The Simon Effect Asymmetry for Left- and Right-Dominant Persons

Robert W. Proctor; Qi Zhong; Jing Chen

doi:10.5334/joc.265

The Simon effect, named after J. R. Simon (1990), has been widely used as a tool to investigate various issues in response activation and selection due to the correspondence of variation on a nominally irrelevant stimulus dimension (usually spatial location) with left and right response alternatives (see Hommel, 2011; Lu & Proctor, 1995; Proctor, 2011, for reviews). Participants are typically instructed to make a left or right keypress response to a stimulus feature such as red or green color; the stimulus occurs in a left or right location, which is not relevant to the task. Yet reaction times (RTs) are shorter and response accuracy higher when the stimulus and response locations correspond than when they do not. This benefit of spatial correspondence is often described as “automatic” since it is not an intention as specified in the instructions (e.g., Luo & Proctor, 2022). The most likely reason why the correspondence between stimulus and response locations affects performance is that the left/right discrimination has been defined as relevant to the responses (Ansorge & Wühr, 2004; Yamaguchi & Proctor, 2012).

In the first decade of the 21^st century, several studies established that a Simon effect asymmetry is evident when left-hand responses are analyzed separately from right-hand responses (Rubichi & Nicoletti, 2006; Spironelli et al., 2009; Tagliabue et al., 2007). For right-handed persons, the Simon effect is larger for the right stimulus location (which corresponds with the right response) than for the left stimulus location (which does not correspond). This relation is reversed for left-handed persons, for which the Simon effect is larger for the left than right stimulus location (Rubichi & Nicoletti, 2006; Tagliabue et al. 2007). These and other findings led the authors of the three studies to propose that the Simon effect asymmetry is due to an attentional bias that favors the operating space of the dominant effector. For right-handers, this bias produces stronger coding of the stimulus location in the right visual hemifield, whereas for left-handers it produces stronger coding of the stimulus location in the left visual hemifield.

There are two ways to analyze the performance data in a Simon task, due to the fact that correspondence on a trial is determined by the relation between the stimulus location and response location on that trial. One common way to code data from a Simon task in an analysis of variance (ANOVA) is with correspondence (C: noncorresponding vs. corresponding) as a factor rather than separating stimulus location and response location (e.g., Proctor & Lu, 1999; Tagliabue et al., 2000). This coding simplifies interpretation of the results because the Simon effect is shown by a main effect of correspondence and its interaction with other variables can be easily evaluated. Rubichi and Nicoletti (2006) performed such a correspondence ANOVA, shown in the left column of Table 1, including the additional factors of stimulus location (StL: left vs. right) and handedness (H: left-handed vs. right-handed). In that type of ANOVA, the C main effect indicates an overall Simon effect and the asymmetry is indicated by the C × StL interaction. Accordingly, a difference in the asymmetry for left- and right-handed persons shows up as the 3-way interaction C × StL × H, which was significant in Rubichi and Nicoletti’s study.

Table 1

Comparison of Two Different ANOVAs for Evaluating the Simon Effect Asymmetry.


	CORRESPONDENCE ANOVA	STL × REL ANOVA

1	StL	StL

2	StL × C	ReL

3	StL × H	StL × H

4	StL × C × H	ReL × H

5	C	StL × ReL

6	C × H	StL × ReL × H

Note: Stimulus Location = StL; Response Location = ReL, Correspondence = C, and Handedness = H. Because the two ANOVAs are performed on the same data set, just partitioning them differently, the two terms on each row yield the same F ratios.

A limitation of using correspondence as a variable when evaluating whether the Simon effect is asymmetric is that the C × StL interaction includes a subtle confound. That is, the Simon effect for the left stimulus (St_left) is calculated by subtracting RT for the left response (Re_left) from RT for the right response (Re_right), whereas the Simon effect for the right stimulus (St_right) is calculated by subtracting RT for the right response (Re_right) from RT for the left response (Re_left). Thus, the dominant effector provides the corresponding response for one stimulus location but the noncorresponding response for the other, which may cause an asymmetry in the data that has nothing to do with correspondence. Consequently, the apparent asymmetry in Simon effects between left and right stimulus locations could instead be due to the difference between the response hands/locations (faster right than left responses for right-handed persons; faster left than right responses for left-handed persons).

An alternative way to analyze the same data from a Simon effect experiment is to code stimulus location (StL: St_left vs. St_right) and response location (ReL: Re_left vs. Re_right) as separate factors, instead of using correspondence as a factor. Tagliabue et al. (2007) pointed out that an ANOVA of the data coded in this manner, with handedness as a third factor, yields the same F ratios as the correspondence ANOVA, just with the relation to the factors changed. With the StL × ReL coding, the Simon effect is indicated by a significant interaction of stimulus location and response location rather than by a C main effect (compare the entries on row 5 of Table 1), with the resulting F ratios being identical. This equivalence of different terms holds true across the entire ANOVA produced using StL × ReL coding vs. correspondence coding: The same F ratios are output for the data set, but with the terms rearranged (as shown for all rows in Table 1).

Of importance, the C × StL interaction in the correspondence ANOVA, interpreted as a larger Simon effect for one of the two stimulus locations, shows up as a ReL main effect in the StL × ReL ANOVA (row 2). That is, this ANOVA shows that the apparent asymmetry in the Simon effect actually reflects faster responding with one hand than the other. Accordingly, the 3-way interaction of StL × C × H in the correspondence ANOVA is the two-way interaction term of ReL × H in the alternative ANOVA, reflecting faster responses with the dominant hand. In other words, because the measure of the Simon effect when stimulus and response locations are coded separately, the StL × ReL interaction, collapses across stimulus and response locations, it is not confounded with either, and the apparent interaction of the Simon effect with stimulus location is revealed to be a main effect of response location (or hand) that may interact with handedness. Note also that (a) even when there is no significant overall StL × ReL interaction, the response location main effect still reflects an asymmetry (opposing Simon effects for the left and right stimulus locations) and (b) the terms representing the stimulus location main effect and the stimulus location × handedness interaction do not change across the two ANOVAs, meaning that those terms are unrelated to the Simon effect.

Seibold et al. (2016) replicated the interaction of the Simon effect asymmetry with handedness found in the earlier studies, although their left-handed participants did not show a complete reversal to a larger effect for the left stimulus location. However, a StL × ReL ANOVA showed that the result was due to the following: Right-handers showed the Simon effect asymmetry favoring the right stimulus location, which was due to their responses being faster overall with the right than the left hand (i.e., a difference in effector efficiency for left and right effectors). Left-handers’ responses showed a nonsignificant opposite tendency, indicating that they did not show the benefit for right responses that the right-handers did.

Although less publicized than handedness, people also have a dominant foot used to perform manipulative actions (like kicking a ball), while the non-dominant foot provides stability (Gabbard & Hart, 1996). Foot dominance is highly correlated with hand dominance (e.g., Peters & Durding, 1979), although it is not quite as strong (i.e., there is more mixed footedness – i.e., no consistent dominance of the right or left foot – than mixed handedness; Tran et al., 2014). Foot-press responses are rarely used in studies of Simon-type tasks. One exception is a study by Proctor et al. (2003) which showed that a left-right Simon effect was obtained with stimuli that also varied along a vertical location dimension when a foot response was paired with a hand response. Another more relevant exception is a study of Phillips and Ward (2002, Experiment 3), which showed correspondence effects with left and right foot-press responses for the side to which object handles were oriented. However, since their analysis did not report the results separately for left and right feet, whether the Simon effect asymmetry was evident for foot-responses cannot be determined. Although not testing the Simon effect, Peters and Durding (1979) demonstrated that tapping responses can be made faster with the right foot than with the left foot for right-dominant persons. Thus, the effector efficiency hypothesis predicts that right-dominant persons should show a Simon effect asymmetry favoring the right stimulus location (i.e., dominant right-hand response) over the left location.

This prediction was tested by Chen et al. (2021) in a study that included only right-dominant persons (due to their availability). Chen et al.’s Experiment 1 demonstrated that pedal-press responses made with the left and right feet yielded a Simon effect asymmetry favoring the right response, like that obtained with the index fingers of the left and right hands. Experiments 2 and 3 showed that the Simon effect asymmetry for RT was evident even when the participants were unable to see the response device and their effectors while responding with the left and right feet or the index fingers. Although both the hands and feet showed a Simon effect asymmetry in RT, the pattern of incorrect responses (a preponderance of errors being incorrectly responding with the right foot on noncorresponding trials for which the left response would be correct) suggested that the asymmetry for the feet could be due at least in part to a bias to make the dominant response, whereas the asymmetry for hands is not due to such response bias. Regardless, for both response modes, the results suggested that, for right-dominant persons, the Simon effect asymmetry is mainly a consequence of comparing conditions for which one response is made by the faster dominant right effector and the other with the slower left effector.

The prediction for left-dominant persons is less clear-cut. Peters and Durding (1979) found that although right-handed persons made tapping responses faster with their right hand and right foot, left-handed persons (the majority of whom were also left-foot dominant) were faster tapping with the left hand but the right foot. Strictly based on effector dominance, as determined by answers to foot-preference questions, the Simon effect for left-dominant persons would be expected to favor the left foot. However, Peters and Durding’s results for tapping responses suggest that left-dominant persons may be more efficient pressing a pedal with the right foot. Thus, effector efficiency, instead of the nominal effector dominance based on subjective responses to questionnaires, may be a better indicator of the Simon effect asymmetry.

A general property of the Simon effect with left and right keypress responses is that it decreases across the RT distribution (De Jong et al., 1994; Proctor et al., 2011). Separate RT bin analyses of the Simon effect and the dominant-hand advantage across the RT distribution have found that for finger-press responses the dominant-hand advantage does not follow the time-course of the Simon effect, remaining flat or even increasing slightly across the RT distribution, suggesting that different processes underlie each (Rubichi & Nicoletti, 2006; Seibold et al., 2016). Such analyses have not been reported previously for foot-press or hand-press responses on pedals. If the difference in time courses is evident for those response modes as well, it will provide additional evidence that the effect of effector efficiency is distinct from the Simon effect.

To summarize, the present experiments were designed to address several issues that remained unresolved after Chen et al.’s (2021) study:

Because left-dominant persons were not tested by Chen et al., whether foot-press responses show the reversal of the Simon effect asymmetry predicted based on effector dominance remained unanswered. This void was filled in the present study by testing both left- and right-dominant persons.
Because the hand responses were made with presses of the index fingers on the outer two keys of a five-row response box and the foot responses with presses of the whole foot on driver-simulator gas pedals, the response keys and forces required to operate them were quite different. These differences were minimized in the present Experiment 1 by having the hand responses made with palm presses on the same pedals as the foot responses.
Due to said differences, the temporal dynamics of the Simon effect and its asymmetry across the RT distribution could not be examined meaningfully in Chen et al.’s study. The designs of the present experiments enabled comparison of the effects obtained with the respective response modes using delta plots of error rate and RT for measures of the Simon effect and its asymmetry across 20-percentile bins (Proctor et al., 2011).
Because, unlike prior findings with index-finger responses, the hand responses for the right-dominant persons did not show a significant Simon effect asymmetry in the present Experiment 1, Experiment 2 compared performance with hand presses of the pedals to that with finger presses on response-box keys.

Experiment 1: Hand Presses on Pedals vs. Foot Presses on Pedals

Experiment 1 evaluated the Simon effect and its asymmetry for left- and right-dominant persons. Hand- and foot-press responses were made by each participant in distinct trial blocks. The effector efficiency hypothesis predicts that right-dominant persons should show an advantage for the right response for both hand- and foot-press responses; left-dominant persons should show an advantage for the left hand-press response but, given Peters and Durding’s (1979) findings for tapping responses, possibly an advantage for the right foot-press response.

Method

Left-handed persons make up only about 10% of the population (de Kovel et al., 2019), which limits their availability as research participants. Prior studies included 20 and 16 participants per hand-dominance group (Rubichi et al., 2006, Experiments 1 and 2), 12 per group (Tagliabue et al., 2007, Experiment 2), and 24 per group (Seibold et al., 2016, Experiment 1). These numbers have been adequate to show the Simon effect and its asymmetry in all cases. In the present study, another limitation was that persons needed to be left- or right-dominant for both handedness and footedness, which restricted the eligible participants further (because the two do not correlate perfectly and mixed preference is more common for footedness; Tran et al., 2014). Thus, we tried to maximize the number of left-dominant persons available at the time and run a similar number of right-dominant persons. Also, to maximize sensitivity for detecting response mode effects, we manipulated responding with the hands or feet within-subjects.

Participants

Sixty-five undergraduates from introductory psychology classes at Purdue University participated for credit toward a course requirement. Preferences for the left or right hand were obtained from subjective responses to the Edinburgh Handedness Inventory (Oldfield, 1971), and preferences for the left or right foot were obtained from the Waterloo Footedness Questionnaire – Revised (Elias et al., 1998). Thirty-one participants were left-dominant (16 female, 15 male; mean age = 19.7 years, SD = 1.7 years): mean handedness score –87.02 ± 15.15 (on a scale of –100 to +100), and mean footedness score –9.23 ± 4.00 [total of answers to 10 questions of left always (–2), left usually (–1), equally often (0), right usually (+1), right always (+2)]. Thirty-four were right-dominant (22 female, 12 male; mean age = 18.6 years, SD = 1.0 years): mean handedness score 90.38 ± 9.23 and mean footedness score 10.63 ± 5.11.

Apparatus, Stimuli, and Procedure

The experiment was conducted on a Dell Optiplex 745 personal computer with a 19-in. LCD color monitor, set on top of a wood-frame box (see Figure 1). Stimulus presentation, response recording, and feedback were controlled by E-prime 2.0 software. Stimuli were red and green rectangles of 1.1° × 0.9° visual angle, presented 3.8° to the left or right of a centered fixation cross (0.7° × 0.7°), displayed on a white background. Responses were made by pressing a left or right Treadlite II foot pedal, placed on the tabletop (hand-press condition; see Figure 1, left) or floor (foot-press condition; see Figure 1, right).

Figure 1

Display Screen and Pedals on Table (Left Panel) and on Floor (Right Panel).

Each participant performed the Simon task for 12 practice trials and 120 experimental trials in each of two trial blocks, responding with the hands in one block and the feet in the other. The order of the response modes was approximately counterbalanced; within each trial block the stimulus colors and spatial locations occurred in a pseudo-random order, with equal numbers of corresponding and noncorresponding trials. Participants were seated in front of the monitor at about 60 cm and responded by pressing the left or right pedal. With the feet, a response was made by pressing downward with the front half of the foot, like pressing a gas pedal. With the hands, a response was made by a downward motion of the part of the palm near the wrist. This manner of responding was necessary to generate sufficient force to operate the pedal.

Each trial started with a 1,100-ms display of the fixation cross. The stimulus then appeared and stayed on the screen until a response was made or disappeared after 1,500 ms if there was no response. Following the response, a 500-ms feedback screen appeared that showed just the fixation cross for correct responses or the word “Error” below the fixation cross for incorrect responses.

Design

The independent variables were stimulus location (left vs. right), response location (left vs. right), and response mode (hands or feet), all within-subjects, and effector dominance (left-dominant vs. right-dominant), which was between-subjects. The dependent variables were RT and percentage error (PE). Separate ANOVAs were performed on the RT and PE data.

Results

Reaction Time

Error trials (2.5%) and trials with RT > 900 ms or < 100 ms (1.0%) were excluded from the RT analysis. For clarity, Figure 2 plots the Simon effect as a function of stimulus location and hand dominance. It shows overall Simon effects for both response modes and asymmetries of the Simon effect favoring the stimulus location corresponding with the dominant hand, except for foot responses of left-dominant participants. Recollect, though, this method of calculation of the Simon effect for the respective stimulus locations includes a confound with hand differences.

Bar graph showing Simon effects for left- and right-dominant persons

Figure 2

Experiment 1: Simon Effect in Reaction Times (RTs) as a Function of Stimulus Location and Response Mode (Hands, Feet).

Note: Error bars represent 95% confidence intervals.

We report the results of a StL × ReL ANOVA that distinguishes stimulus and response locations, as well as effector dominance and response mode, because it does not contain that confound. The mean data from this ANOVA are portrayed in Figure 3. Results showed a stimulus location × response location interaction, F(1, 63) = 103.25, p < .001, $η p 2 = .62$ M1 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .62\] \end{document} , indicating a Simon effect of 26 ms, and a main effect of response location, F(1, 63) = 8.09, p = .006, $η p 2 = .11$ M2 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .11\] \end{document} , reflecting an overall Simon effect asymmetry (right responses 9 ms < left responses; Seibold et al., 2016). This effect of response location interacted with response mode, F(1, 63) = 26.38, p < .001, $η p 2 = .30$ M3 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .30\] \end{document} : RT was shorter for right foot-press responses than for left foot-press responses (Ms = 493 vs. 516 ms), p < .001, but there was no significant difference between the left and right hand-press responses (Ms = 504 vs. 509 ms), p = .273.

Figure 3

Experiment 1: Reaction Times (RTs) for Right-Dominant (Top) and Left-Dominant (Bottom) Persons as Functions of Stimulus and Response Locations, and Response Mode (Hands, Feet).

Note: The error bars reflect Cousineau-Morey 95% confidence intervals (Morey, 2008).

Of primary interest was effector dominance, for which the main effect was not significant, F(1, 63) = 2.08, p = .155, $η p 2 = .03$ M4 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .03\] \end{document} . The interaction of effector dominance with stimulus location and response location was also nonsignificant, F(1, 63) = 2.62, p = .111, $η p 2 = .04$ M5 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .04\] \end{document} , indicating that the Simon effect did not differ reliably between the two effector dominance groups. Effector dominance did interact with response location, F(1, 63) = 5.61, p = .021, $η p 2 = .08$ M6 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .08\] \end{document} : The response-location effect was significant for the right-dominant group (Ms = 525 ms for left responses vs. 508 ms for right responses), F(1, 33) = 11.54, p = .002, $η p 2 = .26$ M7 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .26\] \end{document} , but not the left-dominant group, F < 1.0 (Ms = 495 ms for left responses vs. 494 ms for right responses). Another way to describe this result, thus, is that the right-dominant group showed a significant Simon effect asymmetry, but the left-dominant group did not.

Separate analyses of the respective effector dominance groups showed the following. For right-dominant persons, response mode entered into a 2-way interaction with response location F(1, 33) = 8.30, p = .007, $η p 2 = .20$ M8 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .20\] \end{document} , reflected by a significant difference in response location for the foot responses, p < .001, but not for the hand responses, p = .537. Response mode also entered a 3-way interaction with stimulus location and response location F(1, 33) = 5.38, p = .027, $η p 2 = .14$ M9 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .14\] \end{document} . As shown in Figure 3, these reflect that the Simon effect asymmetry and the Simon effect itself were larger with foot presses than with hand presses for the right-dominant persons. The left-dominant group also showed a significant interaction of response mode × response location, F(1, 30) = 22.84, p < .001, $η p 2 = .43$ M10 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .43\] \end{document} , but not the 3-way interaction, F < 1.0. The left hand-presses were faster than the right hand-presses (Ms = 485 ms vs. 499 ms), p = .012, but the right foot-responses were faster than the left foot-presses (Ms = 488 ms vs. 505 ms), p = .002. Thus, even though the participants in the latter group were classified as left-dominant by their foot preferences, their RT results were like those for right-foot dominant persons.

Returning to the overall ANOVA, effector dominance also interacted with stimulus location, F(1, 63) = 9.34, p = .003, $η p 2 = .13$ M11 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .13\] \end{document} : Right stimuli were responded to faster than left stimuli for the right-dominant group (Ms = 513 vs. 520 ms), p = .010, but there was no significant difference for the left-dominant group (Ms = 492 ms for left stimuli vs. 497 ms for right stimuli), p = .095. No other effect was significant, ps > .05.

Percentage Error

A similar overall ANOVA of PE including all independent variables was conducted. The stimulus location × response location interaction was significant, F(1, 63) = 23.09, p < .001, $η p 2 = .27$ M12 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .27\] \end{document} , reflecting a Simon effect in the PE data (see Figure 4). The 3-way interaction of those variables with effector dominance was also significant, F(1, 63) = 4.86, p = .031, $η p 2 = .07$ M13 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .07\] \end{document} . The PE Simon effect was larger for the left-dominant group than for the right-dominant group, but it was still significant for the latter group, F(1, 33) = 4.50, p = .042, $η p 2 = .12$ M14 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .12\] \end{document} . The main effect of response location was significant, F(1, 63) = 7.56, p = .008, $η p 2 = .11$ M15 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .11\] \end{document} , signaling a Simon effect asymmetry overall. This effect of response location was modulated by effector dominance, F(1, 63) = 8.51, p = .005, $η p 2 = .12$ M16 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .12\] \end{document} . Left responses had a higher PE than right responses for the left-dominant group (Ms = 2.5% vs. 1.0%), F(1, 30) = 20.79, p < .001, $η p 2 = .41$ M17 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .41\] \end{document} . There was no significant difference for the right-dominant group (Ms = 1.8% for left responses vs. 1.9% for right responses), F < 1.0, but that group showed a significant interaction of response mode with response location, F(1, 30) = 4.98, p = .042, $η p 2 = .13$ M18 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .13\] \end{document} . The same interaction was also at the .05 criterion in the overall ANOVA, F(1, 63) = 3.99, p = .050, $η p 2 = .06$ M19 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .06\] \end{document} .

Figure 4

Experiment 1: Percentage Error (PE) for Right-Dominant (Top) and Left-Dominant (Bottom) Persons as a Function of Stimulus and Response Locations, and Response Mode (Hands, Feet).

Note: The error bars reflect Cousineau-Morey 95% confidence intervals (Morey, 2008).

The analysis also showed a main effect of response mode, F(1, 63) = 6.45, p = .014, $η p 2 = .09$ M20 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .09\] \end{document} . Hand-press responses had a lower PE than foot-press responses (Ms = 1.5% vs. 2.1%). The interaction between stimulus location and effector dominance was also significant, F(1, 63) = 9.25, p = .003, $η p 2 = .13$ M21 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .13\] \end{document} . Right stimuli had a higher PE than left stimuli for the left-dominant group (Ms = 2.4% vs. 1.1%), p = .001, but there was no significant difference for the right-dominant group (Ms = 2.0% for left stimuli vs. 1.7% for right stimuli), p = .364. No other effect was significant, ps > .05.

Distribution Analyses

We describe the procedure for the PE bin analysis first because it was performed without omitting trials, whereas the RT bin analysis described after it was performed with error and outlier trials omitted. However, when reporting the results, we report the RT bin analysis first to be consistent with reporting of the main analyses.

All RTs were rank ordered including error and outlier trials per condition for each participant, from shortest to longest. To obtain the Simon effect for each bin, the ranking was conducted for the congruent and incongruent trials separately for each response mode (hand vs. foot). To obtain the response location effect (i.e., the difference in Simon effect for dominant and nondominant effectors), the ranking was conducted for the left and right response locations separately for each effector condition (hand vs. foot). After the trials were ranked by RT, they were equally divided into 5 bins, and PE was calculated for each bin. The Simon effect for each bin was then computed by subtracting the bin mean for congruent trials from that for incongruent trials. Similarly, the response location effect for each bin was calculated by subtracting the bin mean for the dominant trials from that for the non-dominant trials. Separate ANOVAs were conducted for the Simon effect and response location effect, with bin (1 through 5) and response mode (hand vs. foot) being within-subjects.

RT bin analysis was done like the PE analysis, except that error and outlier trials outside [100 – 900 ms] were omitted before ranking the trials and dividing them into bins.

Reaction Time

Figure 5 shows the distribution functions for the Simon and response-location effects in RT. The most obvious patterns from Figure 5 are: The RT Simon and response-location effects show different time courses, with the Simon effect decreasing across bins 2–5 and the response-location effect being relatively constant. Analysis of the Simon effect showed only a bin main effect, F(1.9, 122.3) = 13.17, p < .001, $η p 2 = .17$ M22 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .17\] \end{document} (Huynh-Feldt adjustment applied due to violation of sphericity assumption): The result pattern shows the typical finding of the Simon effect being largest early in the RT distribution and then decreasing. Although the Simon effect tended to be larger with the feet than the hands, the main effect was not significant, F(1, 64) = 2.54, p = .116, $η p 2 = .04$ M23 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .04\] \end{document} , nor was the interaction, F < 1.0. The similar time course for the hand and foot responses indicates that the Simon effect resides in response-selection processes separate from the effectors.

Effect sizes for hand and foot responses across five bins

Figure 5

Experiment 1: RT Simon and Response Location Effects as a Function of RT Bin.

The response-location effect also showed no significant factors, Fs < 1.0. The relative lack of change in the response-location effect across the RT distribution, compared to the decreasing function for the Simon effect, implies that the two effects are due to different factors.

Percentage Error

Figure 6 shows the distribution functions for the Simon and response-location effects in PE. The most notable feature of Figure 6 is that the Simon effect functions (solid lines) show more errors in the first bin than in the others, as verified by a bin main effect, F(2.1, 133.6) = 19.28, p < .001, $η p 2 = .23$ M24 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .23\] \end{document} (Huynh-Feldt adjustment applied). Neither a response mode main effect nor interaction was significant, ps > .10. This result pattern indicates a tendency to make the response corresponding to the stimulus location incorrectly on incongruent trials when responding quickly, regardless of whether responding with hands or feet. It thus provides confirming evidence that the Simon effect has a similar basis for the two response modes.

Figure 6

Experiment 1: PE Simon and Response-Location Effects as a Function of RT Bin.

The functions for the response-location effect (dashed lines) also show the most errors at the early bins, though the pattern is not as strong, as indicated by a nonsignificant bin main effect, F(2.8, 181.6) = 2.34, p = .078, $η p 2 = .04$ M25 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .04\] \end{document} (Huynh-Feldt adjustment applied). The largest effects were at the first and second bins, and the numerical values are negative, indicating a predominance of errors of responding with the non-dominant effector when the correct response should be the dominant one. The functions show no overall difference between feet and hands, F < 1.0, but the interaction of bin and response mode was significant, F(3.4, 219.4) = 2.87, p = .031, $η p 2 = .04.$ M26 \documentclass[10pt]{article} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \[\eta _p^2 = .04.\] \end{document} (Huynh-Feldt adjustment applied). This interaction is due mainly to the response-location effect being restricted primarily to bin 1 for the foot-press responses but lasting through bins 1–3 for the hand-press responses.

Discussion

As expected, a Simon effect was obtained for both hand- and foot-press responses and for left- and right-dominant persons. Of more interest was the Simon effect asymmetry, as reflected in the response location main effect, which was evident in the RT data. For hand-presses, the left-dominant group showed the expected shorter RT for the left response than for the right response. However, the right-dominant group did not show a significant asymmetry of shorter RT for the right response than the left response, as has been found in several other studies. This nonsignificant effect is addressed in Experiment 2. For foot-presses, not only the right-dominant group but also the left-dominant group showed an RT advantage for the right response. These results for hand- and foot-press responses agree with those obtained by Peters and Durding (1979) for tapping responses. This deviation of performance efficiency with responses requiring foot presses from the foot preferences expressed by left-footed persons on the footedness questionnaire suggests that there may be a unique property to press-types of movements made with the feet. We discuss this issue more in the General Discussion.

The error data for the foot responses showed an asymmetry of larger Simon effect for the right stimulus position than for the left stimulus position for both right- and left-dominant groups. In both cases, the PE was high when the stimulus appeared in the right location but required a left response. The hand responses showed a lack of significant asymmetry in the PE data as in the RT data for the right-dominant persons, but an asymmetry favoring the right stimulus location for the left-dominant persons, which is counter to the RT data.

For the bin analyses, the PE data showed that the Simon effect was strongest when responses were fast (i.e., in the first bin). This tendency for errors to be the fast responses was not evident in the Simon effect asymmetry, for which the analysis did not show a significant bin main effect. In the RT data, the Simon effect showed the typical pattern of peaking early in the RT distribution and then decreasing across the remainder of it, and this effect was evident for both response modes. The Simon effect asymmetry was relatively flat across the RT distribution, and again, this pattern was similar for both the hand- and foot-press responses. The different distribution patterns for the Simon effect and the asymmetry imply that they have distinct bases.

Experiment 2: Hand-Presses on Pedals vs. Finger-Presses on Response Box

In Experiment 1, left-dominant persons showed a Simon effect asymmetry in RT for hand-press responses that favored the left hand. However, right-dominant persons did not show the opposite RT asymmetry favoring the right hand, although their foot-press responses did show an asymmetry favoring the right foot. Because the right-hand advantage for right-dominant persons has been obtained reliably when press responses were made with the index fingers, rather than the palms, we deemed it necessary to re-run the hand-press condition, along with a finger-press response condition, in Experiment 2. Direct comparison of those conditions allowed us to evaluate whether the advantage for the right hand is less strong with hand presses than with finger presses, or whether the relation between the Simon effect asymmetry and dominant hand is distinct from how the responses with the left and right hands are implemented.