Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Sep 1.
Published in final edited form as: Psychol Sport Exerc. 2017 Jul 3;32:131–137. doi: 10.1016/j.psychsport.2017.07.001

Introductory Dialogue and the Köhler Effect in Software-Generated Workout Partners

Stephen Samendinger a,1, Samuel T Forlenza b, Brian Winn c, Emery J Max d, Norbert L Kerr e, Karin A Pfeiffer f, Deborah L Feltz g
PMCID: PMC5703210  NIHMSID: NIHMS891376  PMID: 29200958

Abstract

Objectives

This study explored the Köhler motivation gain effect utilizing adults and software-generated partners (SGPs) during an abdominal exercise regimen and compared the type of participant-SGP introductory dialogue as a moderator. The Köhler effect applies interdependent team dynamics in which group performance is dependent upon the weaker member. The third objective was to examine if this motivation paradigm would result in adverse consequences to secondary variables: exertion, enjoyment, and self-efficacy beliefs.

Design

Adults (Mage = 38.8 +/− 7.7) were randomly assigned to 1 of 3 conditions: Interactive Partner SGP (IP), Linear Dialogue Partner SGP (LDP), or individual control (IC), to complete a series of abdominal exercises. The experiment used a 3 (condition) x 2 (gender) ANCOVA design, with a baseline block of exercises as a covariate.

Method

Participants completed abdominal exercises individually and, after a rest, repeated the same exercises with either an SGP programmed to be moderately stronger or individually (IC). Prior to the second exercise block, IP participants interacted with the SGP using a dialog tree optional-response format. The LDP participant introduction was a linear, scripted exchange of basic information.

Results

The LDP and IP conditions persisted significantly longer than IC, generating moderate effect sizes (d = .62; d = .76). The mean difference between partnered conditions was not significant.

Conclusions

The Köhler motivation exercise paradigm resulted in a considerable increase in persistence (Madj = 28.9, SE = 10.6) in the first study to use middle-aged adults with superior SGPs. Differences between introductory dialogue methods were not significant.

Keywords: Motivation, Exergame, Köhler, Software-Generated Partner

Introduction

U.S. adults are not getting enough exercise, including strength training, at the recommended levels to maintain health and reduce the risk of chronic disease (U.S. Department of Health and Human Services, 2008). Motivation is a key issue in the physical inactivity epidemic (Dishman, 2001). Because it takes time and commitment to initiate and maintain a regular exercise program, especially when people’s lives are busy, it is easy to lose motivation. Exergames have become a popular solution to try to boost motivation with games that are entertaining and engaging (Lieberman, 2006). However, as Feltz and her colleagues noted (Feltz, Kerr, & Irwin, 2011), few exergames have taken advantage of the potential of group dynamics to motivate play, such as creating interdependence among exercisers where their progress and/or outcomes are mutually determined.

Recent research has applied the Köhler motivation gain effect to partnered or cooperative exergame play (Feltz et al., 2011; Irwin, Scorniaenchi, Kerr, Eisenmann, & Feltz, 2012; Kerr, Forlenza, Irwin, & Feltz, 2013). This well-studied motivation-gain effect occurs in team tasks when the less capable partner of a team exhibits greater motivation in terms of effort, relative to individual performance, when performing as part of a team on effort-based tasks (Hertel, Kerr & Messé, 2000; Kerr & Hertel, 2011) (See meta-analysis for a review: Weber & Hertel, 2007). The Köhler effect has been strongest in conjunctive task conditions (Steiner, 1972) where the team’s potential productivity is equal to the productivity of its least capable member. The performance gain by the less capable team member is thought to be the result of increased levels of motivation that stem from both being indispensable to the team and making a social comparison to one’s higher-ability partner (Kerr & Hertel, 2011). As instrumentality x value models of motivation suggest (e.g., Vroom, 1964; Karau & Williams, 1993), task motivation is likely to be increased when a team member sees his or her efforts as being highly instrumental (i.e., indispensable) in achieving the team outcome. Because motivation cannot be observed directly and must be inferred (Toure-Tillery & Fishbach, 2014), researchers studying the Köhler motivation effect use behavior in terms of performance effort to infer motivation.

The latest research on the Köhler effect in exergames has shifted toward software-generated partners (SGPs; Feltz, Forlenza, Winn, & Kerr, 2014). Using SGPs provides game designers and players with greater flexibility (e.g., partner availability, easier to change appearance, more programmable exercises). The major question that Feltz and colleagues (2014) asked was whether participants would willingly team up with an SGP or view the SGP as an incomparable other (weakening the Köhler effect). Results showed a significant motivation gain in terms of persistence in a core strength-training game for participants teamed up with an SGP, though this effect was smaller than with human partners (Feltz et al., 2014). These results were consistent with Media Equation research (Reeves & Nass, 1996), which suggests that people will often respond socially to software agents as if they were human.

Additionally, there was no evidence that working with SGPs harmed enjoyment or increased perceived exertion (Feltz et al., 2014), consistent with most prior findings in this area (Feltz et al., 2011; Feltz, Irwin, & Kerr, 2012; Forlenza, Kerr, Irwin, & Feltz, 2012; Kerr et al., 2013). While a promising start, the Feltz et al. (2014) study has been the only one to explore the Köhler effect in exergames with SGPs thus far. Furthermore, this study was conducted with college students; older age groups may react differently to SGPs because they play video games less frequently than college-aged adults (Lenhart, Jones, & Macgill, 2008; Pew, 2013).

In the Feltz et al. (2014) exergame with an SGP, the same protocol was used to introduce participants to their partner as in previous human partner experiments (Feltz et al., 2011). In the protocol, participants were informed that they would be working with a partner, and that his/her gameplay would be visible, projected onto a screen. Before they exercised together, participants met their SGP via a webcam-like connection, during which each introduced themselves to share basic information (e.g., favorite television shows, what they like to do for fun).

While this protocol works well for human partners, it may not be optimal for attempting to build a connection with SGPs (Feltz et al., 2014). Yet, having familiarity with one’s partner is beneficial for improving motivation (Kerr & Seok, 2010). Prior research has also suggested that people may have positive subsequent interactions and higher social attraction with a computer after an exchange of personal information (Kang & Gratch, 2011; Lee, Kiesler, & Forlizzi, 2010; Moon, 2000). One alternative strategy to a simple exchange of spoken information is to use a dialogue tree, which is more interactive and allows the introduction to be back-and-forth, like real conversations. Such interactivity may enhance perceptions of the SGP-human social relationship (Burgoon et al., 2000). These enhanced perceptions within the SGP-participant relationship (e.g., positive feelings, social attraction) may increase a greater sense of team identity and the likelihood that the Köhler effect mechanism of obligation to the team will occur. For example, the participant is less likely to feel indispensable to the team if he/she does not perceive value in doing so.

The purpose of this study was to replicate and extend prior findings based on two questions. First, would being paired with a moderately superior SGP result in significantly increased exercise persistence (i.e., a Köhler effect), compared to individual controls, in a sample of middle-aged adults in a core strength exergame (plank exercises)? That is, would participants persist longer as a result of the conjunctive design of the Köhler effect, utilizing SGPs as moderately better partners? Second, would the type of introduction (standard verbal versus interactive dialogue) moderate the Köhler effect? Our hypotheses were as follows: (1) Participants would persist longer when exercising with interdependent partners (SGPs) compared to exercising alone, (2) A larger Köhler effect would be observed when an interactive dialogue tree was used, and (3) There would be no adverse consequences to secondary variables like exertion, enjoyment, and self-efficacy beliefs.

Materials and Methods

Experimental Design and Participants

After Institutional Review Board approval, participants (N = 90) were recruited from a mid-western city, completed informed consent, and were randomly assigned to one of the three experimental conditions: Individual Control without a partner (IC), Linear Dialogue Partner (LDP), or Interactive Partner (IP). Participants were 90 adults (Mage = 38.8, SD = 7.7), 57 females (IP = 19, LDP = 20, IC = 18) and 33 males (IP = 11, LDP = 10, IC = 12). Participants received a small cash incentive at the end of the session. Power analyses were performed using G*Power software (Faul, Erdfelder, Lang, & Buchner, 2007). An a priori power analysis following f index recommendations for ANCOVA suggested we would detect a moderate Köhler effect (f = .35) with the sample size of 93 with a probability of .85. This projection was consistent with previous conjunctive-partnered studies of the Köhler effect conducted by the research group that have shown moderate-large treatment effects (e.g., Feltz et al., 2014; Feltz et al., 2012).

Exergame

CyBud-X2 was built on the Unity3D game engine, based roughly on the PlayStation 2 EyeToy: Kinetic exergame (though the Kinetic does not employ an SGP) and modified from the Cy Buddy Exercise (CyBud-X) game (see Feltz et al., 2014) with improved design of the SGP based on focus groups with the target audience. The 3-dimensional characters were created with Mixamo’s Fuse software and modified in Autodesk Maya and Photoshop to create a human-looking partner, comparable to those present in modern video games (see Figure 1). The male and female SGPs speak through the voice of prerecorded voice actors and move with human-like motion based on a combination of human capture, key frame animation, and software generated movements. The CyBud-X2 uses a webcam to project the participant’s image onto the screen during plank exercises. Partnered versions of CyBud-X2 included two experimental participant-SGP introduction options. In the IP condition, the conversation was built using a dialog tree where the SGP would speak followed by two or three answer options for the participant. For example, the SGP offered his/her name (Chris) and asked the participant if s/he was ‘from around here’. If the participant selected ‘Yeah, I am’, the SGP replied: ‘Ah, I knew it. You just kind a had that look about you. I am too! I love our sports teams. But, wait, do you even like sports?’ The dialogue tree conversation proceeds through four branches of dialogue based on the choices the participant made and more closely resembled an actual conversation where both parties have some control over how the conversation plays out. In the LDP condition, the participant-SGP introduction is through a linear, scripted exchange of basic information (e.g., name, hometown, occupation, fun pastimes), with no participant influence on what the SGP said.

Figure 1.

Figure 1

Open in a new tab

Experimental Condition Software-Generated Partners (SGPs)

Procedure

Each participant watched an instructional video that provided an overview of the session, an explanation of the exertion scale, and a demonstration of five abdominal plank exercises. Participants were instructed to hold each plank for as long as possible, with 30-s breaks between exercises.

All participants completed five plank exercises (Block 1) individually to serve as a baseline performance. Upon completion of Block 1, participants in the IC condition were told the average time they held the exercises and informed they would complete the same set of exercises again (Block 2) after a 10-min rest. The IP and LDP condition participants were informed they would repeat the exercises with a same-gender SGP, tailored to be slightly more in shape than the participant was, and would work together toward a team score.

Participants in partnered conditions completed the SGP introduction, after which the Köhler Effect manipulations took place. Participants were informed that they, and their SGP, would simultaneously hold the planks for as long as possible and the team’s score would be the number of seconds the first team member to quit was able to hold the plank. This conjunctive task set-up (Steiner, 1972) was followed by veridical performance feedback from the participant’s first set of plank exercises. Participants were then informed that their SGP was programmed to complete the same set of planks for a slightly longer time (roughly 40%) than the participant but was also variable and fatigable over time. This superior SGP performance data invites upward social comparison toward SGPs, tempered with an additional explanation that SGPs were unable to hold the exercises forever and would tire, just like a real person. The moderate performance discrepancy of 40% was chosen based on previous research that demonstrated the perception of too similar or too large of an ability discrepancy is not as motivating as one that is moderate and offers a reasonable achievement goal for the weaker partner (Feltz et al., 2012; Messé, Hertel, Kerr, Lount, & Park, 2002). With abdominal plank exercises, the 40% difference in ability presented to the participants represents a realistic comparison challenge to motivate persistence and has been utilized successfully in previous similar paradigms (Feltz et al., 2012; Feltz et al., 2011; Feltz et al., 2014). Unbeknownst to participants, SGPs were programmed to hold the exercises indefinitely, never quitting first. Hence, the participant always defined the group’s score. Upon completing Block 2, participants completed the remaining questionnaires and were debriefed.

Measures

Persistence

The abdominal plank exercises are simple isometric routines, during which participants use their abdominal muscles to suspend their own body weight. The exercises were chosen as they require very little coordination, skill, or training, are highly effort-based, and have practical importance to core strength fitness. A Block score (i.e., time in seconds) was the average of all 5 plank exercises, each measured from when the position was achieved to when the participant stopped. Block 1 served as the participant baseline performance. Block 2 represented effort with the same exercise subsequent to the manipulations in experimental groups compared to performance effort in the control condition.

Ratings of perceived exertion

The Borg Rating of Perceived Exertion scale (RPE; Borg, 1998) captured ratings of overall exertion immediately upon quitting each exercise. The scale ranges from 6 to 20, with 6 being “no exertion at all” and 20 being “maximal exertion.” Borg correlated the scale’s subjective responses upon exertion to heart rate measurements during the same exercise. (Borg scale response multiplied by 10 provides an estimate of heart rate). Exertion ratings were later averaged over the five planks in each Block.

Self-efficacy beliefs

Participants estimated the number of seconds they were completely confident they could hold each exercise immediately before Blocks 1 and 2, and after Block 2. For each time point, the estimated five hold times were averaged as the self-efficacy score.

Enjoyment

After finishing Block 2, participants completed the Physical Activity Enjoyment Scale (PACES; Kendzierski & DeCarlo, 1991; Raedeke & Amorose, 2013). The 5-item version utilizes a 3-point bipolar scale and responses were averaged for an overall task enjoyment score.

Participant-Partner relationship

As attitudes toward the SGP and perceptions of the validity of a relationship with the SGP may affect Köhler group dynamics (social comparison and indispensability), measures of the participants’ relationship with their SGPs were collected upon completing Block 2. Participants’ feelings toward their partner were surveyed using 4-items on a 5-point rating scale (e.g., “I liked my partner,” “I felt comfortable with my partner”). Exercise team perceptions (5-items, e.g., “I felt I was part of a team,” “I thought of my partner as a teammate”) were also collected using a 9-point scale (Nass, Fogg, & Moon, 1996).

IP and LDP participants also completed the Alternative Godspeed Indices (Ho & MacDorman, 2010), a 19-item semantic differential survey with 3 sub-scales: humanness (e.g., artificial vs. natural), eeriness (e.g., bland vs. uncanny), and attractiveness (e.g., repulsive vs. agreeable). Ho and MacDorman (2010) verified the theoretical structure of the 3 sub-scales with a confirmatory factor analysis. The attractiveness, eeriness, and humanness sub-scales demonstrated sufficient internal consistency (Cronbach’s α values .90, .74, and .92, respectively). This questionnaire attempts to capture participants’ emotional responses to the SGP, to probe whether the SGP might fall into the so-called “Uncanny Valley” (Mori, 1970). This valley represents a negative dip in acceptance with a character’s human likeness when graphing the relationship between comfort and humanness. For example, one’s comfort level with a software-generated character may increase linearly as the character is made to appear more human. However, imperfections in characteristics or movements of the non-human being may result in negative reactions and dramatically decreased comfort level by the human perceiver (Ho & MacDorman, 2010).

Results

Manipulation checks

During the final survey, partnered participants were asked to respond to questions related to the conjunctive manipulation (i.e., SGP was tailored to be slightly more in shape than the participant was, and both would work together toward a team score). To assess the belief that the SGP was relatively more in shape than the participant, responses were gathered on the following: “How do you feel you compared to your partner?” (1 = much less capable, 7 = much more capable). Results indicated participants in both partnered conditions recognized that they were the weaker partner (IP: M = 1.9, SD = .75; LDP: M = 2.2, SD = 1.2; p > .37). The combined mean scores were significantly below the scale mid-point (M = 2.1, SD = .99; t58 = 15.1, p < .001). When asked: “How important did you think your performance would be to the team score?” (1 = not important, 7 = very important), the response means were as follows: IP, M = 5.9, SD = 1.4; LDP, M = 5.28, SD = 1.8. There was no significant difference between the two groups but the combined mean was significantly higher than the scale mid-point (t57 = 7.4, p < .001), suggesting participants believed their performance was important to the team score. Finally, all participants were asked to select from three choices how their total score was determined during the set of exercises: 1 = length I held exercises, 2 = team score, first team member to quit, 3 = team score, last team member to quit. Eighty-five percent of the partnered participants selected the appropriate response (item 2). Twelve percent chose item 1, which may actually be considered appropriate if the participant believed they were indeed the weaker teammate. Taken together, the results of these manipulation checks show that participants believed their partner was more capable, that their own performance would be crucial for the team score, and that they understood the conjunctive task demands, indicating the manipulation was successful.

Persistence

To examine the effect of the Köhler group dynamic on the motivation to persist with the plank exercises, the three group’s average Block 2 plank times were used as the primary dependent variable. There were no significant differences observed between conditions on Block 1 plank times (F2,88 = 0.49 P = .614) and these baseline performance times were used as a covariate in the analysis. Using Block 1 as a covariate in the analysis of Block 2 score comparisons controls for individual differences and equalizes the baseline for all participants. The ANCOVA resulted in a main effect for Condition, F(2,82) = 4.5, P = .014. The LDP and IP experimental conditions persisted longer than the IC (see Figure 2 for adjusted means). The simple contrast showed that the adjusted mean difference between LDP and IC was significant (M = 25.60, P = .017, 95% CI [4.76, 46.44]) and generated a moderate effect size (d = .62). Likewise, the contrast of the IP versus IC demonstrated a moderate-large effect size (M = 28.87, P < .001, 95% CI [7.75,49.99]; d = 76). Contrast results of Block 2 persistence times for the partnered conditions versus IC suggest a Köhler motivation gain effect may explain persistence differences and confirms our primary hypothesis. However, Hypothesis 2 was disconfirmed, as mean differences between the two experimental partnered conditions were not significant (M = 5.49, P = .76, 95% CI [−24.34, 17.81]).

Figure 2.

Figure 2

Open in a new tab

Block 2 Persistence Times, Block 1 as covariate (Adjusted Means & Confidence Intervals)

Ancillary Analyses

No significant differences between conditions emerged for perceived exertion, self-efficacy beliefs, and enjoyment ratings. The average rating of exertion for Block 1 (all conditions) was 15.1 and the average for Block 2 was 15.5, representing “Hard/Heavy” exertion. Differences in self-efficacy were not significant, as participants in all three conditions projected a similar number of seconds they could confidently hold each of the five plank exercises at each time point, but their estimates decreased at each measurement point (T1 M = 48.9 sec, SD = 59.1; T2 M = 34.3 sec, SD = 20.9; T3 M = 27.6 sec, SD = 16.7). Enjoyment ratings were nearly identical for all conditions (F2,85 = 0.001, P = .99). The overall average enjoyment rating (M = 2.4) was significantly higher than the scale mid-point (t(85) = 6.56, P < .001), indicating moderately positive feelings of enjoyment about the physical activity performed. Correlations were run to explore potential useful associations between primary and secondary variables. No relationship was noted between enjoyment and self-efficacy, exertion perceptions, or Block 1 and 2 persistence times.

Participant-Partner relationship

Participants’ feelings toward their partner were nearly identical for both the IP and the LDP conditions (IP M = 3.3, SD = 0.66; LDP M = 3.3, SD = 0.69), with the combined mean significantly greater than the scale mid-point toward positive partner feelings, t(58) = 3.62, P < .001.

There were no significant differences between the two partnered conditions (F1,58 = 0.59, P = .446) in team perceptions. The combined mean (M = 3.96, SD= 1.66) was significantly lower than the mid-point (4.5), suggesting participants in both conditions did not strongly perceive that they were part of a team, t(58) = −4.79, P < .001. Partner perceptions were correlated to team (r = .46, P < .001) perceptions.

Alternative Godspeed Indices (humanness, attractiveness, eeriness) were not significantly different between the SGP groups. Indices for eeriness (M = 2.4, SD = 0.55) and humanness (M = 2.4, SD = 0.94) were negatively and significantly different from the scale mid-point (3), implying that participants in both partnered conditions perceived the partner to be not exactly like a human but also not eerie, t(57) = −8.1, P < .001; t(57) = −5.3, P < .001. The index (M =3.6, SD = 0.57) for attractiveness was also significantly different from the mid-point (3), suggesting the participants in both conditions found the partners to be more attractive than not, t(57) = 7.7, P < .001.

Discussion

This study sought to provide further support that the Köhler Effect can be demonstrated with an exergame using software-generated partners (SGPs) to improve motivation as measured by persistence on a core-strength task in an adult sample. This study also explored whether an interactive, conversation-like meeting with the SGP would strengthen the Köhler effect with middle-aged adults.

Overall, participants working with an SGP persisted significantly longer compared to participants working individually, with the differences achieving moderate effect sizes (Cohen, 1988). Controlling for baseline performances, LDP participants (M = 235.37) persisted 28.9 seconds longer than those in the IC (M = 206.47), resulting in a 14% increase. This increase represents a substantial gain in just one session, practically speaking, for adults who might be working on improving core stability in therapeutic or training settings. This study was the first to use middle-aged adults to partner with superior SGPs in an exergame based on conjunctive-partnered dynamics to test the Köhler motivation effect, suggesting SGPs may be as effective at boosting motivation for an older population as they have been in younger populations.

The second hypothesis was not supported, as there were no statistically significant differences between partner conditions and the results did not align with predictions in that participants who met their partner using an interactive dialogue (IP) would persist longer than linear dialogue introduction (LDP) participants. Based on these findings, the type of introductory dialogue between participant and SGP may not influence performance in the Köhler paradigm.

The third hypothesis concerning ancillary measures was supported, consistent with prior research (Feltz et al., 2011; Kerr et al., 2013; Feltz et al., 2012; Forlenza et al., 2012). Interestingly, even though participants with an SGP performed as the “weak link” in the dyad, beliefs in their ability, exertion, and enjoyment of the game did not suffer compared to the control group.

Participants generally had positive feelings toward their partners regardless of which type of interaction they had, but this did not translate to strong feelings of being part of a team. Even though participants were not particularly affiliated with their partner, they still persisted longer compared to working alone, which means that the SGP influenced persistence. As Kerr and Hertel (2011) suggest, both social comparison and indispensability mechanisms are in play when working under conjunctive task conditions. Social comparison may have played a stronger role than indispensability to the SGP, where he/she may have revised his/her personal performance goal upward or competed with the partner during the second trial block. An intra-group competitive response does not negate the influence of the SGP as a teammate who provides the opportunity for the participant to compare performance levels, which is crucial for the Köhler motivation effect. However, our study was limited to one session. There is evidence that more experience exercising with a conjunctive teammate over time increases the indispensability mechanism of Köhler effect, at least with human partners (Irwin et al., 2012).

Measures of humanness, eeriness, and attractiveness did not differ by condition. Participants judged their partners as not human-like, yet did not find that non-humanness to be eerie or unattractive. Rather, coupled with the positive feelings towards their SGPs, it seems that participants developed a favorable impression of the SGP. These results may explain why the second hypothesis was not supported. An interactive dialogue (versus the standard one-time exchange of information) did not seem to enhance the social relationship between participants and SGPs, hence there were no differences in later performance. Rather than focusing on the type of introductory dialogue, future research could explore the content of the introduction. The information revealed by SGPs in both conditions was generally superficial (e.g., hometown, favorite sports team, hobbies). However, SGPs that reveal ‘intimate’ or ‘deep’ information about themselves may lead to stronger social connections, in turn influencing performance (Kang & Gratch, 2011).

Feltz et al. (2014) explored partner humanness as a potential moderator of the Köhler effect in a college-aged sample, and found that the effect was significantly weaker with SGPs than with human partners. The effect size diminished as the partner became less human, moving from a human (d = 1.41) to a nearly human-like partner (d = 0.76) and, finally to an SGP (d = 0.57). The effect sizes generated in this study (d = .62, d = .76) of middle-aged adults are in line with these previous findings and are encouraging for the use of superior software-generated partners as a source of exercise persistence motivation gains. Thus, the motivation gain achieved so far with SGPs seems to be real, comparable for younger and older people, but smaller in magnitude than has been observed with a human partner in the Köhler exercise paradigm. However, the effect is similar to the large effect sizes noted for both simple physical tasks (e.g., holding a weight over a trip wire) and cognitive tasks (estimated to be g = 0.72 for conjunctive versus control in a 2007 meta-analysis: Weber & Hertel, 2007). This leaves an interesting open question to pursue: how to strengthen the effect with SGPs, particularly in light of growing video and exergame use. We suggest the focus should be on enhancing the relationship between participants and their SGPs, whether through alternative forms of interaction or creative ways to boost mutual trust and identification (e.g., inducing greater mutual interdependence before exergame participation; manipulating common bonds). Strengthening the participant-SGP relationship may increase the likelihood and magnitude of social comparison and team indispensability processes responsible for motivation gains in Köhler effect research designs.

The implications for this research speak to continued efforts to increase motivation toward meeting the recommended physical activity guidelines and potentially reduce health risk associated with inactivity. The Köhler motivation effect is one of only a few group dynamics capable of realizing motivation gains (versus simply reducing motivation losses that may occur, such as social loafing). Further understanding of motivation mechanisms and the application of robust group dynamics principles are vital steps in designing effective group-based physical activity interventions. Adapting research designs to software-generated (i.e., virtual) environments and partners expands the opportunities to apply the Köhler motivation gain effect to a broader scope of populations: home-based active video games; rehab settings; fitness and training regimens. Finally, this study was limited to a single session, in a lab setting, and future research should also explore the motivation effect of the Kohler dynamic in free-living environments over time. This line of study will allow for mobile applications and increase the range of potential exercise interventions (e.g., walking, jogging, cycling) with which these dynamics may be applied.

Conclusion

This study successfully replicated a Köhler motivation gain effect research paradigm previously utilized with exergames and human (confederate) experimental partners. In doing so, the current design also incorporated only software-generated partners with middle-aged adults for the first time instead of human partners. Partnered-condition participants persisted with the exercises significantly longer than those in the no-partner control condition, while the method of partner introductory dialogue did not seem to influence the magnitude of this effect. These findings offer continued support for the robust nature of the Köhler motivation effect, as it has now been adapted to exergame environments and software-generated partners. The relative success of these adaptations may bridge to diverse and practical applications of the Köhler motivation effect to increase physical activity.

Highlights.

  • We studied adult exergame motivation gains with software-generated partners (SGPs).

  • The Köhler motivation gain effect was demonstrated in an exergame using only SGPs.

  • Exercise performance differences achieved a significant moderate-large effect size.

  • An interactive introduction with the SGP did not strengthen the Köhler Effect.

  • Teammates of SGPs performed as the weak link but enjoyment beliefs did not suffer.

Acknowledgments

The authors would like to thank Gregory Kozma and William Jeffery for their help with the development of CyBud-X2.

This study was financially supported by a research grant from the National Heart, Lung, and Blood Institute (National Institutes of Health): 1R21HL111916-01A1.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Borg G. Borg’s perceived exertion and pain scales. viii. Champaign, IL, US: Human Kinetics; 1998. [Google Scholar]
  2. Burgoon JK, Bonito JA, Bengtsson B, Cederberg C, Lundeberg M, Allspach L. Interactivity in human–computer interaction: A study of credibility, understanding, and influence. Computers in Human Behavior. 2000;16(6):553–574. [Google Scholar]
  3. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Lawrence Erlbaum Associates; 1988. [Google Scholar]
  4. Dishman RK. The problem of exercise adherence: fighting sloth in nations with market economies. Quest. 2001;53(3):279–294. http://doi.org/10.1080/00336297.2001.10491745. [Google Scholar]
  5. Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods. 2007;39:175–191. doi: 10.3758/bf03193146. [DOI] [PubMed] [Google Scholar]
  6. Feltz DL, Forlenza ST, Winn B, Kerr NL. Cyber Buddy is better than no buddy: A test of the Köhler motivation effect in exergames. Games for Health Journal. 2014;3(2):98–105. doi: 10.1089/g4h.2013.0088. http://doi.org/10.1089/g4h.2013.0088. [DOI] [PubMed] [Google Scholar]
  7. Feltz DL, Irwin B, Kerr N. Two-player partnered exergame for obesity prevention: using discrepancy in players’ abilities as a strategy to motivate physical activity. Journal of Diabetes Science and Technology. 2012;6(4):820. doi: 10.1177/193229681200600413. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Feltz DL, Kerr NL, Irwin BC. Buddy up: the Köhler effect applied to health games. Journal of Sport & Exercise Psychology. 2011;33(4):506–526. doi: 10.1123/jsep.33.4.506. [DOI] [PubMed] [Google Scholar]
  9. Forlenza ST, Kerr NL, Irwin BC, Feltz DL. Is my exercise partner similar enough? Partner characteristics as a moderator of the Köhler effect in exergames. Games for Health Journal. 2012;1(6):436–441. doi: 10.1089/g4h.2012.0047. http://doi.org/10.1089/g4h.2012.0047. [DOI] [PubMed] [Google Scholar]
  10. Ho CC, MacDorman KF. Revisiting the uncanny valley theory: Developing and validating an alternative to the Godspeed indices. Computers in Human Behavior. 2010;26(6):1508–1518. http://doi.org/10.1016/j.chb.2010.05.015. [Google Scholar]
  11. Irwin BC, Scorniaenchi J, Kerr NL, Eisenmann JC, Feltz DL. Aerobic exercise is promoted when individual performance affects the group: a test of the Kohler motivation gain effect. Annals of Behavioral Medicine: A Publication of the Society of Behavioral Medicine. 2012;44(2):151–159. doi: 10.1007/s12160-012-9367-4. http://doi.org/10.1007/s12160-012-9367-4. [DOI] [PubMed] [Google Scholar]
  12. Kang SH, Gratch J. People like virtual counselors that highly-disclose about themselves. Studies in Health Technology and Informatics. 2011;167:143–148. [PubMed] [Google Scholar]
  13. Kendzierski D, DeCarlo KJ. Physical Activity Enjoyment Scale: Two validation studies. Journal of Sport and Exercise Psychology. 1991;13(1):50–64. [Google Scholar]
  14. Kerr NL, Forlenza ST, Irwin BC, Feltz DL. “… been down so long …”: Perpetual vs. intermittent inferiority and the Köhler group motivation gain in exercise groups. Group Dynamics: Theory, Research, and Practice. 2013;17(2):67–80. http://doi.org/10.1037/a0031588. [Google Scholar]
  15. Kerr NL, Hertel G. The Köhler group motivation gain: How to motivate the “weak links” in a group. Social and Personality Psychology Compass. 2011;5(1):43–55. http://doi.org/10.1111/j.1751-9004.2010.00333.x. [Google Scholar]
  16. Kerr NL, Messé LA, Seok DH, Sambolec EJ, Lount RB, Jr, Park ES. Psychological mechanisms underlying the Köhler motivation gain. Personality & Social Psychology Bulletin. 2007;33(6):828–841. doi: 10.1177/0146167207301020. http://doi.org/10.1177/0146167207301020. [DOI] [PubMed] [Google Scholar]
  17. Kerr NL, Seok DH. “…with a little help from my friends”: friendship, effort norms, and group motivation gain. Journal of Managerial Psychology. 2010;26(3):205–218. http://doi.org/http://dx.doi.org.proxy2.cl.msu.edu.proxy1.cl.msu.edu/10.1108/02683941111112640. [Google Scholar]
  18. Lee MK, Kiesler S, Forlizzi J. Receptionist or information kiosk: how do people talk with a robot?. Proceedings of the 2010 ACM conference on Computer supported cooperative work; ACM; 2010. pp. 31–40. [Google Scholar]
  19. Lenhart A, Jones S, Macgill A. Pew Internet & American Life Project data memo. 2008 Retrieved from www.pewinternet.org.
  20. Lieberman DA. What can we learn from playing interactive games. In: Vorderer P, Bryant J, editors. Playing video games: Motives, responses, and consequences. Lawrence Erlbaum Associates; 2006. pp. 379–397. [Google Scholar]
  21. Messé LA, Hertel G, Kerr NL, Lount RB, Jr, Park ES. Knowledge of partner’s ability as a moderator of group motivation gains: An exploration of the Köhler discrepancy effect. Journal of Personality and Social Psychology. 2002;82(6):935–946. doi: 10.1037//0022-3514.82.6.935. http://doi.org/10.1037//0022-3514.82.6.935. [DOI] [PubMed] [Google Scholar]
  22. Moon Y. Intimate Exchanges: Using computers to elicit self-disclosure from consumers. Journal of Consumer Research. 2000;26(4):323–339. http://doi.org/10.1086/209566. [Google Scholar]
  23. Mori M. The uncanny valley. Energy. 1970;7(4):33–35. [Google Scholar]
  24. Nass C, Fogg BJ, Moon Y. Can computers be teammates? International Journal of Human-Computer Studies. 1996;45(6):669–678. http://doi.org/10.1006/ijhc.1996.0073. [Google Scholar]
  25. Pew Internet & American Life Project spring tracking survey. [accessed January 28, 2014];2013 www.pewinternet.org/
  26. Raedeke TD, Amorose AJ. Journal of Sport & Exercise Psychology. Vol. 35. Human Kinetics; 2013. A psychometric evaluation of a short exercise enjoyment measure; pp. S75–122. [Google Scholar]
  27. Reeves B, Nass C. The media equation: how people treat computers, television, and new media like real people and places. New York, NY, USA: Cambridge University Press; 1996. [Google Scholar]
  28. Steiner ID. Group process and productivity. New York: 1972. [Google Scholar]
  29. Toure-Tillery M, Fishbach A. How to measure motivation: A guide for the experimental social psychologist. Social and Personality Psychology Compass. 2014;8:328–341. [Google Scholar]
  30. U.S. Department of Health and Human Services. Two Thousand Eight Physical Activity Guidelines for Americans: Be Active, Healthy, and Happy. Government Printing Office; 2008. Retrieved from http://www.health.gov/paguidelines/guidelines/default.aspx. [Google Scholar]
  31. Weber B, Hertel G. Motivation gains of inferior group members: a meta-analytical review. Journal of Personality and Social Psychology. 2007;93(6):973–993. doi: 10.1037/0022-3514.93.6.973. http://doi.org/10.1037/0022-3514.93.6.973. [DOI] [PubMed] [Google Scholar]

RESOURCES