical practice or ‘pure’ motor imagery (without AO), limiting inferences about overestimation of ability in MI alone. However, comparing exit ability and confidence ratings after AO + MI in this study, and prior research showing ability overestimation in MI-only conditions (Dahm & Rieger, 2019), suggests that MI, like AO, may also lead to overconfidence. Methodologically, future research should control for potential confidence inflation simply from repeated questioning. While there’s no reason to expect different responses to repeated confidence questions in AO+MI versus AO groups, controlling question frequency is needed to accurately assess how observation or MI alters confidence. Although a keypress task controlled trial duration in the OBS group, interval duration differences between OBS and OBS+MI groups might have contributed to confidence deflation in the latter. While the impact of time alone on confidence is unclear, better timing control is needed in future research.
Considering learning outcomes and potential improvements in delayed retention tests when combining AO and MI, as shown in previous reviews (Eaves et al., 2016, 2022), is important. Compelling evidence suggests that combining MI with AO has additive benefits over either method alone (Eaves et al., 2016; Wright et al., 2014), though long-term learning impacts are rarely studied. In this study, dissociations between confidence measures and actual ability at a later retention date might emerge if MI enriches the learning experience, enabling better skill acquisition and later performance, while moderating short-term confidence.
Manipulating video perspective (first-person vs. third-person) showed no benefit or detriment, and first-person perspective did not differentially affect AO+MI versus AO as predicted. This lack of difference suggests spontaneous MI may not occur during AO, even with a first-person perspective intended to encourage it (Riach et al., 2018). Evidence suggests greater motor system involvement with first-person videos (Alaerts et al., 2009), possibly through spontaneous MI. However, this study found no supporting evidence, despite MI instructions encouraging first-person imagery focusing on movement feel and timing. The absence of measures for spontaneous MI during AO means conclusions about its absence are tentative.
Video perspective was a repeated measure factor, potentially diluting perspective effects when combined over practice halves. The study may have lacked power to compare groups with different perspectives in the first half of practice. Exploratory analysis with video-order as a between-factor showed similar Group X Trial interactions, but also a video-order effect. Confidence was lower when first-person videos were shown first. This order effect might indicate that first-person viewing reduces confidence inflation from repeated observation, especially when not preceded by a third-person view. A first-person view may encourage spontaneous MI or more accurate ability perceptions. Future research should isolate perspective in a between-subjects design with adequate power.
