The Impact of Italian Dubbing on Viewers’ Immersive Experience: An Audience Reception Study

This paper analyses the impact of Italian dubbing on the immersive psychological experience of the American TV series Game of Thrones, using data from post-hoc self-report questionnaires. The show is filled with historical and cultural references to medieval England and the characters are portrayed linguistically through different accents, regional dialects and registers. The research takes the form of an audience reception study. Two audience samples (Italian and English native speakers) watched an excerpt from the ninth episode of season two. English viewers watched the video in the original version, while the Italian sample watched it in the dubbed version. Participants then completed a 40-item questionnaire to self-report their level of immersion, using post-hoc Likert scales to evaluate their presence, perceived realism, character identification, enjoyment and transportation to the narrative world. The results showed significant differences between the immersion levels experienced by the two samples. The English audience reported higher scores on each immersion scale, especially with regard to perceived realism, character identification and transportation, while presence seemed to be the least affected by language transfer.


INTRODUCTION
In recent years, there has been an unprecedented boom in reception studies in the field of audiovisual translation (AVT). Reception studies explore the forms and modes of consumption and reception of audiovisual texts, thus going "beyond descriptivism" (Di Giovanni and Gambier, 2018, p. x). A combination of experimental methods is used to examine the various cognitive aspects that influence the reception of translated audiovisual texts (Díaz Cintas and Szarkowska, 2020, p. 4). An important concept in the context of AVT reception is the phenomenon of psychological immersion, that is "the sensation of viewers of being transported into the story world, and of being swept up in this world to such an extent that it takes priority over their experience of their immediate surroundings" (Wilken and Kruger, 2016, p. 258). The term is "used more or less interchangeably with concepts such as presence, involvement, and engagement" (Nilsson et al., 2016, p. 109). In fact, this notion also comprises additional offline measures, such as the concepts of transportation and character identification (Tal-Or & Cohen, 2010;Green et al., 2004), presence, flow and enjoyment (Wissmath et al., 2009;Wissmath & Weibel, 2012), as well as perceived realism (Cho et al., 2014).
In Audiovisual Translation Studies, scholars have used immersion as a measure to rate the impact of subtitles on an audience's ability to become immersed in the story world of film Kruger, Doherty, & Soto-Sanfiel, 2017;Kruger, Doherty, Fox, & De Lissa, 2018;Kruger, Soto-Sanfiel, Doherty, & Ibrahim, 2016). Immersion has also been tested to analyse the reception of audiodescribed audiovisual texts (Romero Fresco & Fryer, 2013;Walczak & Fryer, 2017) and audiovisual products with subtitles for the deaf or hard-of-hearing (Romero Fresco & Fryer, 2016). More recently, levels of immersion have been assessed to study the impact of voice-over (Flis et al., 2020). Although dubbing is the preferred form of access F. Raffi to foreign language audiovisual content for millions of viewers, this translation modality has mainly been studied in conjunction with subtitling to analyse the impact of subtitling vs. dubbing on viewers' immersion in audiovisual products (Wissmath et al., 2009;Wissmath & Weibel, 2012).
However, dubbing remains an under-researched area with regard to reception and, more specifically, immersion, with a few notable exceptions. Ameri, Khoshsaligheh, and Khazaee (2018) examined the reception of Persian dubbing among Iranian lay viewers using focus group interviews and a questionnaire. The results obtained from the mixed methods indicated that technicality was the most highly valued dimension for rating the quality of dubbing. Similarly,  findings suggested that the Iranian audiences surveyed considered translation quality to be less important than technical issues related to dubbing production, such as voice-acting and synchronisation. In one of the first attempts to apply eye tracking technology to dubbing, Di Giovanni and Romero Fresco (2019) conducted an experiment to compare the eye movements of viewers of original English and dubbed version clips from the film The Grand Budapest Hotel (2014). The study revealed a 'dubbing effect' whereby viewers avoid looking at the actors' mouths, thus allowing them to suspend disbelief (Coleridge, 1817(Coleridge, /1985 and be transported into the fictional world. Romero Fresco (2020) took a step further by focusing on three groups of participants: native Spanish viewers watching a clip dubbed into Spanish; native English participants watching the same clip in English; and Spanish viewers watching an original (and comparable) clip in Spanish. The study confirmed that dubbing viewers do indeed perform unconscious eye movement strategies in order to become immersed in the audiovisual product. To further explore the reception (including enjoyment) of Iranian dubbing viewers, Ameri and Khoshsaligheh (2020) employed netnography (based on ethnographic research) to analyse the posts and comments of an online community. The study showed that Persian dubbing suffers from a range of issues, which has led to a decrease in the number of people enjoying dubbed audiovisual products in Iran.
Given that the analysis of psychological immersion can provide a reliable indication of the impact of audiovisual translation on viewer reception  and that dubbing, the main AVT modality for fiction content in Italy, remains largely underinvestigated, this paper aims to analyse the psychological immersion experienced by Italian-and English-speaking audiences when viewing the American TV series Game of Thrones in its original and dubbed versions, to explore the impact of dubbing using data from posthoc self-report questionnaires (i.e. offline measures), which have been found to be valid, reliable, and nonintrusive (Paas et al., 1994). The two audience samples were presented with a video selected from the ninth episode of season two. All participants of the Italian sample watched the clip in its dubbed version, while English-speaking participants watched the original version. After viewing the extract, participants completed a 40-item questionnaire 1 to self-report their level of immersion, using post-hoc Likert scales to evaluate their presence, perceived realism, character identification, enjoyment and transportation to the narrative world. Data were processed through a quantitative analysis in order to obtain statistical mean values to quantify the immersion levels of both samples of participants. A complementary translation analysis of the two versions of the video (i.e. original and Italian dubbed version) serves as a supplementary tool to interpret the data, specifically focusing on elements related to register, dialect, and regional accent. Considering the linguistic and cultural differences between the two versions and the translation choices made in the passage to Italian, the main hypothesis of this study is that dubbing may decrease audience immersion in the fictional world.
In the following section, the notion of immersion and its main dimensions (presence, perceived realism, character identification, enjoyment and transportation) as measures to rate audience engagement are presented. Then, the case study is illustrated in detail before illustrating the principle results stemming from the questionnaires. The translational analysis of the video extract is used as a tool to help interpret the findings. Finally, the results are discussed and some concluding remarks are drawn.

IMMERSION: MEASURING AUDIENCE ENGAGEMENT
When we watch a film or a TV product, our attention is captured in such a way that we often lose awareness of the environment around us. We want to be transported, to immerse ourselves for a short time in a different reality, to relax and be entertained (Lothe, 2000). From psychological and cognitive points of view, there is great similarity between the ways in which we process the world generated by an audiovisual product and the real world. The audiovisual product is constructed in such a way as to favour the spectator's cognitive involvement. Carroll and Seeley (2013), in their theoretical model of attentional engines, argue that the scenes that make up a narrative sequence act as recognitional prompts, i.e. recognitive stimuli "sufficiently structured to enable a perceiver to identify its depictive content by matching its formal / compositional content to declarative knowledge of the structure and function of target objects and event types" (p. 58). In other words, the sequences in an audiovisual product encourage involvement and understanding because the narrative scheme itself provides the viewer with the information necessary to interpret those sequences. This process works because the human mind tends to select and concentrate only on those aspects of the narrative that make cognitive contact possible.
According to Carroll and Seeley (2013), cinema simply needs to provide the audience with sufficient details to recognise the narrative intent and model their emotional, perceptive, and semantic involvement by inserting characters and events within that narrative context. Clearly, this does not mean that viewers experience the audiovisual dimension as if it were real, but only as a simulation of reality. Viewers are always aware that what they are watching is not real, despite remaining sensorially involved, but this involvement creates a cognitive connection between real and virtual, between the everyday world in which we live and the imaginary universe of audiovisual products.
Immersion is thus a sensitive cognitive state, which can be stimulated or inhibited by external factors, such as the physical context in which the viewing experience takes place (at home or at the cinema) or its social context (alone, with friends, or with strangers) (Kruger, Doherty, & Soto-Sanfiel, 2017). Self-assessment scales relating to five specific dimensions (presence, perceived realism, enjoyment, character identification, and transportation) are commonly used to rate the level of immersion, as discussed below.

The Five Dimensions of Immersion
As extensively demonstrated by previous studies (Kruger, Soto-Sanfiel, Doherty, & Ibrahim, 2016;Kruger, Doherty, & Soto-Sanfiel, 2017;Kruger, Doherty, Fox, & De Lissa, 2018), the concept of immersion is mainly associated with the principles of 'transportation', 'character identification' (Green et al., 2004), 'presence', 'flow' and 'perceived realism' (Wissmath et al., 2009). Presence or 'telepresence' (Minsky, 1980) can be defined as the feeling of being 'inside' the audiovisual context itself (Wissmath et al., 2009), or the impression of having been transported into the virtual context, as though awareness of one's surrounding environment falls away. There is a sort of self-projection into the audiovisual reality (Wirth et al., 2007). As Kim and Biocca (1997) explain, presence is a process made up of two phases: first, the departure, or the process of 'vanishing' from one's physical environment; and then the arrival, when viewers descend into the media environment.
While presence is concerned with the medium, 'transportation' is mainly related to the content and can be defined as the experience of cognitive, affective and imaginary involvement in a narration (Green et al., 2004). For Wissmath et al. (2009), the concept of transportation matches that of suspension of disbelief. A fundamental pre-condition for being able to experience transportation is identification with the characters in terms of adopting their thoughts, goals, emotions, and behaviours. To achieve this, viewers must abandon their physical and social reality and embrace the world created by the audiovisual product (Green et al., 2004).
Another dimension related to immersion is 'perceived realism' (Green et al., 2004), that is the question of how much the world represented in the narrative reflects the real world in the viewer's perception when experiencing a given narrative context (Gerbner & Gross, 1976). However, since the world represented in audiovisual products is multidimensional, the evaluation of perceived realism cannot be unidimensional (Busselle & Greenberg, 2000;Hall, 2003;Pouliot & Cowen, 2007). In particular, Hall (2003) and then Cho Shen and Wilson (2014) identified five sub-dimensions of perceived realism: 'plausibility', 'typicality', 'factuality', 'narrative consistency', and 'perceptual quality'. While plausibility refers to the idea of fictional events and behaviours seeming possible to believe, typicality refers to the extent to which the fiction comes close to the real experiences, past and present, of viewers. Therefore, an event can be plausible without being typical, if it does not fall within the spectator's range of experiences. Factuality aims to identify whether a given narrative represents specific events or characters from the real world. Factuality differs from typicality in that a given F. Raffi narrative may represent factual events which, however, would hardly be evaluated as typical by viewers. Narrative consistency defines the degree of congruence and linearity of narrative elements; a narrative may be coherent (or not) regardless of the plausibility, typicality, and factuality of the events represented. Finally, perceptual quality measures how far the sound and visual elements of a given audiovisual product help to create a convincing and persuasive representation of reality, regardless of whether or not this is reflected in viewers' own experiences (Hall, 2003).
For Tal-Or and Cohen (2010), another layer of immersion is 'character identification', a multidimensional concept which consists of cognitive empathy, emotional empathy and the ability to experience an imaginary reality from the perspective of fictional characters. This process of emotional rapprochement towards a given character can arouse a strong emotional response among viewers (feelings of joy, sadness, tension, and anger that can manifest physically, for instance the reaction of crying). Thus, the perception of an audiovisual product could radically change depending on the type of emotional relationship established between the viewer and a given character (Cohen, 2001;Livingstone, 1998).
'Enjoyment' differs from the above-mentioned dimensions in that it is not an independent variable. However, it is closely linked to transportation and character identification, which are considered to be the factors that most influence the degree of enjoyment of an audiovisual product (Tal-Or & Cohen, 2010). On the other hand, presence and perceived realism may influence the level of immersion experienced by the viewer without necessarily making the experience pleasant and enjoyable.
In what follows, the case study is presented. The main objective is to compare the psychological immersion experienced by English-and Italian-speaking audiences when viewing the American TV series Game of Thrones in its original and dubbed versions, respectively. Particular consideration is given to the linguistic and cultural differences between the two versions and the translation choices made in the passage to Italian.

STUDY DESIGN
A convenience sampling method was adopted for the distribution of questionnaires. Participants were not chosen randomly but constituted an opportunity sample of 140 regular viewers of the TV series: 70 for each sub-group (Italian and English as their first languages, respectively) 2 . The average age was 26 years for the English-speaking group and 24 years for the Italians 3 . All participants came from one of three main fields of study (1) humanities, (2) arts, and (3) medical sciences. Therefore, the two samples were fairly homogeneous and comparable.
The questionnaire was created in two language versions (Italian and English) using LimeSurvey (https://www.limesurvey.org), a free and open source on-line survey application, accompanied by a cover letter presenting the study and providing simple instructions 4 .
The questionnaire consisted of 48 questions grouped into two macro sections: demographic information and viewing experience. The first part, proposed before watching the video, was composed of eight closeended questions with a list of optional answers to choose from; if none were applicable, participants were free to add further options. This macro section aimed to collect information on participants' habits related to the consumption of TV series, in general, and Game of Thrones, in particular, and, most importantly, which language version they usually opt for 5 : original-language version, subtitled with original audio, or dubbed version. As for viewing habits, both samples showed a preference for watching the series on Netflix (68% of the English sample and 64% of the Italian sample) and more than 80% of respondents had seen the whole show at least once (86% of the English sample and 81% of the Italian sample). Unsurprisingly, the English sample mainly opted for the original version (90%) while the Italian sample favoured the dubbed one (64%) 6 .
After watching the video, participants were asked to rate their viewing experience. This second part of the questionnaire contained 40 items (immersion questions) on seven-point Likert scales with answers scaled from 'Completely disagree' to 'Completely agree' and from 'Never' to 'Always' 7 , as presented in the following sections.
Before moving to the analysis of the viewing experience and discussion of the results, the audiovisual stimulus is described next with the aim of disclosing which translation solutions were adopted to render the dialects, regional and foreign accents and different registers employedby certain characters. The results of this analysis will be used to interpret the data obtained from the macro section of the questionnaire, in particular with regard to the Italian sample, and to investigate whether potential translation problems surrounding linguistic and socio-cultural elements result in lower immersion scores among Italian viewers.

Language Variety in Game of Thrones
A video excerpt 8 from the ninth episode of the second season of Game of Thrones was embedded into the online questionnaire. The episode, titled 'Blackwater' in the original version and 'L'assedio' in the Italian version, recounts the final events of the second season, namely the succession war that has broken out in Westeros following the death of King Robert Baratheon. The selected video, lasting 23 minutes, covers the first half of the episode: it starts with the preparations of the siege, shown from the viewpoint of various characters, and ends with the first part of the battle. A striking feature of the episode, and particularly the selected excerpt, is the presence of several British dialects, as well as marked exotic and foreign accents. Furthermore, different registers come into play, thanks to the heterogeneity of characters, from the eloquence of royal family members to the formality and deference of the King's staff and the colloquialism of the soldiers. The variety of linguistic registers, dialects and accents are key elements of the entire series, providing realism, authenticity and clear characterisation while moving the narrative forward. In this episode, the events preceding the war are set in different social contexts. We move from the royal rooms of the Red Fortress to a tavern frequented by soldiers, commoners and prostitutes, thus confirming the tendency of Game of Thrones to offer a very broad narrative and a meticulous and complex 'social picture' of the Seven Kingdoms.
In the opening scene, Captain Sir Davos Seaworth and his son Matthos discuss the imminent attack on King's Landing. Sir Davos speaks with an authentic 'working-class' accent and adopts a low register, in contrast to Matthos's more correct and formal style of speaking: clear indicators of different social backgrounds. However, the Italian dubbed version does not provide any type of linguistic characterisation to reflect this difference. Dialects and accents are language-specific, and therein lies the difficulty of bringing them into the receiving culture. One strategy could be to replace one dialect or accent with another domestic 'equivalent'. However, this type of translation can only result in a loss or a relocation of characters' identities (Heiss, 2004).
In the next scene, the exotic accent of the character Shae is also lost in the Italian dubbing. In this case, the character speaks with a markedly non-British, foreign accent, influenced by the Turkish origins of the actress who plays the role of Shae (Sibel Kekilli). Would it have been helpful to reproduce an accent of this kind in the dubbed version (i.e. the stereotypical portrayal of a Turk speaking in Italian)? Arguably, this would have helped to characterise Shae and reinforce the meaning and coherence of certain scenes, especially those in which Shae's accent is precisely the main focus (See Table 1).
Queen Cersei has just met Shae for the first time and immediately picks up on her accent, thus leading her to wonder how a foreign woman, without the slightest experience of court life, could have come to live at court. Cersei discovers that Shae is actually her brother's secret lover. Since in the Italian version Shae does not use any type of foreign accent, both the Queen and the viewers lose this linguistic clue to Shae's identity. This further demonstrates that linguistic elements are not only used in the series to reinforce recognition of characters, but also to create key narrative dynamics that are difficult to replicate in the dubbed version.
In the scene that follows, the Italian version respects the degree of formality and elegance of the dialogue in the original English version. Queen Cersei and Grand Maester Pycelle speak with received pronunciation and use a high register that could be expected of such wise and powerful figures. Then, when the narrative switches to a tavern full of soldiers and prostitutes, the register changes accordingly. Lord Bronn, a mercenary soldier at the service of Tyrion Lannister and commander of the city guard, talks to Armeca, a prostitute.
In the original, the woman speaks with a marked foreign accent and uses a low register. She also makes grammatical mistakes (e.g. 'How many times you break it?', my emphasis). Meanwhile, Bronn speaks with a strong Yorkshire dialect, which is particularly distinctive, and also uses a low register; these linguistic elements help to create coherence in character identity: in the case of Bronn, a sarcastic man with a black sense of humour and a pragmatic, amoral philosophy on life. In the dubbed version, both the prostitute and Bronn speak in standard Italian, without any kind of accent. The register is also higher than in the original, as shown in Table 2. Firstly, the grammatical inaccuracy of Armeca is not reproduced in the dubbed version, since the expression 'Quante volte te lo sei rotto?' represents a correct way of speaking. Secondly, the choice to translate Bronn's line 'Got in a scrap' as 'Litigai' raises the level of formality of the dialogue: the verb 'litigare' (i.e. 'to quarrel') does not reflect the idiomatic flavour of the English expression. Furthermore, a remote past tense has been chosen, which is not typically used in colloquial Italian but reserved mainly for formal and written language.
The following dialogue, one of the last exchanges of conversation before the battle begins, takes place between King Joffrey and Queen Sansa (See Table 3  Armeca: E mi piace il tuo naso. Quante volte te lo sei rotto?
Armeca: And I like your nose. How many times have you broken it?
Bronn: The second time I was nine years old. I quarreled with some guys who were older than me. I got beaten.
The two characters here, both members of ancient and powerful noble families, speak in an impeccable Southern English accent and use a very high register. As with the previously mentioned conversation between Queen Cersei and Grand Maester Pycelle, when the register is high, and in the absence of dialectal inflections and exotic accents, the translated dialogue generally stays closer to the original. In the first part of the dialogue, Joffrey even talks about himself using the third person ('Your king rides forth to battle', 'You should see him off with a kiss') and the Italian adaptation successfully reproduces Joffrey's linguistic choice. However, in the second part of the dialogue, Sansa addresses the king as 'Your Grace' in the original (showing deference), but in the Italian version no corresponding term is employed. She even uses the second person form 'tu' ('Perdonami, hai ragione', 'sarai nell'avanguardia'), which is commonly used in informal situations between people of the same age or social rank. Therefore, the Italian version does not faithfully reflect these markers of respect, even if it maintains the high register.
To conclude, what emerges from the comparative analysis of the original and dubbed versions of the video is that the translation of linguistic elements tends to be less problematic when characters use a high register without any particular dialectal inflections or exotic accents. However, as previously discussed, the rich variety of dialects, exotic accents and registers that characterise Game of Thrones is precisely what helps to shape the characters' identities. Failure to mirror this could have a negative impact on the degree of immersion reported by the Italian sample.

ANALYSIS AND RESULTS
In this section, the results provided by the two samples on the five perceptive dimensions of Immersion will be discussed. Once collected, answers were converted into numerical values (e.g. 1 for 'Completely disagree' or 'Never'; 6 for 'Completely agree' or 'Always'). Subsequently, these values were used to calculate a weighted average for each group of questions regarding a specific dimension of Immersion. The weighted average obtained for each section is the first reliable indicator for the evaluation and comparison of the different levels of immersion experienced by the two audience samples.

Presence
Regarding the evaluation of presence (Section A), seven items adapted from Kim and Biocca's (1997) model were used. Table 4 shows the average scores. Table 4, although the average score for the English sample is higher than that of the Italian group, the difference is not significant (4.25 against 3.86). In general, there are no significant differences in terms of presence experienced by the two audience samples across all seven items proposed. As discussed later, the average score for presence is the lowest of the five dimensions for the English sample. Perhaps this is due to the abstract nature of the very concept of presence and the difficulty of assessing perception in relation to the questions posed, which may have led participants to provide more neutral answers. Sansa: Forgive me, you're right, I'm a fool. Of course you will be in the vanguard. I know Robb always goes where the fight is toughest. And he's just an impostor.

As shown in
Joffrey: Robb's turn will also come. And you will also lick his blood on the Hearteater.

Perceived Realism
Perceived realism (sections B and C) included 19 items adapted from the models developed by Green and Brock (2000), Tal-Or and Cohen (2010), and Cho et al. (2014) and divided into five sub-sections (plausibility, typicality, factuality, narrative consistency, and perceptual quality) (See Table 5).
The English sample reports consistently higher scores in relation to almost all the items proposed in this section of the questionnaire. Italian participants provided a slightly higher score for one item only, i.e. item 5 of the typicality section. The data are presented in Table 6, which reports the general averages of the five subdimensions and the overall average of the entire section of perceived realism. 5. The video portrayed a typology of events that might have happened to a lot of people. 3.28 3.00 6. Events similar to those portrayed in the video might have taken place in the real world.
2.84 4.24 Factuality 7. The events portrayed in the video are based on or inspired by real facts.
3.28 3.76 8. The typology of events shown in the video reminded me of a specific cultural and geographical context. 15. The audio elements (voices, sounds, music) of the video were realistic and consistent. 4.72 5.12 16. The acting in the video was consistent and convincing.
5.16 5.20 17. The accents and dialects reproduced by the actors were realistic and authentic.
4.00 4.88 18. The costumes, scenography and settings in the video were realistic and convincing. 5.12 5.28 19. I felt that the overall production elements of the video were realistic and convincing and helped to make the portion of story consistent and enjoyable. 4.84 5.28 As anticipated, the weighted average of the English sample is significantly higher than that of the Italian sample (+0.48). Indeed, significant differences emerge across all sub-dimensions of perceived realism, in particular in relation to plausibility, in which the average of the English sample is almost one point higher (+0.9).

Character Identification
'Character identification' (Section E) was rated using 4 items adapted from Tal-Or and Cohen's (2010) model. Participants reported the averages in Table 7.
The English sample reports significantly higher averages in relation to all items, +0.57 with respect to the Italian sample. This seems to further support the research hypothesis that English viewers feel more involved in the show and manage to establish a greater level of empathy and identification with the characters than the Italian audience, thanks to the greater authenticity provided by the original version of the TV show (i.e. not dubbed).

Enjoyment
The dimension of enjoyment (Section D) was analysed using 4 items (Tal-Or and Cohen 2010) and the two audience samples reported in Table 8.
The English sample scored higher averages on all four items. These differences are indeed significant, since in three out of four items the discrepancy between samples is greater than 0.50. It is also interesting to note that, for both the Italian and English groups, the average score for enjoyment is the highest of all five dimensions of Immersion.

Transportation
Transportation (Section F) was rated using 6 items adapted from Tal-Or and Cohen's (2010) model. The data provided by the two audience samples in relation to the last dimension are presented in Table 9. The English sample once again reports significantly higher scores than the Italian sample in relation to all of the items proposed. In comparison to the other dimensions, the averages of both samples are significantly lower, as with presence (see section 4.1). Additionally, the average reported by the Italian sample is the lowest of the five dimensions of Immersion.

Results: Immersion
The data presented above are summarised in Table 10.
An independent t-test confirmed that the levels of immersion recorded by the two samples reached statistical significance: t(8)=2.57, p<0.03. More specifically, the English sample reported a higher score than the Italian group as regards the dimension of presence. In general, the scores obtained were low compared to the other dimensions: 4.25 for the English sample (i.e. the lowest score of the five dimensions) and 3.86 for the Italian sample (the second lowest result, after transportation). As for perceived realism, the British sample reported significantly higher scores on all sub-dimensions with respect to Italian viewers. This can be explained by the fact that the five areas of perceived realism are strongly linked to the linguistic and cultural dimensions of the series. In terms of character identification, the averages for the English sample show significantly higher results than those reported by Italian participants. The reasons for this difference may be found in the process of translation from English to Italian. The largest gap between the two samples was reported for the category of enjoyment. This dimension takes into account a wide range of aspects of the audiovisual product, encompassing a general assessment of what the viewer has seen. Considering that the English sample obtained higher scores for all previous subsets of immersion, it is not surprising that enjoyment followed suit: the result, surely, of a more nuanced understanding and therefore deeper appreciation of the fictional reality. In terms of transportation, considering the gap between the two samples, it is conceivable that the perception of emotional involvement is somehow influenced by the linguistic transfer and that dubbing, somehow, negatively impacts on the level of transportation for Italian viewers. In the following section, these findings are further discussed and conclusions are drawn.

DISCUSSION AND CONCLUSION
Significant differences emerged in terms of the emotional and sensorial involvement experienced between the English and Italian viewers after watching the excerpt from Game of Thrones in its original and dubbed versions, respectively. The data collected from the questionnaire demonstrate that the Italian sample reported significantly lower scores on all five scales of immersion compared to the English audience. The only dimension that seems to be less affected is presence, with a gap of around 0.39; the other four dimensions showed gaps close to or greater than 0.50 which, on a six-point scale, constitute a rather significant difference. These results show that the English sample managed to reach higher levels of suspension of disbelief, transportation (stemming from emotional involvement) and identification with the characters, as well as greater perceived realism than the Italian audience.
It has been argued that the linguistic adaptation for the Italian version is one of the main factors to motivate these differences. As discussed in section 3.1., Game of Thrones is driven by peculiar historical and cultural dynamics as well as a richness of linguistic features. The plot is inspired by events that really happened in England in the Middle Ages, and the characters present a complex linguistic heterogeneity, unlike most television and cinema products that typically adopt standardised British English. Each character is given a precise linguistic identity, with the use of accents and regional dialects. If, on the one hand, these characteristics allow viewers of the original version to reach a high level of immersion, the perception changes significantly when the show is adapted to another linguistic and cultural context. Indeed, a consistent loss of linguistic typicality and socio-cultural elements has been detected in the Italian adaptation (see section 3.1.). This could have negatively affected the perceived authenticity and realism from the perspective of the Italian audience, as well as the viewers' appreciation of the plot, quality of acting, and production. Although dubbing can be effective in its primary function of linguistic transfer, in this case, it flattens out the target text or, in Whitman-Linsen's (1992) words, imposes a sort of "linguistic whitewashing" that results in "devitalized" (p. 118) dialogues. The linguistic and social differences between characters is diluted.
In addition to the issues related to language transfer, it should be taken into consideration that the Italian audience has a completely different cultural background. Moreover, factors such as familiarity with the television series and viewing frequency in the original language (see section 3) are also decisive factors which seem to have influenced the level of immersion experienced. If presence is defined as the feeling of 'being there', inside the world generated by the audiovisual product, this dimension may well be influenced by viewing frequency. The study shows that the level of presence increases for those participants who are familiar with the TV show (+0.53 and + 1.15 for the Italian and English samples respectively). Character identification is also strongly influenced by familiarity with the show (+0.93 and +1.5 for the Italian and English samples respectively). Familiarity is influential, too, in terms of transportation and enjoyment. In relation to transportation, the gap between the 'familiar and 'less-familiar' viewers, in both samples, is greater than 1 (+1.36 for the English group). Enjoyment reports the biggest difference in relation to the variable of familiarity with the show, with both samples reporting a gap of more than 2 between those who regularly watch Game of Thrones and those who seldom watch it.
With regard to the Italian sample, the study has also shown that the preferred language version for viewing Game of Thrones plays a fundamental role. Participants accustomed to watching the TV show in its original English version reached lower levels of immersion when they watched the excerpt in its dubbed version compared to those participants who were already accustomed to the Italian adaptation: -0.38 for perceived realism, -0.37 for character identification, and -0.18 for transportation. For this group of Italian viewers, dubbing seems to limit rather than facilitate the viewing process.
Additional situational variables may have influenced the level of immersion. The study aimed to recreate the emotional and sensory experience of watching a television series; but immersion, of course, is greater when the experience occurs 'naturally' rather than when viewers are asked to purposefully watch an audiovisual product under specific conditions (i.e. to complete a survey). Hence, the findings should not be considered representative, but the results may generate hypotheses for future studies. Continued research could incorporate eye tracking technology, which has recently been used for similar purposes (Orrego-Carmona, 2016;Di Giovanni & Romero Fresco, 2019). Nevertheless, considering the scarcity of studies on the impact of dubbing on immersion, especially with the English-Italian language pair, this study takes a step towards further developing this underinvestigated research area with a specific spotlight on the global phenomenon of Game of Thrones.