Game Studies

Towards a Comprehensive Model of Mediating Frustration in Videogames

Abstract

The article presents the findings of a research project exploring how players react to frustrating game segments in inherently enjoyable gameplay and what keeps players motivated in these scenarios. The presented study found its motivation in the fact that the prominent literature is lacking a deeper understanding of micro changes in motivation during videogame play. The research utilized semi-structured interviews and a focus group with 9 young (age M=26) male participants and template analysis as its methodological framework. The result of the study is a theoretical model of a process of orientational shifts in situational motivation, which outlines how the motivational system of Self-Determination Theory could be understood in terms of the attentional and informational processing system, presented in Flow Theory. The paper outlines the further research avenues based on these findings and argues for the benefit of extrinsic motivation in suboptimal gameplay scenarios. The model provides a clear vocabulary that can complement the fast pace and complex environment of videogames, while help comprehending the players' ever-changing subjective experience and intentions.

Keywords: attention, flow, frustration, information processing, motivation, self-determination, videogame

Introduction

This article presents a theoretical model of how players mediate their motivation under highly frustrating gameplay scenarios. The presented model is the result of a small-scale qualitative study based on the Hierarchical Model of Motivation (Vallerand & Ratelle, 2002) (a theory expanding on Self-Determination Theory (Ryan & Deci, 2000)) and Flow Theory (Nakamura & Csikszentmihalyi, 2002).

Although the question of motivation in gameplay is a well-researched topic (Bartle, 2004; Rigby, 2004; Salen & Zimmerman, 2004; Yee, 2005; Ryan, Rigby & Przybylski, 2006; Rigby & Ryan, 2007, 2011; Przybylski, Rigby & Ryan, 2010; Lafrenière et al., 2012), this paper wishes to present a new angle on the subject. The bulk of the contemporary literature focuses on either social impact of videogame use (Vallerand et al., 2003) or aiming to understand what makes videogames fun (Bartle, 2004, p. 129). This latter approach often takes a practical approach, trying to "crack" good game design (Bartle, 2004, p. xix; Salen & Zimmerman, 2004, p. 1).

While such contributions are valuable to the fields of game research and design, there is still a lack of studies in game psychology (Rigby & Ryan, 2011, p. xii). Even though psychological theories have been used to understand and analyze game design (Jones, 1998; Sweetser & Wyeth, 2005; Cowley et al. 2008) and predict player behaviour (Bartle, 2004; Yee, 2005), there is still research to be done to understand the shifts in the players' experience during long, continuous play sessions.

Since most studies are preoccupied with questions like why people play on a macro level (Yee, 2005) or enjoy playing games in the first place (Sweetser & Wyeth, 2005), the focus is predominantly on intrinsically motivating experiences and their effect on the players' behaviour and wellbeing (Ryan, Rigby & Przybylski, p. 348; Rigby & Ryan, 2011, p. 10). This also means less attention is spent on the research of frustration and studies that do are often focusing on societal and psychological outcomes rather than the gameplay experience (Przybylski et al. 2014; Breuer et al. 2015).

The purpose of this article is twofold. Its main goal is to lay down the groundwork for future research by presenting a theoretical model of orientational shifts in situational motivation during frustrating game segments. Its secondary goal is to provide a vocabulary that builds on and integrates aspects of Flow Theory and the Hierarchical Model of Motivation (and Self-Determination Theory), providing novel material that so far has been overlooked in the field of videogame research.

The presented pilot research of frustration and motivation in videogames found patterns of player behaviour, in which players — after initial intrinsically motivating game segments — persisted playing through frustrating scenarios to recapture the initial positive experience. After they successfully escaped the frustrating segment, they reverted to their initial state of intrinsic interest. This suggests a process, in which the perceived locus of causality of players' motivation is situationally shifting from intrinsic to extrinsic, while their prior interest — or focus of play — and intrinsic contextual motivation stays the same.

The next section will expand on the theoretical background of the motivational system presented by Self-Determination Theory and the Hierarchical Model of Motivation and how these concepts could be understood in the terms of the attentional and informational system of Flow Theory. The second half of the article will build on this vocabulary and present a theoretical model of the findings. The paper will close with briefly discussing possible future research avenues.

Theoretical Background

The following section will provide the basic vocabulary through the relevant elements of Self-Determination Theory and the Hierarchical Model of Motivation, and also present the underlying attentional and informational framework of Flow Theory. This section will close with an interpretation of how the two systems can be connected to produce a coherent framework, in which we can discuss continuously changing moment-to-moment motivation.

Self-Determination Theory and the Hierarchical Model of Motivation

Self-Determination Theory (SDT) was chosen as a fundamental basis for the research because of its versatility and universality (Deci & Ryan, 2002, p. 26; Ryan, Rigby & Przybylski, 2006, p. 348). It has been proven that the motivational system presented by SDT has greater predictive power than other motivational models (Rigby & Ryan, 2007). The Player Experience of Need Satisfaction (PENS) framework showed the utility of SDT in the field of videogame studies (Ryan, Rigby & Przybylski, 2006, p. 349; pp. 351-352), although it mainly focuses only on intrinsic motivation and evaluates games on a macro level.

SDT describes motivation based on so-called basic psychological needs (Ryan & Deci, 2000) and the perceived locus of causality of the motivation (Ryan & Connell, 1989). The basic psychological needs, competence (innate need to grow and gain mastery), autonomy (desire of volition), and relatedness (meaningful connection) (Rigby & Ryan, 2011) are the basis for the formulation of self-determination that in turn inspires intrinsic motivation (Deci & Ryan, 2000, p. 230; Vallerand, 2000, p. 316). Contrary to the motivational theories that came before (Hull, 1943; Skinner, 1953; Ryan & Deci, 2000, p. 57), SDT regards motivation as a non-unitary phenomenon. While early theories were focusing on what separates a motivated state from an unmotivated one (the motivation-amotivation dichotomy), SDT paints a more intricate picture (Ryan & Deci, 2000).

Ryan and Deci (2000) posit that motivation does not only have a certain level, but also an orientation, based on the subjective relation of the person's motivation and her phenomenological self (Ryan & Connell, 1989). This perceived locus of causality draws a fine line between true intrinsic and extrinsic motivation and explains the basic process of internalisation. This process can be viewed as a gradual change in the perceived locus of causality of extrinsic motivation (Ryan & Connell, 1989, p. 750-751) that leads to more self-determined behaviour.

Extrinsic motivation describes activities that are affected by rewards or pressures (Przybylski, Rigby & Ryan, 2010, p. 155) and are outside the actor's phenomenological self. It moves on a scale from being controlled to being compelled to an activity (Ryan & Deci, 2000, p. 61, Fig. 1). The distinction between the poles of this scale is not trivial. While being controlled is generally considered to diminish positive psychological outcomes, being compelled to act — referred to as internalized motivation — does produce positive outcomes, even though the motivation is not intrinsic (Vallerand, Pelletier & Koestner, 2008, p. 259).

Intrinsic motivation "exists in the nexus between a person and a task" (Ryan and Deci, 2000 p. 56). Because it depends on the psychological needs of competence, autonomy, and relatedness, it is "catalysed rather than caused" (Ryan & Deci, 2000, p. 58). Although intrinsic motivation can be supported by the novelty, challenge or aesthetic value of situations (Ryan & Deci, 2000, pp. 59-60), it is not directly linked to them. Intrinsic motivation is strongly associated with volition, positive psychological outcomes, and wellbeing (Ryan & Deci, 2000; 2017).

The two different kinds of motivational orientation, however, are not additive. A meta-study (Deci, Koestner & Ryan, 1999) showed that any expected "tangible reward, threat, deadline, directive and competition pressure [sic]" can shift the perceived locus of casualty and undermine intrinsic motivation (Ryan & Deci, 2000, p. 59).

Frameworks for the Measurement of Motivation in Gameplay

The Player Experience of Need Satisfaction (PENS) framework (Ryan, Rigby, & Przybylski, 2006) explains how the universally applicable psychological needs — as discussed in the beginning of this section — can be supported by and manifested during videogames. Competence is supported by the proximal and distal goals of the player and her competence and self-efficacy; autonomy manifests in meaningful choices and strategic decisions; and relatedness is realised through interactions with other players and believable computer agents.

The Gaming Motivation Scale (GAMS) (Lafrenière et al., 2012) focuses on how intrinsic- and the different shades of extrinsic motivation can manifest during gameplay. Lafrenière et al. (2012) identify the following concepts – based on the work of Ryan and Deci (2000, p. 61, Fig. 1):

• Intrinsic Motivation: playing for the pleasure of experience
• Integrated Regulation: playing for the enhancement of life, self-image, or values
• Identified Regulation: playing for the development of skills or social bonds
• Introjected Regulation: playing out of compulsion
• External Regulation: playing for rewards or prestige
• Amotivation: playing despite a lack of motivation (or stopping because of it)

We can see — as discussed before — that unlike intrinsic-, extrinsic motivation lies across a wider scale (from integrated to external regulation). This article used GAMS to formulate a part of the a priori codes for its analysis. Intrinsic and — different types of — extrinsic motivation were identified in the participants' accounts using the items of the GAMS questionnaire (Lafrenière et al. 2012, p. 829, Table 1) as cues and label descriptions.

Although SDT, PENS, and GAMS are great frameworks for identifying different types of player motivation, their scope is usually left vague (Deci & Ryan, 2000, pp. 230-231). Some parts of the literature observe games on a macro-level (full sessions or playing in general), while others focus on situationally occurring structures. One explanation for this could be that Deci and Ryan (2002) presents SDT as aa universal theory, which can be applied on both a micro and macro-level.

The Hierarchical Model of Intrinsic and Extrinsic Motivation

Fortunately, the Hierarchical Model of Intrinsic and Extrinsic Motivation (HMIEM) provides a solid structure, building on the basic ideas of SDT (Vallerand & Ratelle, 2002). The model posits that motivation — apart from having a level and orientation — has three levels of generality. These are the global (personality), contextual (life domain), and situational (momentary) levels. The theory explains their connection as a strong top-down influence and a weak "recursive bottom-up relationship" (Vallerand & Ratelle, 2002, p. 51).

The Interpretation of HMIEM in the present study

The presented research significantly scaled down the two lower levels of this model (contextual and situational) to one gameplay session. In their original article, Vallerand and Ratelle only cite "education, leisure and interpersonal relationships" as the most prominent contextual domains (2002, p. 44). Although in the original terminology, the contextual level of videogames would be "leisure" and the situational level would be "gameplay session", this article justifies its interpretation of HMIEM with its more focused scope. Fortunately, Vallerand and Ratelle also suggest that the typology is flexible (2002, p. 46). In this new setup, the contextual level became the "gameplay session" and the situational level became the separate "gameplay scenarios". It was not the focus of the research to determine whether this interpretation is a new application or expansion on the original model of HMIEM, and thus, the article will continue to refer to gameplay session and scenario as contextual and situational levels and frame players' motivations accordingly.

Frustration in Flow Theory

Along with SDT, Flow Theory was chosen as the other fundamental pillar for the research because of its direct link to challenge and action centric emotional states. Though the original theory was born from observations of the creative process (Nakamura & Csikszentmihalyi, 2009, p. 195), its main concept, flow, became very popular in videogame research (Jones, 1998; Sweetser & Wyeth, 2005; Cowley et al. 2008).

Flow Theory focuses on the "optimal experience" of total and effortless engagement that many early research subjects called "the zone" (Csikszentmihalyi, 1997, p. 29). Flow was later described as a feeling of enhanced autonomy, intrinsic joy, and total control over an activity (Nakamura & Csikszentmihalyi, 2002; 2009). Flow emerges during intrinsically interesting challenges that the players are barely capable of overcoming.

It is easy to understand the interest of game research and design in such a phenomenon. Since most videogames are composed of a series of challenges, many researchers applied Flow Theory to build an effective framework to evaluate and guide game design (Jones, 1998, p. 3; Salen & Zimmerman, 2004, p. 347; Sweetser & Wyeth, 2005). Although this approach was successful in creating great frameworks for the evaluation of game design, it generally only focused on the macro-level and fell short in evaluating the players' moment-to-moment experience (Cowley et al. 2008, p. 16). The main reason for this is that flow — in the original theory — is described as a specifically subjective, autotelic, and situationally emerging phenomenon, however, frameworks building on the "eight characteristics of flow" (Nakamura & Csikszentmihalyi, 2002) tend to focus only on the objective (environmental) requirements (p. 90). This helps creating guidelines to identify design elements which can facilitate flow, but tell us less about the players' personal — and often more diverse — experience.

This article does not follow in the footsteps of such studies but focuses on the attentional and informational system that Flow Theory uses to define different emotional states (Nakamura & Csikszentmihalyi, 2002, p. 95, Figure 7.1b). Instead of a tight focus on flow, this has the advantage of being able to describe a wider spectrum of the changes in the players' subjective experience.

Flow Theory speaks in terms of entropic and negentropic psychic states (also labeled as negative and positive emotions) based on the disturbance or aid they bring to the actor's attentional focus (Csikszentmihalyi, 1997, p. 22). Whereas psychic negentropic states are arousal, flow, control, and relaxation, psychic entropy groups together apathy, boredom, and anxiety.

Entropic sates, similarly to flow, are "functions of how attention is being structured at a given time" (Nakamura & Csikszentmihalyi, 2002, p. 92). Psychic entropy in particluar, describes a state when there is "noise in the information processing system" (Cowley et al. 2008, p. 22) that prevents the actor to focus on the activity at hand.

If the information processing system fails to balance the increase in novelty of the stimuli (also known as the external complexity) and the "confirmation" of the environment based on the cognitive model (also known as the internal complexity) (Cowley et al. 2008, p. 20, Fig.10) then the psychic disorder increases. This causes the incongruity (Rauterberg, 1995) of the system to rise to an unmanageable level and players experience the game as either too boring (in the case of negative –) or too difficult (in the case of positive incongruity) (van Lankveld et al. 2008). The psychic entropy caused by either of these situations is experienced through negative emotions, which the players try to escape by avoidance, either by seeking new stimuli (Rauterberg, 1995; van Lankveld et al. 2008) or reaching a negentropic state by restructuring their internal model and resolving the disorder (Nakamura & Csikszentmihalyi, 2002, p. 90; Cowley et al. 2008, p. 22).

In videogames, similarly to Csikszentmihalyi's anxiety (Nakamura & Csikszentmihalyi, 2002, p. 95, Figure 7.1b), players generally experience frustration as a high arousal, negative valence emotion (Mandryk & Atkins, 2007, p. 342, Fig. 15.). Frustration is not always disruptive to the play experience. Initially, it can help players focus by introducing non-specific arousal to the information processing system, which forces them to abandon peripheral objects in their attentional field (Bessiere et al. 2006). However, frustration gradually causes an informational overload, which eventually shifts the focus of attention from the actions at hand to the cognitive processing system, causing "performance dysfunction" (Bessiere et al. 2006, p. 944). We call the former form of frustration adaptive and the latter maladaptive.

By describing players' emotional states based on the function of their attentional system, we see that players experience fun when they are in negentropic psychic states – emotions with positive valence but varying levels of arousal (Mandryk & Atkins, 2007, p. 342, Fig. 15). This does not always mean flow, but a positive experience nevertheless. Rising frustration, on the other hand, increases the noise of the attentional system, drawing resources from the task (Bessiere et al. 2006) and eventually disrupting the negentropic mental state (Cowley et al. 2008, p. 22), which leads to negative feelings, usually signified by high arousal (Mandryk & Atkins, 2007, p. 342, Fig. 15).

The Compatibility of the Attentional and Motivational Systems

Although there were studies comparing these frameworks, they mostly observed how the satisfaction of basic psychological needs (as described by SDT) correlates with the flow (Kowal & Fortier, 1999). The findings were unsurprisingly positive, but they did not delve into finding a shared vocabulary. This section provides a rundown on how the formulation and maintenance of motivational structures (as described by SDT) can be connected to the attentional system behind Flow Theory.

Flow Theory regards every affective state as a function of the attentional system (Nakamura & Csikszentmihalyi, 2002, p. 92). This description makes it easy to describe flow in the terms of information processing (Cowley et al. 2008) as it simply explains motivational structures as "manifestations of psychic negentropy" (Csikszentmihalyi, 1997, p. 22).

SDT "is specifically framed in terms of social and environmental factors that facilitate versus undermine intrinsic motivation" (Ryan & Deci, 2000, p. 58) and expands on the facilitation of basic psychological needs (Ryan & Deci, 2000, pp. 58-60). However, Ryan and Deci (2000) hint towards how the structures of SDT might work in the terms of information processing. They posit that intrinsic motivation can be supported by "novelty, challenge, aesthetic value" in a given situation (Ryan & Deci, 2000, pp. 59-60). Incidentally, these same phenomena — "novelty and complexity" (Cowley et al. 2008, p. 21) — also govern the attentional system. The "environmental factors" of Ryan & Deci (2000, p. 58) can be viewed as properties of the "environmental complexity" (Rauterberg, 1995, p. 56), which, along with "bodily complexity", create the "complexity of the context" (p.56), or "external complexity" in the terms of Cowley et al. (2008, p. 22). This means that the presented informational system can either support or disrupt intrinsic motivation. The "internal complexity" (Cowley et al. 2008, p. 22) then could also refer to not just the mental model (van Lankveld et al. 2008, p. 228), but the motivational structure, because the mental ordering of a situation follows the lines of goals and motivations (Csikszentmihalyi, 1997, p. 22). Since focusing attention by structuring the situation reduces noise and individuals "could process information by (unconsciously) optimizing it according to personal goals" (Cowley et al. 2008, p. 22), this means that engaging in not just intrinsic but also extrinsic motivation could help to resolve psychic entropy (Csikszentmihalyi, 1997, p. 23). This notion is further supported by Csikszentmihalyi, who specifically states that amotivation (and not extrinsic motivation) is what leads to the highest level of psychic entropy (Csikszentmihalyi, 1997, p. 23).

The presented system also works in reverse. Because motivation is a "manifestation(s) of psychic negentropy" (Csikszentmihalyi, 1997, p. 22), underlined by a strong connection between negentropic psychic states, intrinsic motivation and self-determined extrinsic motivation (Kowal and Fortier, 1999), any disruption in the attentional system which leads to the rise of psychic entropy will disrupt the motivational structure as well. As intrinsic motivation is particularly dependent on the subjective mental state (Ryan & Connell, 1989, p. 749), it is also more prone to the negative effects of psychic entropy. In summary, the facilitation of intrinsic and extrinsic motivation can be expressed through the attentional processing system by considering the motivational structures as mediators in the mental ordering of attention as well as the presented information. These structures can lower the noise and internal complexity in an information system, which in turn lowers the psychic entropy (Csikszentmihalyi, 1997; Cowley et al. 2008). However, emerging mental states can also affect the motivational structure. Frustration in particular raises non-specific arousal levels (Bessiere et al. 2006), which takes attentional resources, and disrupts the mental ordering and the motivational structure. Intrinsic motivation is especially sensitive to the entropy in the attentional system because it relies more on the subjective ordering and thus the internal– rather than the external complexity (Nakamura & Csikszentmihalyi, 2002; Cowley et al. 2008). It can be expected that intrinsic motivation provides a stronger mental ordering than extrinsic motivation since it supports highly negentropic psychic states like flow (Nakamura & Csikszentmihalyi, 2009, p. 196). However, extrinsic motivation, even though it can be considered suboptimal, is still a viable strategy to avoid high levels of entropy (in the form of amotivation) when a purely intrinsic structure is not viable (Vallerand, 2000, p. 315).

The second part of the paper will lean on this conceptual framework to explain its results. The presented findings and model are reinforcing this interpretation.

Methodology

The main topic of the research was player persistence during videogame play. More specifically, the study focused on the frustrating parts of inherently enjoyable gameplay. The main research questions were:

• How do players react to frustrating situations arising during gameplay?
• How do players keep themselves motivated during frustrating scenarios?

It was hypothesized that players keep a vague prior interest — or focus of play — intact through the frustrating scenarios. Because videogame play is inherently intrinsically motivated (Lafrenière et al., 2012, p. 827), it was also hypothesized that players' persistence is fuelled through contextual intrinsic motivation and their desire to restore their situationally intrinsic state.

The research was a pilot-like study that used a focus group with three, and semi-structured interviews with six young male participants (nine in total, age M=26), who all identified as gamers. The methodological framework of the study was template analysis (King, 2012). The participants were gathered first with purposive (Bryman, 2012, p. 418), then snowball sampling (Bryman, 2012, p. 424). The first participant's criterion was to find a player that frequently plays frustratingly hard games. To ensure the collection of critical data, subjects were asked to nominate other players who routinely played games which were considered hard or on hard difficulty.

Template analysis has been chosen as the method of research because it provided a flexible framework that was specifically designed to uncover underlying themes and patterns in a sample (King, 2012). The method uses any qualitative or quasi-qualitative data — usually interview transcripts (Brooks & King, 2014, p. 4) — to construct a continuously evolving template of codes (King, 2012, p. 426) that are later interpreted by the researcher (King, 2012, p. 446; Brooks & King, 2014, p. 8).

The interviews had been transcribed and used to construct a template of codes, which developed over time, going through a few iterations before arriving at the final variation. A priori codes were created based on items of the GAMS scale (Lafrenière et al., 2012) and descriptions of flow (Nakamura & Csikszentmihalyi, 2009). These codes helped identifying intrinsic and extrinsic motivational structures, their facilitators, and the shifts between them. During the coding process, these a priori codes were evaluated and grouped together if necessary. The final template contained a hierarchical list of descriptive codes and an unordered list of interpretative, integrative themes. These parallel codes were developed during the coding process and uncovered underlying patterns across several subjects. This dataset provided the basis for later analysis and interpretation, which produced the theoretical model of the observed processes. Though the initial research questions and hypotheses led the investigation, the study had a highly inductive nature due to the chosen methodology, which promotes flexibility in the development of the theme and interview templates (Brooks & King, 2014, p. 7).

Prominent Patterns

The template had an extensive rubric of codes, too long to include or meaningfully discuss here in its entirety. As the codes developed over time, they led the investigation into new avenues as they highlighted prominent patterns across several participants. In this subsection, these integrative themes (King, 2012, p. 432) are presented both as preliminary results and tools for the complete investigation (discussed in the next section).

While some participants actively sought out challenges, others focused on relaxation. The template showed that this focus of play could vary from session to session. Although cross-session play is out of the scope of this article, these changing interests are noted here because they provide the initial source of motivation for playing. Interestingly, while this resulted in some differences in tackling and structuring challenges — even within the same behavioural pattern (notably challenge-seeking players made use of metagames to create new challenges, while immersion seeking players used metagames to acquire new information or rewards, reducing the overall complexity of the game) — when it came to frustration, the adaptation strategies were very similar.

The main identified reasons of frustration could be tracked back to an unordered or noisy informational system caused by either too hard or too easy challenges. Both of these can lead to either mismatched challenge levels, which present the players with an informational entropy, or repetition, which leads to the loss of novelty. Both cases affect the attentional system negatively (Cowley et al. 2008, p. 20) and contribute to rising frustration, which introduces non-specific arousal that further rises the entropy.

Interestingly, depletion of the rewards that the game can provide (in the form of new levels, story elements, challenges, etc.) led the participants to abandon the game, but the perceived loss of psychological need support did not. Though frustrating scenarios generally do not facilitate the basic psychological needs, these depleted games were characterized quite differently, often described as empty or boring, compared to the high arousal state of frustration.

Because the loss of the players' foci of play and not frustration was the main reason of session exits and game abandonment, it was theorized that players' prior interests play a prominent role in the formulation of their contextual and situational motivation. It was also concluded that the mediation of frustration across different player types (challenge or relaxation seeking) was controlled by roughly the same process.

Limitations

The main limitation of the study was the scope and size of the sample, which was relatively small and had little variety (Bryman, 2012, p. 425). To counteract this, purposive and snowball sampling was used consciously to lower the risk of redundant or shallow interviews (Bryman, 2012, pp. 418-419).

The participants' profiles showed vague diversity in terms of classical player profiles (Bartle, 2004; Yee, 2005), which increases the dependability of the study (Bryman, 2012, p. 392) as it shows that the participants — to a certain extent — acted as critical cases (Bryman, 2012, p. 419) for young male gamers. Despite this, subjects were all players who were actively seeking out challenging game experiences. From their accounts (sometimes recalling tens of retirals) it was evident that they had a high threshold for frustration. Admittedly, this limits the results of the study to serious gamers and might not generalise over to a broader and more casual population.

However, the article refers to the study as a "pilot", because of its setup and inductive, exploratory nature; it merely lays down the foundation for further research, but does not lend itself to broad generalization. In the same manner, the result of the study is a theoretical model which needs further testing and development through future research.

Results

The ultimate result of the study is a theoretical model of a process of orientational shifts in situational motivation (Figure 1). This model is the result of the analysis and interpretation of the gathered data, filtered through template analysis. The data gathered from the research sample shows a peculiar pattern of how players mediate frustrating bits of the game across all participants.

Figure 1. Model of orientational shifts in situational motivation during frustrating videogame segments (click to expand).

Although the sample — due to its limited size — cannot be considered fully representative, the very similar way all the different interviewed players chose to tackle rising frustration indicate the possibility of a strong general pattern. This suggests that a motivational model for videogame play can largely benefit from a hierarchical view.

The model describes the players' changing motivation and psychic entropy, and the relation of the two systems. It follows the players as they segment the game into discrete chunks, choose their actions to reinforce their contextual intrinsic motivation through situational segments, face and deal with frustration, and "normalize" the experience by returning to the initial intrinsic situational state.

Segmenting the Game

The study regarded the gameplay session as the contextual level, while separate segments or gameplay scenarios were considered situational (Vallerand & Ratelle, 2002). Players were conscious about the underlying structure of games, often describing experiences in the form of levels, challenges, story elements or memorable parts. For example, Subject 4 described[1] game "parts" based on frustrating and non-frustrating segments:

Only parts become frustrating, not the whole game. There is a frustrating part, like in every game. There is a frustrating part in every game. Like in Neverhood (1996), when you have to cross the room, and it takes forever. (Subject 4)

It is possible that the shift in situational motivation happens especially easily because of this pronounced segmentation and awareness of situational chunks. The experience consciously facilitated by the players, who decide to avoid or push through frustrating segments.

However —in line with HMIEM (Vallerand & Ratelle, 2002) — these frustrating situations are not oozing automatically into the whole (contextual) game experience. Several continuous frustrating episodes must follow each other to make the whole context of a game frustrating, as illustrated by these accounts from Subject 2 and Subject 3:

[Question: What makes a game frustrating?] If you fail, you have to go through multiple times on the same part. […] The worst is when you had to repeat the part before. (Subject 2)

[Playing Skyrim (2011)] I try many times before I quit. […] For example, in one of my first play-throughs, I thought it would be cool to go with non-direct combat and went with necromancy. I ran into an enemy who was literally beating me into the ground. I think I did more than a hundred tries before I gave up. (Subject 3)

Situational Reinforcement

The analysis found that the interviewees played for challenge (which would lead to flow) or relaxation. In the context of Flow Theory, the former is a low challenge scenario, which is based on the intrinsic values of immersion (Nakamura & Csikszentmihalyi, 2002, p. 96). In both cases, players start intrinsically motivated, both contextually and situationally. They decide to go into frustrating situations only after their initial intrinsic motivation is situationally reinforced. Participants initially looked for content which facilitates intrinsic motivation, and only moved forward to frustrating segments, if the game has already established itself as intrinsically rewarding. As Subject 4 and Subject 6 put it:

[Neverhood] built up its image in a way that I know that I could solve it. The game already won me over and I knew that I could solve it. But if it opens with a challenge on this [very hard] level, I probably won't bother to play at all no matter if I can do the first puzzle in the end or not. (Subject 4)

If the first ten [challenges] is good and then comes two bad, you push it through. If the first ten are bad, you leave (Subject 6).

The players' prior interest or focus of gameplay — a vague gravitation towards a desired mental state — also plays a part in determining how they orient within the game world, what kind of challenges or game elements they seek, and what they wish to avoid. The focus of the contextual motivation can change from session to session, as evidenced by these accounts from Subject 3 and Subject 5:

It depends on the mood, interest, game type. Each of these can influence what is more important, story or game play. (Subject 3)

I enjoy the story and I enjoy the gameplay. […] It actually depends on the mood. […] If I haven't played in quite some time, I rather play a challenging game. [… But] if I am in the mood to prefer a story-oriented gameplay it is mainly because I want to relax and enjoy casual gaming and not the challenging part of a game. (Subject 5)

This initial interest created a context for their play session that was not necessarily true for each occasion. What controls this is hard to say based on the current research, but participants' recollection generally direct to mood and vitality (Ryan, Rigby & Przybylski, 2006 p. 352), which in turn suggests that the nature of the contextual focus of gameplay is affected by the global (personality) level motivational state (Vallerand, 2000 p. 314). Unfortunately, the current study had insufficient scope to measure and analyze the global level motivation of the participants.

If the game fails to reinforce the players' contextual intrinsic motivation by giving them situational scenarios early on, in which they can stay interested (intrinsically motivated), the players will most likely abandon the game.

Facing Frustration

Games usually build up tension and challenges over time (Salen & Zimmerman, 2004a, p. 321). This curve is hard to match for each player because personal skills are depending on too many subjective factors (Nakamura & Csikszentmihalyi, 2002, p. 91). Thus, players likely experience frustration eventually in one of their play sessions. The participants' two general strategies to tackle high levels frustration was to either simply avoid them or — more predominantly — to mediate the experience with extrinsic motivation.

It is important to note that few players, however, took a third path to mediate their experience. These players usually had "challenge" as their primary focus and were less likely to find extrinsic motivators when faced with frustrating game segments. Their main strategy instead was to lower the complexity of the system by using outside resources, taking advantage of different game mechanics, and mentally framing frustrating situations as interesting challenges instead. However, they also had a certain threshold, after which they also resorted to the two prominent strategies cited above, as supported by this account from Subject 1, playing Dark Souls II (2014):

I start to look at online wikis and look up data on spells and weapons and thinking about how to advance my character. Sometimes I figure out something nice, so I start the game again and I try the new thing until I get frustrated. Sometimes it doesn't work but I'm still trying for a half hour before quit for good. (Subject 1)

As the rising level of frustration becomes maladaptive and disrupts the initial negentropic state, players cling to their prior interest. However, this focus that previously led them to intrinsically interesting segments, with the loss of that intrinsic state, becomes important to reach. The best example for this change of perspective was given by Subject 4:

If there is a huge challenge that you have to get past and you are in the flow, then [you] get to this challenge and it just breaks the whole thing. And there is the idea that the moment you get past that you'll be put back into the flow and everything will be enjoyable again. (Subject 4)

This signals a shift in the orientation of their situational motivation because while a feeling of interest alludes to intrinsic motivation, a feeling of importance indicates "well-internalized" extrinsic motivation (Deci & Ryan, 2000, p. 230). Every participant expressed this in some form or another, no matter if their focus was to experience challenge (and flow) or just to relax. They did not just want to escape the entropy, but to restore their intrinsically motivated state (and psychic negentropy).

However, the disruption of their intrinsic motivation means that their level of psychic entropy grows higher, causing additional disturbance in their motivational and attentional systems. To avoid amotivation that leads to the highest levels of psychic entropy, players resort to suboptimal structures. Unfortunately, these frustrating scenarios leave their attentional system overloaded and cognitive resources preoccupied. They also limit players' autonomy and do not give enough competence feedback to facilitate intrinsic motivation.

The next best option in this scenario is to resort to extrinsic motivational structures (Vallerand, 2000, p. 315) to escape the highly entropic state. Players thus keep their prior interest but shift their situational motivation from intrinsic to extrinsic to avoid a sudden lack of motivation. They prioritize their actions in a way that the proximal pressure of the newly risen psychic entropy takes priority (Nakamura & Csikszentmihalyi, 2002, p. 90; Cowley et al. 2008, p. 22). This means a clear, well-defined, and reachable goal set by the players themselves. The new structure can aid the focusing of the attentional system and reduce the entropy, helping to resolve the noise in the information processing system. This was a very prominent theme across almost all participants:

[Frustrating parts are] only inspiring in the sense that I want to push through it, so I can go forward. […] It's work. (Subject 2)

Just do it. No matter the cost, just get past. […] I can do it. I must do it. Come on this game cannot get past me like this. (Subject 3)

If I look at a part of a game that is frustrating and I decide to do it, I do it to advance the story. (Subject 4)

Well I consider it as work, as a job to be done. (Subject 5)

If you are not buffing your character up, you cannot complete the game. You have to do some repetitive stuff. [… In Witcher 3 (2015)] I can get a power point [… when] for example climbing a mountain and the view is beautiful. But I do this while grinding my teeth. […] But I need the skill point because it is valuable. (Subject 6)

Interestingly, among many participants, this type of mediation usually involved one of the less integrated types of extrinsic motivation (Ryan & Deci, 2000, pp. 61-62), such as introjection or external regulation. As it can be seen in the quotes above, this is often manifested as a sense of pride or rewards in terms of progression. However, this can be explained by the sudden shift and lack of autonomy support, which does not allow more integrated motivation than introjected regulation (Ryan & Deci, 2000, p. 64).

In summary, players have a certain prior interest that initially supports their intrinsic motivation. Under frustrating segments, this object becomes the focus of their extrinsic motivation, because they start to regard it as a reward if they manage to pull through. Their perceived locus of causality between their phenomenological self (Ryan & Connell, 1989, p. 759) and focus of their motivational structure shifts towards an extrinsic relationship. This can explain why players who have a genuine interest and intrinsic motivation towards the game on the contextual level are experiencing situational parts as "work", "suffering" and "necessary evil" to their gaming experience.

Normalizing the Experience

The last piece of the process is what happens after the frustrating situation has been averted. One might assume that, since extrinsic motivation works against intrinsic in a very impactful way (Deci, Koestner & Ryan, 1999; Przybylski, Rigby & Ryan, 2010), the players' initial motivation suffers. However, participants seemed to get back to their intrinsically motivated state quite easy and saw frustratingly hard segments only as interruptions of their otherwise intrinsically motivated gameplay:

If I have to retry four, five times, and I only get this hard challenge about only every hour, then I can enjoy it. (Subject 2)

[I am] coming back to play a game, again and again because of the flow I get. Pre-set challenges, leaderboards and achievements one of the main things that can drag me back to a game. […] But this is only because of the immersion that I actually want. […] Having a longer, overarching goal is more important for me. Because once there is a larger goal, you can set your own smaller goal on how to reach it. (Subject 3)

When I completed Mass Effect (2007) on "insane" [difficulty], I felt the external goal, I felt that I had to reach my goal when I reached a frustrating part. Other than that, I just played and didn't care about the difficulty. (Subject 6)

The main reason for session exit was not frustration, but the depletion of need support, which happened when the game was not able to upkeep the players' competence and autonomy needs. This was generally characterized by boredom and loss of interest in the activities the game could offer, or, as Subject 2 described it:

Sometimes I find a few things bothersome. Like the crafting and gathering can get boring or if you are not using fast travel, riding around can burn you out a bit. But then I just take a break. (Subject 2)

In contrast, depletion of elements which could potentially be regarded as extrinsic rewards for playing the game were the main reason of total game abandonment. For example:

As I finished the story, I haven't really played much. It wasn't really interesting to me anymore. I was mainly invested in the story so there were no other things to do. (Subject 5)

Although some of these aspects are also facilitators of intrinsic motivation (like the narrative, which can support intrinsic immersion), as evidenced by the previous subsection, these elements were also used by the players to facilitate an extrinsic motivation when faced with frustration.

However, codes that were associated with the causes of frustration were generally not connected to session exits. There were a few instances when unresolved frustration led to "rage-quits" (leaving the game with a highly negative emotion after a major failure), but in general, the main reasons were boredom, fatigue, and loss of perceived competence and/or autonomy support. Or, as Subject 2 put it:

I feel like a zombie. […] I don't feel the game, don't enjoy it anymore. (Subject 2)

These feelings are often associated with amotivation (Ryan & Deci, 2000, p. 61). However, frustrating segments were instead characterized by high arousal (Mandryk & Atkins, 2007, p. 342, Fig. 15.) and extrinsic motivation.

Players — after pushing through the frustrating bits — usually revert to intrinsic motivation in a new situational segment. Over the scope of the whole game, they experienced the frustrating bits only as interruptions but not meaningful disruptions. An explanation for this phenomenon can be the top-down influence of contextual motivation (Vallerand & Ratelle, 2002, p. 51). As the extrinsic motivation is resolved by the players earning their respective rewards through ending the frustrating segment, the intrinsic, contextual motivation creates the new situational frame. The new segment starts as intrinsically motivated because the psychic entropy was lowered at the end of the previous segment and the players received competence and/or autonomy feedback. It can be inferred that this segment is negentropic, as players are not framing them as "the frustrating bit" anymore and have enough need support because the game already proved it could reinforce intrinsic motivation. Moreover, the players' extrinsic motivation can only be resolved in a meaningful way, when they achieve the goal that they set as a reward, which, in this case, is ending the frustration and returning to an intrinsically motivated state.

Another pattern that can support this interpretation is the "Zeigarnik effect," explained by Rigby and Ryan as an innate drive to completion (2011, p. 109). The game might be harder to leave in an unresolved state, thus players who already experienced the positive, psychic negentropy were compelled to play up to a point where the psychic entropy is resolved. However, this interpretation can also be categorized as introjected regulation under extrinsic motivation.

Summary

Summarizing the process that can be inferred from the interpretations of the participants' accounts, the paper arrives at a theoretical model. This model describes a hypothetical scenario of orientational shifts in situational motivation that can account for player persistence through highly frustrating gameplay.

The model (Figure 1) is based on the notion that players segment the game into chunks based on the game innate structure and their own emotional experiences. While the contextual motivation represents their alignment to the whole gameplay session, the situational instances of motivation follow these segments. The shifts in motivation are happening in and between these situational frames.

The described process requires players to arrive into the game with a contextually intrinsic motivation. This motivation is based on an often vague focus of play described as a prior interest in an experience or an orientation towards a mental state. The first step in the process is the establishment of the situational and reinforcement of the contextual motivation. Players seek out elements of the game that are in line with their interest and provide them with enough psychological need support to formulate intrinsic motivation. These segments facilitate focused attention and psychic negentropy and help to reinforce the contextual motivation.

Upon arriving at an inescapable and frustrating game segment, players with reinforced intrinsic motivation will persist to push forward. However, as the level of frustration rises and the information processing system becomes cluttered with noise, the players have a hard time allocating their attention. As the ordering of the attentional and information processing system becomes harder, psychic entropy rises. This disrupts the intrinsic motivational structure, which is ordinarily a "manifestation(s) of psychic negentropy" (Csikszentmihalyi, 1997, p. 22). Because the intrinsic situational motivational structure is gone, psychic entropy rises even further, reaching its high point (Csikszentmihalyi, 1997, p. 23; Cowley et al. 2008, pp. 20-21). This segment is described by anxiety and frustration.

While the situational motivation is disrupted, the contextual motivation stays intrinsic. This can be explained by the negative bottom-up effect of the situational frustration being so weak that it cannot meaningfully disrupt an already reinforced intrinsic contextual motivation. The foci of the players also stay the same, even though their situational motivation has lost its original structure. In this event, where finding another situational frame is not viable because of the preoccupied attentional and cognitive resources, the quickest and easiest solution to overcome the entropy is to establish a suboptimal, extrinsic structure. Thus, players shift the perceived locus of causality of their situational motivation from intrinsic to extrinsic.

Players set their object of interest that led their intrinsic motivation to an exact goal to reach. The newfound motivation helps to prioritize and to structure the attentional and informational system, lowering the level of entropy. This helps to overcome the challenge, which will lower entropy even further.

In the end of this segment, the information processing system is ordered; the entropy is low; and the players reached their goal. On the one hand, this scenario nourishes negentropic psychic states. On the other hand, it leaves the players without a motivational structure once again. This happens because extrinsic motivation diminishes with its rewards achieved and pressures avoided. However, in the ordered system, it is easy for a new intrinsically motivated frame to emerge. This is supported up by the top-down influence of the — still intrinsically motivated — contextual frame, the competence feedback gained from beating the challenge, and the return to an intrinsically interesting segment that facilitates the basic psychological needs.

Of course, the rundown described above is the optimal path for encountering and adapting to frustration. Prolonged or frequently repeated frustrating episodes can give enough bottom-up influence that it meaningfully disrupts the contextual motivation. In other scenarios, it might be viable to change the situational frame, looking for other scenarios and challenges, and gain back the intrinsic motivation that way.

This theoretical model of orientational shifts in situational motivation during videogame play can account for players' behaviour during frustrating gameplay scenarios. It makes use of a hierarchical view of motivation that is beneficial, given the complex nature of videogames. It also utilizes the attentional and information processing system, as seen by Flow Theory, which has the benefit of using a clear vocabulary of beneficial and detrimental mental states. This framework combined with the conceptualisation of the players' experiences, as the allocation of attentional and psychic energies, can benefit not just researchers, but designers alike.

Discussion and Further Research

The study presented through this paper has its limitations. The small sample size and inductive nature of the research make it akin to a prototype project. Nevertheless, the results of the study are promising and point towards new directions. Thus, the model is worth further development and research. The obvious next step is to revise the methodology and duplicate the study on a larger scale with both "hardcore" and "casual" players. Unfortunately, experimental research is very hard to conduct, because such setups can easily interfere with intrinsic motivation (Vallerand & Ratelle, 2002, pp. 45-46). However, analyzing recorded play-throughs and the text of the game could lead to better understanding of how the process of structuring and segmenting the game unfolds.

One of the underdeveloped areas of the project is the process of shifting back to intrinsic motivation. The analysis interprets the accounts of the participants about continuous play with the use of literature to figure out the process. While this approach produces consistent results with the rest of the model, further development is needed.

The current model does not make a difference between negentropic psychic states when it comes to reinforcing intrinsic motivation. Although the current data is not suggesting any large variance, because different affective states have different arousal levels (Mandryk & Atkins, 2007), they could interact differently with the attentional system. Another question regarding reinforcement is to which extent players need reinforcement before persisting through a seriously frustrating challenge without abandoning the game.

The model only uses a modified view on the two lower levels of HMIEM, but the data hints towards possible development in extending the framework to a global level. What this article calls focus on play or prior interest falls very close to the interests in classical player types (Bartle, 2004; Yee, 2005). It is worth investigating how these typologies correlate to contextual and situational goals and global level motivations. Players' disposition towards certain game elements seems to affect their strategies to tackle challenges. Different attitudes towards the game might define different levels of arousal that, in turn, could affect the attentional system.

The presented theoretical model is — even in its infancy — able to account for a part of player behaviour that has yet to be mapped in depth. Both researchers and designers can benefit from this development as the model leads into the direction of better understanding of players' moment-to-moment motivation, frustration, and adaptation strategies. Especially worthy of further development is the role of extrinsic motivation in adaptation strategies. Despite extrinsic motivation being generally sidelined by contemporary research in favour of intrinsic motivation (Ryan, Rigby & Przybylski, p. 348; Rigby & Ryan, 2011, p. 10), it seems that the former can play a substantial positive and supportive role in gameplay. In sub-optimal scenarios, extrinsic motivation can help players stay engaged, resolve the situation, and return to intrinsic motivation. The model also integrates SDT, a very useful motivational framework (which is already being employed in videogame research) with the very popular Flow Theory and its background of attentional and informational processing.

Acknowledgements

Special thanks to Andreas Lindegaard Gregersen, PhD, my former professor at the University of Copenhagen, whose encouragement and advice led me throughout my research.

Endnotes

[1]Interview excerpts were corrected grammatically to help ease of reading; however, efforts have been made to preserve as much meaning from the original transcripts as possible.

References

Bartle, R. A. (2004). Designing virtual worlds. New Riders.

Bessiere, K., Newhagen, J. E., Robinson, J. P., & Shneiderman, B. (2006). A model for computer frustration: The role of instrumental and dispositional factors on incident, session, and post-session frustration and mood.Computers in human behavior, 22(6), 941-961.

Breuer, J., Scharkow, M., & Quandt, T. (2015). Sore losers? A reexamination of the frustration–aggression hypothesis for colocated video game play. Psychology of Popular Media Culture, 4(2), 126.

Brooks, J., & King, N. (2014). Doing Template Analysis: Evaluating an End of Life Care Service. Sage Research Methods Cases.

Bryman, A. (2012). Social research methods. Oxford University Press.

Csikszentmihalyi, M. (1997). Finding flow: The psychology of engagement with everyday life. Basic Books.

Cowley, B., Charles, D., Black, M., & Hickey, R. (2008). Toward an understanding of flow in videogames. Computers in Entertainment (CIE),6(2), 20.

Deci, E. L., & Ryan, R. M. (2000). The" what" and" why" of goal pursuits: Human needs and the self-determination of behavior. Psychological inquiry,11(4), 227-268.

Deci, E. L., & Ryan, R. M. (2002). Handbook of self-determination research. University Rochester Press.

Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological bulletin, 125(6), 627.

Hull, C. (1943). Principles of behavior: an introduction to behavior theory. Oxford, England: Appleton-Century.

Jones, M. G. (1998). Creating Electronic Learning Environments: Games, Flow, and the User Interface.

King, N. (2012). Doing template analysis. Qualitative organizational research: Core methods and current challenges, 426.

Kowal, J., & Fortier, M. S. (1999). Motivational determinants of flow: Contributions from self-determination theory. The Journal of Social Psychology, 139(3), 355-368.

Lafrenière, M. A. K., Verner-Filion, J., & Vallerand, R. J. (2012). Development and validation of the Gaming Motivation Scale (GAMS).Personality and Individual Differences, 53(7), 827-831.

Mandryk, R. L., & Atkins, M. S. (2007). A fuzzy physiological approach for continuously modeling emotion during interaction with play technologies.International journal of human-computer studies, 65(4), 329-347.

Nakamura, J., & Csikszentmihalyi, M. (2002) "The concept of flow." Handbook of positive psychology, 89-105.

Nakamura, J., & Csikszentmihalyi, M. (2009). Flow theory and research. Handbook of positive psychology, 195-206.

Przybylski, A. K., Rigby, C. S., & Ryan, R. M. (2010). A motivational model of videogame engagement. Review of general psychology, 14(2), 154.

Przybylski, A. K., Deci, E. L., Rigby, C. S., & Ryan, R. M. (2014). Competence-impeding electronic games and players' aggressive feelings, thoughts, and behaviors. Journal of personality and social psychology, 106(3), 441.

Rauterberg, M. (1995). About a framework for information and information processing of learning systems. In Information System Concepts (pp. 54-69). Springer US.

Rigby, S. (2004). Player Motivational Analysis: A model for applied research into the motivational dynamics of virtual worlds. motivation research group, University of Rochester, Rochester, NY.

Rigby, S., & Ryan, R. (2007). Rethinking carrots: A new method for measuring what players find most rewarding and motivating about your game. Gamasutra. com. Retrieved December, 3, 2007.

Rigby, S., & Ryan, R. M. (2011). Glued to games: How videogames draw us in and hold us spellbound: How videogames draw us in and hold us spellbound. ABC-CLIO.

Ryan, R. M., & Connell, J. P. (1989). Perceived locus of causality and internalization: examining reasons for acting in two domains. Journal of personality and social psychology, 57(5), 749.

Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary educational psychology, 25(1), 54-67.

Ryan, R. M., & Deci, E. L. (2017). Self-Determination Theory: Basic Psychological Needs in Motivation, Development, and Wellness. Guilford Publications.

Ryan, R. M., Rigby, C. S., & Przybylski, A. (2006). The motivational pull of videogames: A self-determination theory approach. Motivation and emotion,30(4), 344-360.

Salen, K., & Zimmerman, E. (2004). Rules of play: Game design fundamentals. MIT press.

Skinner, B. F. (1953). Science and human behavior. Simon and Schuster.

Sweetser, P., & Wyeth, P. (2005). GameFlow: a model for evaluating player enjoyment in games. Computers in Entertainment (CIE), 3(3), 3-3.

Yee, N. (2005). Motivations of play in MMORPGs. In Proceedings of DiGRA 2005 Conference: Changing Views