Can video game subtitling shape player satisfaction?

ABSTRACT

This paper focuses on video game translation from the vantage point of player experience. In our study player experience is operationalised with the use of the Game User Experience Satisfaction Scale (GUESS-18) which comprises nine experiential sub-scales. Drawing on an experiment with control and treatment groups, this article examines the (in)stability of player satisfaction by involving two groups who played the game Distraint: Deluxe Edition – either with or without typographic spelling errors deployed in the Polish translation of the in-game texts. The participants remotely installed and executed the game program on their local computers, then filled in our online questionnaire on their experience with the title. We find statistically significant differences between the two conditions in terms of Usability/Playability and Personal Gratification. Importantly, typos were also found to impact the composite score of gamer Satisfaction. Given these results, we conclude that user satisfaction is relatively malleable and can be shaped by aspects of translation. At the same time, a caveat is that we have not found evidence to indicate that typos specifically shape most of Satisfaction’s sub-constructs – like Audio and Visual Aesthetics, Play Engrossment, or Enjoyment – when considered separately.

KEYWORDS:

• Player experience
• spelling errors
• GUESS
• reception studies
• translation
• cognition

Acknowledgements

We thank the developer of Distraint, Jesse Makkonen, for his consent to use the game in our study. More information about the developer can be found under JesseMakkonen.com. We moreover thank Dr Aleksandra Matysiak and Dr Michał Kornacki for helping us recruit participants.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 https://partnerhelp.netflixstudios.com/hc/en-us/articles/215758617-Timed-Text-Style-Guide-General-Requirements

2 Following the minimal system requirements (Makkonen, 2017) – CPU: Dual Core 2.0 GHz; GPU: 512 MB card capable of shader 3.0; Memory: 2 GB RAM; Microsoft Windows OS and DirectX 9.0c. The no-direct-contact protocol of the experiment was dictated by the COVID-19 pandemic. However, subjects were effective in transitioning from the online survey to the self-installed game and back to the digital survey, perhaps due to the experimental design geared towards approachable instructions. For more information on our remote experimentation see Deckert & Hejduk (2022, p. 5).

3 Individual participant supervision was not feasible given the anonymity and remote protocols of the study, as well as its scale.

4 The two playable versions were identical besides the 25 experimental typos. We have proofread the official Polish translation of Distraint’s English subtitles before inserting the experimental typos. The original translation is attributed to a game-polonisation website Ikskoks.pl (for more information refer to Deckert & Hejduk, 2022, p. 17).

5 This seems to have been the case because respondents’ comments from both groups indicated certain semantic issues with the translation (cf. Deckert & Hejduk, 2022, p. 15). Since both groups indicated those, we do not consider this an extraneous variable.

6 While similar intuitive measurements have seen use in reception research already (e.g. Di Giovanni, 2018, p. 34), we are aware that it was also possible to rely on standard language competence scales, such as the Common European Framework of Reference for Languages (CEFR). Given that our experiment was executed remotely and that not all of our participants would have been similarly acquainted with language ability description standards, we chose to prioritise understandability of the question, while keeping the number of questions to a minimum, which still served the purpose of the study.

7 In this study, we utilise the Mann-Whitney-Wilcoxon instead of Student’s t test following de Winter & Dodou’s (2010, p. 5) findings into comparing two independent samples of Likert data. They found that ‘the power differences between the t test and MWW were minor’, but MWW was superior for skewed distributions and inferior for multimodal ones. In our datasets we detected no cases of multimodal distributions, only of skewed left and right, or potentially symmetric distributions.

8 See also the point in Section 4 about the results for Social Connectivity being consistently and justifiably lower across conditions than for other constructs.

9 This is one of two open-ended questions in our survey. The other one read ‘If you would like to share any thoughts with us directly, you can type them below’. Only some answers were relevant to the study, but 37 of those were collected in total.

10 The reason for this was that we were unsure while designing the experiment if typos can affect subtitle comprehensibility (for findings cf. Deckert & Hejduk, 2022, p. 12). The game’s tutorial needed to be comprehensible since subjects were not supervised.

11 This opinion likely rests on the respondent’s knowledge that, before character encoding was more standardised, localisation teams would have to change game text typefaces into local alternatives (sometimes quite dissimilar to the original ones) so that translated dialogues could (e.g.) display Polish diacritics. Distraint uses the same font in both language versions.

12 What we mean is that, upon registering the possibility of formal errors in the subtitles, players might have chosen to report some identification in case there were errors to which they had paid no attention, which too implies that they had instead allocated cognitive resources to other aspects of gameplay, facilitating the idea that it was e.g. engaging or challenging.

Additional information

Notes on contributors

Mikołaj Deckert

Mikołaj Deckert is associate professor at the University of Lodz, Institute of English Studies, in Poland. His research is primarily in interlingual translation, AVT/MA, and more broadly in language and cognition. He serves as peer-review editor for the Journal of Specialised Translation (JoSTrans), has recently co-edited The Palgrave Handbook of Audiovisual Translation and Media Accessibility (Palgrave Macmillan, 2020, with Łukasz Bogucki) and co-authored On-Screen Language in Video Games: A Translation Perspective (Cambridge University Press, 2022, with Krzysztof Hejduk).

Krzysztof Hejduk

Krzysztof Hejduk is a doctoral student at the University of Lodz. Awarded by the Polish Minister of Education & Science, he is now involved in several research projects, including one endowed by the National Science Centre of Poland. He is member of the Polish Cognitive Linguistics Association (PCLA) and has worked for the Common Language Resources and Technology Infrastructure projects (CLARIN). He presented the results of his research into game localisation at international conferences and has recently co-authored a Cambridge Element monograph.