This document summarizes a study on the effects of language fluency and subtitles on film comprehension. The study tested 90 participants with varying English language fluency levels who viewed an excerpt from the film North by Northwest under three conditions: original version, intralingual subtitles, and interlingual subtitles. The results showed that subtitles had different effects depending on language fluency. For advanced viewers, both subtitle types reduced comprehension scores. For beginners, interlingual subtitles improved comprehension scores but intralingual subtitles did not. Intermediate viewers were not significantly affected. The study provides new insights into how subtitles differentially impact comprehension based on viewers' language skills.
Effects of language fluency and subtitles on film comprehension ISB7 Utrecht The Netherlands10 july 09
1. Effects ofEffects of Language FluencyLanguage Fluency
andand SubtitlesSubtitles
onon Film ComprehensionFilm Comprehension
Jean-MarcJean-Marc LAVAURLAVAUR & Dominique& Dominique BAIRSTOWBAIRSTOW
Utrecht 8-11 July 2009Utrecht 8-11 July 2009
2. Audiovisual messagesAudiovisual messages
• Overcoming the barriers of languageOvercoming the barriers of language
Evolution of technology and extension of the Internet withEvolution of technology and extension of the Internet with
programs in different languages.programs in different languages.
• ExposureExposure to audiovisual messagesto audiovisual messages
Increasing with number of TV or computer screens in ourIncreasing with number of TV or computer screens in our
everyday environment (home, workplace, towns).everyday environment (home, workplace, towns).
3. Types of informationTypes of information
Contextual information:
- derived from the two previous types and added to the viewer’s
own knowledge.
Non-linguistic Linguistic
Auditory
Music, background sounds,
Dramatic effect
Auditory
Dialogues, song
lyrics
Both
Visual
Images in movement
Visual
Subtitles, signs
4. Multi-modality and multilingualism:Multi-modality and multilingualism:
The case of subtitlingThe case of subtitling
• Multi-modality: transmitting or receiving information on
more than one mode.
• Multilingualism: communicating in more than one
language, be it actively (through speaking, writing and/or
signing) or passively (through listening, reading and/or
perceiving).
• In audiovisual messages: subtitling
5. Types of subtitlesTypes of subtitles
• Interlingual subtitles
language different to that of
the dialogues (generally a
condensed translation)
see picture
• Intralingual subtitles same
language for the dialogues
and the subtitles (gen. a
condensed version of what is
said)
« Call my mother
straight away ».
6. • Other types
Double subtitles (used in the countries
when two languages or dialects are in
current use)
Inverted subtitles (mainly used for second
language learning / subtitles in the non
dominant language)
Types of subtitlesTypes of subtitles
7. PreviousPrevious studiesstudies
• Subtitles lead to an automatic reading
behaviour even when they are not necessary to
the viewer’s comprehension
D’Ydewalle, Praet, Verfaillie & Van Rensbergen,
1991.
• With eye movements technique, automatic
reading leads to a loss of visual information
processing
Koolstra, Peeters & Spinhof, 2002.
8. Previous studiesPrevious studies
• Disturbing effect on global understanding for
bilinguals, against a global facilitating effect for
monolinguals.
Lavaur & Bairstow, 2008.
• Visual input while language processing is
beneficial since it adds to the meaning gained
through the auditory input.
Marian, 2009.
9. Fluency levelsFluency levels
• The previous studies agree:
– Intralingual subtitles distract the viewer’s attention from
the images to the subtitles;
– Interlingual subtitles also distract the viewer’s attention
but can also help monolingual viewers.
• What has not been taken into consideration?
– Language fluency levels must be clearly defined and
range from beginner to advanced learners;
– Relative effects of both types of subtitles must be
measured on a same population to apprehend their
interaction with fluency levels.
10. HypothesesHypotheses
• Different effects depending on subtitles and
fluency levels:
- Original version: the level of comprehension of an
audiovisual message improves with the fluency level
- Intralingual version Subtitles cause distraction and
more so if they are in the viewer’s dominant language.
- Interlingual version Subtitles cause a distraction and
more so when they imply that both of a viewer’s known
languages are on-screen.
11. Preliminary study-MethodPreliminary study-Method
• Participants:
12 secondary French school students (16 to 18 years old)
• Material:
Excerpt of the film “North by Northwest” (A. Hitchcock,
1959), lasting 8’40, in two versions.
• Procedure
(1) Viewing of the silent version, then free recall task.
(2) Viewing of the dubbed version, then free recall task.
(3) Filling in of a short questionnaire.
13. Experimental studyExperimental study
• Method:
- Participants 90 new participants, selected with a
language criterion.
- Material
(1) Questionnaire made up from the recalled
information (1st
part), 42 items concerning either
images or dialogues;
(2) 3 Versions of the same movie;
(3) Word translation task;
(4) Personal data and auto-evaluation questionnaire.
14. Experimental studyExperimental study
• Design:
- Three independent variables:
(1) English fluency level (Low Vs. Intermediary Vs. High);
(2)Version of the film (Original Vs. Intralingual subtitles Vs.
Interlingual subtitles);
(3) Type of information (Visual Vs. Linguistic).
GROUPS Original Intralingual Interlingual
Beginners 10 10 10
Intermediary 10 10 10
Advanced 10 10 10
15. Experimental studyExperimental study
• Procedure:
(1) Word translation task
(2) Personal data questionnaire
(3) Viewing of the film extract
(4) Computerised Film comprehension questionnaire
Each participant was tested individually .
16. Results
• Data analysis design:
(1) Group attribution method validity test ;
(2) Analysis depending on the version of the film;
(3) Analysis depending on the language fluency;
(4) Analysis depending on the type of information.
18. Results
• Global interaction
The global comprehension scores reveal an interaction
between language levels and version of the film.
• Analysis depending on the version of the film
Significant effect of the participants’ fluency level
- Original version: comprehension increases with
language level
- Intralingual: global reduction of gaps between scores
- Interlingual: further reduction of gaps.
19. Results
• Global comprehension score depending on the version of
the film
10
15
20
25
30
35
Original Version Intralingual Interlingual
Version of the film
Globalcomprehensionscore
Beginners
Intermediary
Advanced
20. • Analysis depending on language fluency:
Presence of subtitles affected two of the groups.
- Beginners: global comprehension increases with
interlingual subtitles, not intralingual
- Intermediary: no significant difference of global
comprehension depending on the version
- Advanced: deterioration of global comprehension
scores with both types of subtitles.
Results
21. Results
• Evolution of global comprehension score for each
fluency level depending on version
15
18
21
24
27
30
33
Original Version Intralingual Interlingual
Version of the film
Globalcomprehensionscore
Beginners
Intermediary
Advanced
22. • Analysis depending on the type of information:
- Visual information
Original version Best scores for Beginners and Advanced
participants;
Intralingual version All three groups are very close as
Intermediary level increases;
Interlingual version Scores very close, Beginners group
most affected.
Visual information (results)
23. Visual information (results)
• Visual information processing depending on the
version and the fluency level
4
6
8
10
12
14
16
Original Version Intralingual
Version
Interlingual
Version
Version of the film
Meanscoreforvisual
information
Beginners
Intermediary
Advanced
24. • Analysis depending on the type of information:
- Dialogues
Original version Scores increase with language fluency
levels;
Intralingual version Beginners’ score increases slightly,
Advanced group’s score diminishes;
Interlingual version Beginners’ score increases
significantly, other groups stay level.
Dialogue information (results)
25. Dialogue information (results)
• Dialogue information processing depending on the version
and the fluency level
0
2
4
6
8
10
12
14
16
18
20
Original Version Intralingual
Version
Interlingual
Version
Version of the film
Meanscorefordialogue
information
Beginners
Intermediary
Advanced
26. General discussion
• Aims
(1) Confirming the general effects of subtitles on film
comprehension
(2) Studying the relative effects of different subtitles
(3) Showing the importance of fluency levels
• General effects of subtitles
When necessary subtitles have a positive effect on
global comprehension
When superfluous subtitles have a distracting effect on
global comprehension
27. General discussion
• Expected effects of subtitles on film comprehension:
- Distracting effect when unnecessary for any type of
subtitle and for both types of tested information:
Fluent viewers’ perception, comprehension and memorisation of
visual and dialogue information is reduced in the same manner
when subtitles are on-screen.
- Contradictory effects:
Viewers that only partially need them (in one language or another)
seem either unaffected by subtitles or else in an unmeasured
way.
28. • Expected effects of subtitles on film comprehension:
- Globally facilitating effect when very necessary to
film comprehension:
but concealing partial effects depending on the type of
information:
(1) Disturbing effect for visual information processing
(2) Facilitating effect for dialogue-based information
processing
General discussion
29. • New findings:
- Concerning advanced viewers:
The global distraction provoked by subtitles extends to
both types of subtitles, regardless of the language
used.
Detailed results show that it is the same for dialogue-
based information, but that the effect is uneven when
visual information is analysed.
General discussion
30. • New findings:
- Concerning beginner viewers:
The global comprehension scores increased in the
presence of subtitles.
This effect was found even with both foreign subtitles
and foreign dialogues.
General discussion
31. • Benefit provided for Beginners
Incidental language learning through subtitles?
• The case of “reversed” subtitling
Koolstra & Beentjes, 1999.
• Eye movement technique
Comprehension of what really happens in the
presence of different types of subtitles.
Issues
Good morning, my name is Dominique Bairstow and I have got a PDH position in cognitive psychology at the University of Montpellier 3, in the South of France. I study the influence of subtitles in film perception, comprehension and their effects in language learning. The work I am going to present today, Effects of language Fluency and Subtitles on Film Comprehension, was done in collaboration with Mr. Jean-Marc Lavaur who is an assistant professor in cognitive psychology, also at the University of Montpellier 3.
Audiovisual messages can be of many different types including films, documentaries, news items or even video-clips. The evolution of technology and the extension of the Internet allow a flowing exchange of information across the world, overcoming the barriers of language. What’s more, the TV and computer equipment of homes, but also workplaces and even public places in towns, mean that we are continuously exposed to these audiovisual messages, that can be either translated or not (depending on the language in which they are, but also on the context of presentation).
An audiovisual message puts forth different types of information: On one hand, there is the non-linguistic information that may be gathered through the images, the music, the background sounds or even through the way all these elements are linked to create a dramatic effect; On the other hand, there are the linguistic or verbal elements which can be introduced either on an auditory modality, with dialogues and song lyrics, or on a visual one using written text. Fairly often, the two modalities are used at the same time. ( in the case of subtitles repetition of the same information) A last type that must not be forgotten is contextual information, which is derived from the two previous types added to a certain amount of the viewer’s own knowledge.
Coming back to what we are interested in here, we must first of all link audiovisual messages to the theme, which is multi-modality and multilingualism. Broadly, “Multi-modality” means …(definition), and in the same manner, “Multilingualism” implies… (definition). In the case of audiovisual messages, multimodality takes the shape of subtitles, added to the information already present, and multilingualism refers to the languages used both for the dialogues and for the subtitles.
Subtitling is a technique used in audiovisual translation. This technique can be used to reach various aims, depending on the type of subtitles which is used, thus leading to various language combinations: Interlingual subtitles consist of text written in a different language to that used for the dialogues. They are generally used when the viewer does not sufficiently understand the spoken language. Intralingual subtitles on the other hand are in the same language as the dialogues and can be used for hearing-impaired viewers, language learning or to give additional information bearing on the situation depicted in the audiovisual document.
Other types of subtitles include Double subtitles and inverted subtitles. As their name indicates, double subtitle consist of presenting two sets of subtitles in different languages, generally one at the top and one at the bottom of the screen. Concerning inverted subtitles, the practice is to use the viewers known language in the oral dialogues and a foreign language for subtitling. The use of these two methods is spreading in various fields due to the effects they are found to have. Effectively, the results of recent studies seem to indicate that they could be very useful in certain situation, one of which being the case of foreign language learning. We will come back to this further on.
Previous studies have taken subtitles into account in order to examine the effects they might have on the viewer’s processing of audiovisual information. To begin, D’Ydewalle & Coll. Showed that the presence of subtitles, whether useful or not, leads to an automatic reading behaviour which is not influenced by a habit formed from long-term experience from reading subtitles. Next, Koolstra & Coll. Used the eye movement technique in order to show how the presence of subtitles distracts the viewer, thus leading to a loss of visual information.
In the 2008 Lavaur & Bairstow study, we found that the effects of subtitles depend on the participants: bilingual participants are very much disturbed by unnecessary subtitles, whereas monolinguals are much helped by them, so much that it seems to compensate the negative distracting effect they usually produce. The scores for both visual and linguistic information processing were higher in the presence of subtitles. To finish, Marian has very recently suggested that a visual input while language processing is taking place could be beneficial to the viewer’s understanding, since it gives additional meaning.
All these studies have looked into the effects of various types of subtitling on either visual or linguistic information, but non really took into account different fluency levels. What’s more they all examine the different types of subtitles individually. Considering all these findings, it seems that fluency levels must have an important role in the measured effects and that these effects are different depending on the type of subtitle which is used. This leads us to the present study.
In this study we expected to find different effects of subtitles on film comprehension, depending on the type of subtitle used and on the viewer’s fluency level. When the Original Version in English is shown, the level of comprehension should increase with the viewer’s fluency level, giving the best score for Advanced participants, followed by the Intermediary group and last the Beginners. With the addition of intralingual subtitles (both spoken and written words in English), the gap in between comprehension scores should decrease, due to a distraction of the Advanced group’s attention and a certain amount of help for the Intermediary group and possibly for the Beginners. Finally, when interlingual subtitles are added (spoken language in English and written language in French), we ought to observe a further reduction of the gap in between global comprehension scores, since the Advanced group would be in presence of both their known languages on two different modalities and the Beginners would access linguistic information in their own language. As for the Intermediary group, results should be conflicting due to both the help and the distraction these subtitles would produce.
The participants were recruited in a French secondary school, by asking them to take part in an experiment bearing on film perception and comprehension. No language criteria was necessary, apart from them being native French speakers. An excerpt of A. Hitchcock’s “North by Northwest” was chosen in two versions: a silent and a dubbed version. The participants were asked to watch the silent version attentively after which they were asked to recall as much information as possible on a sheet of paper and using short sentences. Next, they were shown the dubbed version and then asked to recall as much information as possible that is different to that recalled after the first version. To finish, they had to fill in a short questionnaire bearing on their personal characteristics.
After the first part, 24 pages of free-recalled information were obtained (2 per participant). The information was given in the form of short sentences and was used in the making of a questionnaire regarding the film for the second part of the study. This table represents the mean scores and standard deviations of information recall after each viewing. The information given by the participants was divided into 2 categories after the first viewing (visual and other) and into 3 categories after the second viewong (visual, dialogues and other). In both cases, « other » includes erroneous and inference based information. As can be seen, the implicit aim of our study was achieved: the silent version was supposed to induce the recall of visual based information, whereas the presentation of the dubbed version added to the instruction to recall different information, induced the recall of dialogue based information.
For the second part of the study, we selected 90 participants from the same school as previously, using a language criterion in order to create 3 language fluency levels: low, intermediary and high. Using the information recalled from the first part of the experiment, we created a questionnaire bearing on the sequence, concerning either visual or dialogue based information. This questionnaire contained 42 items, each of which was presented with 1 correct answer, 3 wrong answers and one possibility of answering “I don’t know”. Three versions of the same movie extract were used: the original version (English dialogues), an intralingual version (dialogues and subtitles in English) and an interlingual version (English dialogues and French subtitles). The participants saw only one of the versions. Next we created a word translation task containing 15 words to translate from French to English and 15 from English to French. Half of these words were considered to be of medium difficulty and the other half more complicated. This was used to evaluate each participant’s fluency level. To finish, and in order to know the profile of our population, we created a personal data questionnaire containing a linguistic level auto-evaluation task.
This work contains 3 independent variables, creating a 3 by 3 by 2 factorial design. By crossing language fluency and version of the film , 6 experimental groups were obtained (viewing either of the subtitled versions) and 3 control groups (seeing the non-subtitled version), each group containing 10 participants.
In the second part, each participant was asked to go through the word translation task and fill-in the personal data questionnaire with the auto-evaluation task. Following this, they were asked to view the film extract attentively, having been warned beforehand as to which version they would see. The film was shown on a laptop computer screen using the Media Player Classic software. Immediately after seeing the film, they were asked to answer the film comprehension questionnaire, on the same computer and using the Sphynx Lexica software.
The results of the second part of this study were analysed following 4 dimensions: First, we ran a correlation test between the language translation task and the auto-evaluation in order to check whether the group attribution method was valid. Next, data analysis was performed following each of the 3 independent variables: version of the film, language fluency levels and type of information.
In this table we can see the scores each participant obtained in the word translation task (out of a possible maximum of 30) and the score they attributed themselves during the auto evaluation task (out of a possible maximum of 10). The Bravais-Pearson test showed a positive global correlation between the two scores, meaning that the more the participants evaluated their language fluency level highly, the better was their score to the word translation task. Therefore, we can say that the fluency level measured by the translation task was comforted by the subjective evaluation of this level that the participants attributed to themselves. This analysis confirms that the attribution of each participant to a certain group depending on his language fluency level is valid.
To begin, the analysis of global comprehension scores showed an interaction of the language fluency levels with the versions of the film. Next, the partial analysis taking into account the three versions of the film showed a significant effect of the participants’ fluency level. Effectively, when the original version was presented, the comprehension score increased with the viewer’s knowledge of the on-screen languages. When intralingual subtitles are added, the gap between each group’s comprehension score is slightly reduced, as the subtitles disturb when not needed and help a little the less fluent groups. Finally, when interlingual subtitles are added, the gap between global comprehension scores are further reduced as the presence on-screen of both known languages distracts fluent participants but helps non-fluent participants. We will see these effects in more details I the next graph.
As can be seen in this graph, when the original version is presented the highest comprehension score is obtained by the group with the highest level of language fluency, the advanced group, followed by the intermediary group and then the Beginners. When intralingual subtitles are added, the most affected group is the Advanced group, whose global comprehension score diminishes of a faire amount. The two other groups, on the other hand, are hardly affected at all. Finally, when interlingual subtitles are added, the gaps in between each group’s comprehension are further reduced, though the global difference remains significant. As can be seen on this figure, the Beginner and the Intermediary groups cease to differ significantly in the presence of these subtitles.
Concerning language fluency levels, the presence or not of subtitles had a significant effect on two groups out of the three tested. Indeed, while the Beginners’ comprehension level increases with the addition of interlingual subtitles, the Advanced group is very much distracted by them. As opposed to these two groups, the Intermediary participants’ global comprehension doesn’t seem affected by any type of subtitles. These results can be seen in more detail on the next graph.
As we can see in this figure, two groups out of three are affected by the presence of subtitles, since the Intermediary group’s level (in blue) hardly varies. The Beginners comprehension level (in yellow) remains stable when intralingual subtitles are added, but increases significantly when interlingual subtitles, in their mother tongue, are added. As far as the Advanced participants (in red) are concerned, both types of subtitles have an effect. The intralingual subtitles, presenting the same information on the visual modality as on the auditory one, have a disturbing effect on their global comprehension. This disturbing effect gets stronger when both this group’s known languages are on-screen simultaneously, that is to say with interlingual subtitles.
In order to check whether the global comprehension scores were not hiding particular relative effects depending on the type of information, we analysed the processing of visual and dialogue based information depending on each fluency level and on the version of the film presented. Concerning visual information processing, the beginner and advanced group obtained the same comprehension score, whereas the intermediary’s score was much lower. When intralingual subtitles are added, the scores level out somewhat, as the intermediary group’s score increases and the score of the other groups diminishes. Finally, with interlingual subtitles, the visual processing scores are similar for all three groups. The details of these results are shown in the next figure.
When the non-subtitled is shown, the Intermediary group got the lowest score for visual information processing, whereas the two other groups obtained equal scores. The addition of intralingual subtitles strongly diminishes the Beginner and Advanced groups’ scores, whereas it increases the Intermediary group’s visual information processing. When interlingual subtitles are added, the beginners processing is the most affected, although all three groups seem to be distracted by these subtitles.
Concerning dialogue information processing, the results followed a general pattern when the original version was shown, with the scores getting better as the language fluency level gets higher. When intralingual subtitles are added, the Beginners’ score increases slightly whereas the advanced group’s score diminishes. With interlingual subtitles, the Beginners’ dialogue comprehension increases significantly, while the other two groups don’t seem affected either way.
With the non-subtitled version, the dialogue processing scores follow the same pattern as that of the global comprehension scores, meaning the lowest score for the Beginners, followed by the Intermediary group and finally the Advanced participants who get the best comprehension of the dialogues. When English (intralingual) subtitles are added, the advanced group’s score diminishes while the Beginners dialogue comprehension increases and the Intermediary’s stays level. The Interlingual (French) subtitles prove to be of greatest help to the beginners, whose score becomes higher than the Intermediary’s one, while the Advanced group’s understanding stays the same as with the other kind of subtitles.
This study set out to investigate the effects of 2 different types of subtitles on film comprehension depending on the viewer’s fluency level and on the type of information contained in the film. Our first aim was to confirm the general effect of subtitles, depending on their degree of necessity. Next, by presenting both types of subtitles in the same study and to the same type of population, we wanted to measure the relative effects of these subtitles on audiovisual information processing. The last aim was to show the importance of language fluency levels in film comprehension. The general results of this study confirmed the previous findings concerning subtitles and film comprehension: when subtitles are necessary, they increase the viewer’s global comprehension and more so when they’re in the viewer’s own language. On the other hand, when they are not needed because the viewer understands the language, then subtitles in general have a disturbing effect on global comprehension.
Our first aim was achieved, since the global results of this study confirm what was expected. To begin, when subtitles are not needed because the viewer fully understands the language of the film, the presence of subtitles could not have any other effect than interfering in their attentional processes, thus affecting both types of tested information (visual and dialogue). Next, as far as intermediary leveled viewers are concerned, it seems that their comprehension level stays more or less the same throughout the study. This effect was also expected due to the contradictory effects subtitles must have on their understanding of a film.
To finish with these first results, subtitles were a great help to those viewers who had an extremely poor knowledge of the language used in the dialogues, that is English. This help they provided for global comprehension was however concealing a partial effect depending on the type of information tested: that is to say that while they helped in dialogue based information comprehension, they also produced a disturbing effect on visual information processing, due to the viewer’s attention being attracted by the subtitles area.
Following this confirmation of the general role of subtitles, the results wielded some newer and more interesting results. First and concerning the most fluent participants, the global results showed the expected distraction effect of subtitles, but they also put forward the fact that this effect happened whatever language was used in the subtitles, as if it little mattered. Further analysis put to light the partial effects that the global results were covering: each type of information processing was in fact affected in the same measure, but not in the same manner. Both visual and dialogue information processing lost more or less 3 points in quality between the original version and the interlingual one. But whereas the dialogue information lost these points as soon as subtitles were added to the sequence, the visual information processing lost them more gradually, being slightly affected by intralingual subtitles and then more affected by interlingual subtitles. This interesting finding means that unnecessary subtitles affect both types of information processing in a same manner when they use both the viewer’s known languages simultaneously; but they affect visual information in a smaller amount when only one known language is on-screen on two modalities. One could conclude therefore that crossing multimodality with multilingualism requires more attention to achieve a same level of comprehension as far as linguistic information is concerned, but that the cost of this is a lower attention for other information such as image input.
The next important finding of this study was the effect of subtitles on the comprehension of viewers with a very poor level of fluency. Obviously, we had predicted that these viewers would be helped by the presence of subtitles in their own language since they could not possibly have understood the dialogues. But what interests us here is the fact that these viewers were not only helped by the interlingual subtitles, but also by the intralingual ones. This finding is important because it implies that even if presented in a foreign language, subtitles can help in extracting information from an audiovisual document, both visual and linguistic.
The first set of results enable us to better understand how the presence of subtitles can influence film comprehension, depending on the fluency level of the population to which it is intended. The last results were most important, especially when considering the benefit subtitles provide for the group with the lowest fluency level. This leads us to think subtitles may produce an incidental learning of the language of the film, provided these viewers dispose of a minimal knowledge of the language beforehand. This idea is backed by the fact that even intralingual subtitles seem to help in their understanding of the film, implying that these viewers not only pay attention to these foreign subtitles but also manage to extract a little meaning from them. The theme of language acquisition is much developed in the field of audiovisual studies, as researchers such as Koolstra and Beentjes have already began to show the benefits of reversed subtitling. This method consists of showing a film with the viewer’s mother tongue presented orally in the dialogues and the secondary language in the subtitles. Their findings indicate that this situation enables the viewer to find satisfactory links between lexical elements of both languages. This method should therefore be looked into more deeply in order to be better understood and applied in all fields of language learning. To finish, applying the eye movement technique to studies taking into account all these different types of subtitles would help to discover more precisely what happens when each type of subtitle is on-screen, depending on the viewer’s fluency, and therefore enable a deeper understanding of their relative effects on each viewer’s attention processes.