Why tweet when you can bubble ? Students ’ perceptions of a voice microblog for the development of their L 2 spoken production skills

Hiroshima Bunkyo Women’s University renauddavies29@gmail.com This study explores the role voice microblogging-based activities play in developing learners’ spoken production skills. The study involved 33 low-intermediate level university students majoring in English at a private university in Japan. The voice microblogging application used in the study was Bubbly, a cross-platform social networking application that allows users to share short 90-second voice messages with fellow users. Findings, based on an exploratory factor analysis coupled with group interviews, suggest that tasks using Bubbly are beneficial for students’ oral performance as they help to provide extra practice for speaking. Furthermore, results clearly show that the application’s social and technological affordances help to raise students’ level of motivation, engagement and confidence. It can be concluded that voice microblogging tasks along with Bubbly’s various affordances offer several benefits to the blended oral communication classroom.


Regular Paper
Why tweet when you can bubble?Students' perceptions of a voice microblog for the development of their L2 spoken production skills

Renaud jonathan Davies
Hiroshima Bunkyo Women 's University renauddavies29@gmail.comThis study explores the role voice microblogging-based activities play in developing learners' spoken production skills.The study involved 33 low-intermediate level university students majoring in English at a private university in Japan.The voice microblogging application used in the study was Bubbly, a cross-platform social networking application that allows users to share short 90-second voice messages with fellow users.Findings, based on an exploratory factor analysis coupled with group interviews, suggest that tasks using Bubbly are beneficial for students' oral performance as they help to provide extra practice for speaking.Furthermore, results clearly show that the application's social and technological affordances help to raise students' level of motivation, engagement and confidence.It can be concluded that voice microblogging tasks along with Bubbly's various affordances offer several benefits to the blended oral communication classroom.
Keywords: Voice microblog, Social Net working, Oral Communication, English as a Foreign Language introduction The number of social networking system (sns) users continues to increase signifi cantly (Brenner, 2013) along with a growing body of research concerning the application of snss in language teaching.Numerous studies reveal that social networking tech nology has the potential to enhance lan guage acquisition and increase students' motivation and engagement (Khoshnoud, 2014;Hsu, 2013;Kabilan, 2010).In particu lar, microblogging tools such as Twitter are gaining global attention (Bennett, 2012).As a result of this social networking revolution, an increasing number of educators are endeavoring to incorporate Twitter into their courses and research into the benefits of utilizing Twitter to enhance language learning is ever growing (Lakarnchua, et al 2014;Hattem, 2012;Kim, E., et al. 2011;Lomicka & Lord, 2012).In the case of Japan, Twitter has been gaining tremendous attention.According to comScore, an Internet analytics company, Twitter has more reach than any other social network in Japan, including Facebook (Singh, 2013).
However, as Twitter is limited to only communication through text, it is not an ideal tool for helping students develop their second language (l2) oral proficiency.Bubbly, another free microblogging application, may be considered an excellent alternative to Twitter as it enables users to not only post short messages of 140 characters, but also leave short 90second voice comments or bubbles.Unsurprisingly, Bubbly is often described as "Twitter with a Voice."The findings in this study reveal that tasks using Bubbly are beneficial for students' oral performance as they help to provide extra practice for pronunciation, fluency, intonation, and accuracy.Furthermore, technological affordances such as the ability to add sound effects, background music and voice filters to a voice comment help to increase both the students' level of engagement and confidence when speaking online.To the author's knowledge, there have been no investigations of the role voice microbloggingbased activi ties play in developing learners' oral proficiency.In this paper the author aims to address this gap by providing an account of how voice microblogs were integrated as a pedagogical tool in the EFL classroom and reports on the students' feedback and perceptions of the use of this social medium to enhance their l2 spoken production skills.

The microblog
This paper will be examining the use of a voice microblog to enhance learning and engage students through increased social interaction and connection.Thus, it is important to first define what constitutes a microblog and what makes it different from a regular blog.A blog, also known as a weblog, is a regularly updated online journal, which enables a user to post their thoughts and opinions and often allows visitors to leave a comment.Ebner and Schiefner (2008) argue that the main difference between blogs and microblogs is speed.Microblogs tend to be used to write about users' thoughts and to give very short reflections.These brief messages are then shared within the users' chosen community.For example, Twitter is a social network in which the members of a community share their current activity by answering the question, "What are you doing?" Users are restricted to typing messages called "tweets" that are capped at 140 characters and spaces.It comes as no surprise that this shorter and faster form of blogging is highly conducive to social network ing.Benefits of using Twitter in language education include the ability to foster a positive social presence (Dunlap, et al., 2009), help encourage collaboration, feedback and concise writing (Mork, 2009) as well as support informal Learning (KassensNoor, 2012;Aspden, & Thorpe, 2009).However, it is important to note that microblogging tools such as Twitter are not designed for conversations, but rather, are tools which provide users with a forum to display updates to their status (Borau et al., 2009).Therefore, a common concern among language educators who wish to utilize Twitter in the language classroom is the required brevity of tweets, which may result in shallow exchanges and bad grammar usage (Grosseck & Holotescu, 2008).Lastly, as Twitter is limited to only communication through text, it is not an ideal tool for helping students develop their l2 oral proficiency.

Asynchronous voice communication
Asynchronous voice communication tools can be an excellent way to help students to improve their oral communication skills outside of class.As is the case with microblogging, the integration of asynchronous voice communication activities into language teaching practice is a relatively new area of study and one that warrants exploration.Research has shown that voicebased asynchronous technologies can increase l2 motivation and col laboration as well as help to create an anxietyfree environment where learners have more confidence when speaking and can communicate at their own pace (Yaneske & Oates, 2010;Sun, 2009;Hsu, et al., 2008;McIntosh, et al., 2003).A survey carried out in June and July, 2012 on the use of digital voice tools among language teachers in uk higher education insti tutions (heis) revealed improved oral skills and language learning strategies, increased learner confidence and motivation, and enhanced feedback opportunities (Sadoux, 2013).Finally, voice has also been shown to be more effective than text in nurturing trust and cooperation (Jensen, Farnham, Drucker, & Kollock, 2000) as well as communicating expres sive meaning and nuance (Doran, 2010;Ice et al., 2007;Kraut, Galegher, Fish, & Chalfonte, 1992).However, it is important to note that research has shown some challenges to using asynchronous voice communication tools in language teaching.For example, a study by Gleason and Suvorov (2012) revealed that participants were unable to edit their recordings once they were posted and therefore, could not correct their errors.In other studies, partici pants reported feeling embarrassed to record their voices for others to hear (Marriott, 2002;McIntosh, et al., 2003;Yaneske & Oates, 2010).We will return to these shortcomings in the next section of the paper when we discuss the benefits of voice microblogging with Bubbly.

Microblogging with voice: Bubbly
Thus far, this paper has looked at microblogging with text as well as online asynchronous voice communication and their advantages and disadvantages when utilized in language teaching.It is now time to merge these two forms of online communication and examine the benefits of asynchronous voice communication within a social networking environ ment.This brings us to, Bubbly, a free microblogging application that enables users to leave short 90second voice comments, or rather, bubbles.As stated earlier, the required brevity of messages in Twitter can be problematic.However, Bubbly's 90second voice comment ing feature allows for longer messages with more depth.A 90second voice comment can have the equivalent of anywhere from four to ten tweets, depending on speaking speed.It is important to note that Bubbly has many of the same features of Twitter.For example, Bubbly allows users to post short messages of 140 characters.Bubbly also supports the use of hashtags.Hashtags begin with a "#" symbol and are used to indicate the topic of a bubble.For example, "#springvacation" or "#bestfriend".As with Twitter, microbloggers on Bubbly are also able to forward bubbles to other microbloggers using mention, which is in the format of "@username".Bubbly users may also easily add links to videos, pictures and web pages, which can be seen within the application eliminating the need to exit the app.Like Twitter, Bubbly is also crossplatform, allowing students to download it to their iPads, iPhones or Android phones.This is paramount as it helps students to stay connected more easily once they are out of range of campus WiFi.Bubbly's strong similarity to Twitter helped to dramatically drop the learning curve for the students in this study as the majority of the students were already avid Twitter users, which is of no surprise as Japanese is the most tweeted language after English (Wauters, 2010).
Another noteworthy feature of Bubbly is the ability to send private voice messages.This is integral as it allows the teacher to give private oral feedback to students as well as have students ask their teacher questions oneonone.Not to mention, users are able to listen to their posts as many times as they like before making the post public and may delete a voice post at any time.Bubbly sets itself apart from other asynchronous voice communi cation tools such as VoiceThread and Wimba Voice in that it enables users to add sound effects, background music and voice filters all of which allow for increased engagement and confidence when speaking online.As pointed out earlier, one of the big disadvantages of asynchronous voicebased communication is that it can be demotivating for students who are embarrassed to record and post their voice comments for others to hear.As will be seen in the results of this study, Bubbly's technological features can help to lower a voice blogger's speaking anxiety and consequently, help to increase confidence as well as make speaking online a motivational and engaging experience for language learners.

Research questions
Research into the use of online voice communication tools to enhance various aspects of language teaching and learning is still quite scarce.In particular, research in the area of online voice communication to enhance l2 spoken production skills has failed to examine the effects of asynchronous voice communication tools within a social networking envi ronment and the role technological affordances such as sound effects, voice filters and soundtracks play in raising students' level of motivation, engagement and confidence when speaking online.Accordingly, the aim of this research was to identify which forces positively influence students' perceptions of voice microblogging as an effective tool for the develop ment of their l2 spoken production skills and further, to investigate the extent to which social and technological affordances help to raise student motivation, engagement and confidence.With the above in mind, the following three research questions were posited: 1.What are students' perceptions of Bubbly's effectiveness as an asynchronous social net working oral communication tool for the development of their l2 spoken production skills?2. What forces positively influence students' perceptions of voice microblogging as an effective tool for the development of their l2 spoken production skills?
3. What features of tasks on Bubbly help to raise students' motivation, engagement and confidence when speaking online?

Participants
The study involved 33 students majoring in English at a private women's university.The speaking proficiency of the students was lowintermediate.The students were part of a second year general English communication class that was split between two teachers, one teacher being the author.There were 17 students in one class and 16 students in the other.All students were between the ages of 19 and 20 years old.

Selecting the Bubbly application
In choosing an appropriate application for the research project, the researcher utilized Puentedura's (2012) samr model, which supports teachers as they design, develop, and integrate online activities into their curriculum.The model, shown in Figure 1, enables educators to classify technology used in teaching depending upon whether it substitutes, augments, modifies, or redefines a task.Substitution and augmentation refer to technol ogy that enhances the learning, and modification and redefinition refer to technology that transforms the learning, in other words, tasks that would be impossible without the technology.Research suggests that students are most engaged in activities considered to be in the redefinition and modification categories of the samr model (Bloemsma, 2013).
Therefore, educators should aim to integrate technology that will help to transform learn ing, not just enhance it.Consequently, as the Bubbly application allows for students to communicate asynchronously within a social network with various technological affor dances such as voice filters, background music and sound effects, the author considered the mobile tool to have the potential to redefine language learning tasks while engaging students outside of class.

Introducing the Bubbly application
As for introducing the application, the teachers combined the aforementioned classes for a single 90minute period and dedicated the time to assisting students with downloading the Bubbly application, creating an account with a simple profile containing no personal information and then making sure to have students follow all other Bubbly participants in the class, which resulted in each student gaining 32 followers by the end of the session.
Students were then shown the basics of the application and were introduced to the first Bubbly task, which is discussed in detail in the next section.All students were equipped with an iPad mini, as the university started issuing iPad minis to all faculty members and students in the previous year.Students who did not have WiFi access at home were encour aged to download the app to their smartphones as well.As stated earlier, for most students, a very small learning curve was involved since the majority of students were already quite familiar with Twitter, which has a similar user interface with many of the same functions.
In order to keep the class social network private, students were asked to delete all automati cally added followers and to block any outside followers who may accidentally discover the students' profiles.Students were also asked to only follow students involved in the project.This was an important step as Bubbly does not offer the option to create a private group.

Task design
The main purpose of the Bubbly project was to motivate students to improve their spo ken production skills through engaging speaking activities outside of class.Accordingly, tasks were designed with "flow", a theory of motivation, at the very bedrock.Flow may be described as a mental state in which a person performing an activity is completely absorbed in a feeling of energized focus, involvement, and enjoyment that leads to improved per formance on a task (Csikszentmihalyi, 1990).As online technology advances it is becom ing increasingly clear that online interactive tools can enhance input and facilitate flow (Tractinsky, 2013).Hoffman and Novak (1996) and Chen (2006) have aptly stated that quali ties of online technology such as control, immediate feedback and interactivity help to evoke flow experiences.Furthermore, research shows that online social networking services can facilitate flow (Barker et al., 2013;Mauri et al., 2011).
According to Egbert (2003), "flow and language acquisition occur under many of the same conditions" (p.506) and "findings suggest that teachers can theoretically facilitate the flow experience for students by developing tasks that might lead to flow" (p.513).Thus, Egbert's (2003) model of flow in language acquisition was used as a reference when attempting to develop engaging and motivating tasks to be used on Bubbly.Figure 2 is an adaption of Egbert's (2003) model.In this model, social and technological affordances have been added to better suit its purpose as a framework for developing activities to be used with the Bubbly application.Consequently, the model shows how a welldesigned task, along with learners' skills in the target language (tl) coupled with social and technological affordances help to engage learners and facilitate flow, which in turn, can result in students' improved performance followed by changes in the learners' skills.According to Egbert, flow houses four important dimensions: a balance between task challenge and student ability (shown in figure 3), a need for focused attention on the task, an absorbed interest in the task and the material, and lastly, a sense of control when doing the task.As Bubbly allows users to control content, receive instant feedback, and interact asynchronously within a social network encompassing a myriad of technological affordances, the author considered Bubbly an appropriate tool to be used with the above model.The author would like to clarify that this study did not examine participants' flow experiences and that this model was used for the sole purpose of designing engaging and motivating tasks.(Schell, 2008, p. 118) The Bubbly voice project contained four tasks, which were designed using the above model and with the goal of building on content learned in the course textbook, World Link 1 by Cengage (Stempleski, 2011).The tasks also served as a way for students to practice for the final speaking tests at the end of each semester.Tasks 1, 3, and 4 appeared in similar form on their speaking tests.The first task served as the final project for the 1st semester and was worth 5% of their semester one grade.The other three tasks were done during second semester and counted as 10% of their final grade.For each task, the two teachers created examples on Bubbly, which the students could refer to at any time.The tasks were as follows: Task one: My Hero (5%).Students were asked to post a 60 to 90second voice bubble intro ducing a person they admire followed by three separate text bubbles describing the person's background, appearance, and why they chose the person.The post header had to contain a picture of the person and when possible, a link to a related video clip or online article.Students were then asked to read and listen to their fellow classmates posts and leave a voice comment on at least three different posts, which was also a part of their grade.For this assignment, the teachers took screenshots of each post, printed them off, graded them and gave them back to students with written feedback.
Task two: My Summer (2%).Each teacher posted a 90second bubble describing what they did over summer vacation.Students were then asked to describe what they did over summer vacation and to comment on at least three other students' posts.This task was designed to give students a little more freedom to talk about and comment on things of immediate interest to them.
Task three: Where am I? (4%).For this third activity, each teacher posted a link to a map on their respective Bubbly blogs and added a voice bubble giving directions to a mystery location on the digital map.Using the teachers' examples, students were then required to post their own voice bubbles giving directions to their own mystery place on the same map.Students were then asked to listen to their classmates' directions and guess the mystery destination.Students did this latter part of the activity in class where they were given a simple worksheet for recording how many times a student had to listen and guess until they guessed their mystery destination correctly.Students were then asked to go back and correct their verbal directions by writing a clearer version on paper based on their class mates' and teacher's feedback.
Task four: What's wrong?(4%).For this final speaking assignment, students worked in pairs to roleplay a phone conversation in Bubbly.The conversation content requirements were to have an invitation, a refusal with a health excuse, and lastly, advice for the health problem.Both teachers recorded an example conversation in their respective Bubbly blogs to serve as an example.Students then worked in pairs to create their own unique conversa tions using vocabulary studied in class and from the World Link textbook.For this fourth speaking task, the teachers gave personal voice feedback via Bubbly to each pair of students.

Survey
The researcher developed a survey that consisted of a 24item, sixpoint Likert scale (see Appendix), which was completed online during class time.The survey questions were in Japanese to make sure students fully understood the content of the questionnaire.It included three sections: (a) students' perceived gains in spoken production skills (6 items), (b) motivation and confidence (9 items) and (c), affordances (9 items).In an attempt to increase measurement precision and avoid a middle category, which can cause statistical problems, a sixpoint likert scale was chosen in accordance with Nemoto and Beglar's (2014) guidelines for developing a Likertscale questionnaire.Participants were asked to rate their agreement on the scale (1 = strongly disagree, 2 = disagree, 3 = slightly disagree, 4 = slightly agree, 5 = agree, 6 = strongly agree).At the end of the survey students were given the option to join a group interview.A total of 10 students volunteered for the interviews, which were held within two weeks of completing the survey.

Procedure and analysis
To answer research questions 1 and 2, the researcher ran a factor analysis on the 24 vari ables contained in subcategories (a) through (c).To strengthen reliability, the following two items were excluded from further analysis: 1. Listening to my classmates' voice comments is interesting.

Bubbly helped me practice class content.
A KaiserMeyerOklin (kmo) measure of the sample adequacy and a Bartlett's test of sphe ricity validated the fitness of the data for factor analysis (fa), performed based on a factor loading of 0.5 or higher and an eigenvalue greater than 1.For these data, Bartlett's test is highly significant (p < .001)and therefore, fa is appropriate (Field, 2005).Due to the small sample size, the kmo value was 0.5, which increased to 0.6 upon removal of the aforemen tioned two items.According to Hutcheson G., & Sofroniou N. (1999), a kmo value of 0.6 is considered a mediocre level result.However, it should be noted that each of the four factors extracted had high clean loadings with all but one loading being greater than 0.6.Thus, the data may be considered reliable despite the small sample size (Guadagnoli & Velicer, 1988).

Results: a quantitative and qualitative interpretation
A principal components analysis was conducted using a promax rotation procedure to see how the 22 items grouped together.Four factors were extracted with an item loading greater than 0.5 as the criterion of importance.The factors accounted for 73.7% of the total variance.Table 1 shows the factor loadings as well as descriptive statistics listing the mean and standard deviation for each item.

Student's perceived gains in spoken production skills
As indicated in Table 1, Factor 1 received strong loadings from the 6 items with 38% of the variance.All items were directly related to students' perceived gains in spoken production skills.Among the items within Factor 1, item six's mean was the highest at 4.67.Group interviews revealed that students spent a lot of time planning what they were going to say, writing everything down, rehearsing, and then making changes to the script several times before finally recording their voice comment.Students interviewed reported recording 3 to 5 times before being satisfied with their comment.This process even applied to short voice comments and responses.One student pointed out, "Bubbly helped me with writing.So much time is spent on planning and writing."Consequently, speaking activities on Bubbly may be said to help learners to reflect and improve on not only their speaking skills such as fluency and pronunciation, but also their writing skills.This increase in selfcorrection and l2 selfawareness can be seen in the following student statements: Before I record Bubbly I prepare script, so I can practice and pay attention to my pronunciation.
If I don't like it, I can record again and again.
Bubbly helped me to notice my mistakes.I can record again if not right.

I have to plan by writing and choosing words
Bubbly made me more conscious of my English skill Sometimes I asked teacher how to pronounce.
It (Bubbly) made me think more about my speaking skill.
My pronunciation is not good more than I thought

Motivation and confidence
Loadings for factor 2 resulted in five items (7, 11, 12, 13 and 15), all of which relate to moti vation and confidence.Factor 2 accounted for 17.9% of the total variance.Items 12 and 13 revealed that participants see activities on Bubbly being superior to activities in class.When asked why, students expressed that they wanted to perform well as their comments were being shared on a social network, which enables their classmates to listen to their com ments as many times as they like.One student stated, "Bubbly helped more than speaking in class because we thought more carefully" and another stated, "Everyone can hear my voice, so I try to speak well."Another student stated, "In front of my friends in class I have too much pressure and cannot fix my mistake" Item 11, "I feel motivated by listening to classmates' voice comments," had the highest weighted average of 4.6.One student stated, "I enjoy listening to everyone's comments.Especially, I liked listening to everyone talk about their hero".It would therefore seem that the social networking aspect of Bubbly gives birth to an environment where students feel socially motivated to perform well, which makes them work harder and consequently, makes these tasks in some ways more motivating than traditional tasks in the classroom.As for item 7, which asked students if Bubbly increased their confidence in speaking English, students reported an increase in both their confidence to speak online and to give presentations.Writing down their scripts, rehearsing, and then finally recording were all very helpful steps, which students felt prepared them for presentations in English.However, students did not feel that the project helped as much to increase their confidence when speaking in person in English, as tasks did not involve much spoken interaction.Thus, students interviewed still voiced the importance of having speaking tasks in class, as they tend to lend themselves more to real conversation due to being facetoface and in real time.

Social affordances
Factor 3 received appreciable loadings all relating to social affordances, accounting for 10% of the total variance.Items 18, 19 and 20 highlight the importance of social identity.Social identity theory suggests that individuals work towards maintaining or enhancing their selfesteem and are intrinsically motivated to achieve a positive selfconcept (Tajifel & Turner, 1979).This can be seen in that the majority of students interviewed, including those with low selfperceived proficiency, reported that they cared about looking good in front of their peers and were therefore intrinsically motivated to produce good English and inter esting comments.Students also reported that their level of motivation and engagement increased when students left voice comments on their Bubbly posts.Several of the students interviewed mentioned that one of the most rewarding aspects of the Bubbly project was receiving personal comments on their page.In fact, item 21, "I am pleased when others leave a comment for me", received the highest mean in the survey with 5.30.This is of no surprise as brain research supports the engaging nature of social networking environments.Sherman (2013) articulates this well when stating, Brain research also suggests why the use of digital technology, especially social media, can be so addictive.When we receive a small burst of information, such as a tweet or a status update, our brains release the same pleasureproducing chemical dopamine as is released when we eat chocolate, fall in love, or use cocaine.(p.8) An examination of student comments showed them to be all positive, serving as a form of peerencouragement.One student said that she enjoyed listening to personal comments on her page as they helped to confirm that her classmates were in fact listening to her voice bubbles and this gave her more confidence in her English ability.Students also responded quite positively to private voice feedback from the teacher.Voice feedback was only given for task four, but students stated that they would have liked voice feedback for the other tasks as well.One student indicated that it would have been nice to receive voice feedback for the first task as she was unsure of her pronunciation of some words.

Technological affordances
Lastly, Factor 4 received 4 loadings, which all pertain to technological affordances.This final factor accounted for 6.8% of the total variance.8 out of the 10 students interviewed felt that the soundtrack and voice effects/filters on Bubbly helped them to feel less anx ious and gain more confidence.In particular, students liked the ability to rerecord their comments until they reach a satisfactory result.The following comments all relate to the technological affordances of Bubbly.
Soundtrack makes my voice comment more high quality.
Sounds can hide my English skill.
Soundtrack makes my comment like a commercial.
When I listen to friend's comment with music it's nice.I don't like my voice, so add music and change my voice, yes, I like.
Only voice is a little nervous.
No music is clear.Easy to hear.
However, two students interviewed did not agree that features such as adding sound to the background or effects made any difference.One student stated that she didn't care if there were sound effects or not and that it did not change her level of confidence.The sec ond student stated that she did not like to use sound effects because sometimes it made it difficult to hear the speaker.They both; however, did agree that the ability to rerecord was very beneficial.

Students' perceptions of tasks
To answer research question three, students interviewed were asked to rate the tasks based on their level of perceived enjoyment and engagement.As discussed earlier, Bubbly project tasks were designed using an adapted version of Egbert's model of flow in language acqui sition with the purpose of motivating and engaging students outside of class.Interviews revealed that Task 1 was seen as being the most enjoyable and engaging of the four tasks followed by Task 4, 2 and then Task 3. Students expressed a lot of interest in the topic of heroes and enjoyed not only sharing their heroes, but also enjoyed listening to other stu dents talk about their heroes.Of the four tasks, this first task was by far the richest in terms of content, encompassing pictures, videos and website links.Consequently, it also gener ated the longest voice comments.Students stated that they spent a lot of time searching for pictures, videos and information about their hero and admitted spending much more time on this task than any of the other three tasks.One student admitted to spending two hours researching, writing and recording her post on top of the initial hour of class time that was given to all students.With that said, it should be noted that as Task 1 was the first task students did for the Bubbly project, there was most certainly a strong sense of novelty, which may have boosted engagement.Further, this task, being a final project, was worth more points than any of the other tasks.Nevertheless, the popularity of Task 1 helps to shed light on what makes a task engaging and may be considered a good example of the type of task that can help to facilitate flow as well as both intrinsically and extrinsically motivate students.
Task 4 held second place in terms of popularity.Students reported that they enjoyed working with a partner and creating a roleplay together.Students also liked receiving pri vate voice feedback from their respective teachers.However, students admitted to rushing this task a little as it required them to meet with a classmate on campus to complete the activity and they did not want to take up too much of their partner's time.This reluctance to take up another's time may be argued to be a culturally influenced concern.Similarly, another student pointed out that doing a task alone increased the amount of time spent on the task as she could easily rehearse in the privacy of her home at a time convenient for her.Task 2 was considered semiengaging, but a little too easy and all students interviewed felt that task 3 was too difficult.Despite this perceived high level of difficulty, students felt that Task 3 was important as it helped them to prepare for the speaking test later in the semester, which had a similar task.In short, one could argue that, unlike tasks 1 and 4, tasks 2 and 3 did not balance challenge and skill appropriately and thus, failed to help learners enter the flow channel, depicted in Figure 2 of this paper.Interviews with students suggest that Tasks 1 and 4 did balance challenge and skill appropriately and thus, these tasks helped students to enter the flow channel, which helps to partly explain students' perception of these tasks being more engaging.In the case of Task 1, which better suited students' personal interests, allowed for online information searching, sufficient time, balanced challenge and skill well, and was more conducive to sharing pictures, videos and website links, this task may be considered the most successfully designed of the four.

Discussion
Results revealed that social and technological affordances, motivation, and confidence are very much intertwined and together positively influence students' perceptions of their l2 spoken production skills.In fact, several studies have stressed the important relationship between motivation, confidence, and ability (Butler & Lumpe, 2008;Phillips & Lindsay, 2006).This interrelated relationship has also been found in studies utilizing computer mediatedcommunication (cmc) (Wu, et al., 2011;Wu & Marek, 2010).The findings in this study echo this same important relationship cycle.However, it would seem that the combined role of social and technological affordances and their positive influence on this relationship cycle have not been examined until now.Therefore, the findings in this study bring exciting new insights into the benefits social and technological affordances have to offer within a voice microblogging environment.The following figure may serve to illus trate these findings and provide answers to the research questions posited in this study.As seen in figure 4 below, technological affordances help to lower anxiety, consequently boosting a language learner's confidence and level of motivation.The social networking element helps to raise students' levels of engagement and students are motivated to work even harder when what they produce is shared within a social network among classroom peers.Together, this results in students' perceived speaking ability increasing.This relationship can also be seen in Egbert's (2003) adapted model (Figure 2), which was used to design the tasks in this study.Moving from left to right in the model, one can see that the first section, which encompasses an appropriate challenge, control, sufficient time etc., along with social and technological affordances, may be considered salient components that help to raise learners' confidence.This increase in confidence helps to motivate and engage learners.The social networking aspect helps to further increase this engagement, which in turn, facilitates flow.This increase in motivation and/or the state of flow leads to improved performance on a task, which can lead to changes in competence and/or perfor mance, once again raising students' level of confidence.As stated previously, task one may be considered a very good example of a task that fulfills the requirements of this adapted model, allowing for increased motivation and engagement, and consequently, a perceived improvement in performance.Lastly, results from Factor 1, which examined students' perceived increase in spoken production skills, showed that speaking activities on Bubbly helped learners to reflect and improve on both their speaking and writing skills.In other words, Bubbly greatly facili tates reflection and consequently, fosters the development of metacognitive strategies.This can be seen by students' reported use of strategies such as brainstorming and rehearsing before recording a comment, evaluating their voice post, and finally rerecording if neces sary.The fact that students' published blogs were within an online, publicallyshared space, encouraged students to utilize such strategies more and as a result, raised the quality of their learning outcomes.conclusion Findings suggest that speaking tasks which fulfill the requirements of Egbert's (2003) model of flow in language acquisition and which offer both social and technological affor dances such as the ones discussed in this paper, are extremely beneficial to learners as these affordances help to elevate students' level of confidence, motivation and engagement.However, due to the small number of participants (n = 33), the results of this study should be interpreted with some caution.A larger sample would be needed to further test and confirm the validity of the findings.Whether the findings would hold true with learners from other courses, with different academic abilities, or from other cultures is unknown.
Despite these limitations, the findings in this study are considered useful to other researchers and educators who are interested in examining the use of a voice microblog to help develop students' l2 spoken production skills and raise student motivation, engage ment and confidence in a blendedlearning environment.During the duration of this study, no problems occurred in relation to unwanted followers joining the network.However, as Bubbly is an open social network, educators should be aware of this possibility when con sidering integrating such a tool into their language learning classroom.
For future research, the author suggests a greater variety of tasks over a longer period of time.In particular, tasks which house more discussion would better serve as a way for students to practice their spoken interaction skills.Furthermore, as Bubbly is a social net work, it may be a good idea to stretch some activities beyond the classroom network.For example, students could be given the opportunity to participate in language or culture exchange with native speakers of English from another university.This would add a more authentic element and therefore further increase motivation and engagement.As for the level of engagement, an examination of participants' flow experiences is beyond the scope of this paper; however, as several students reported being highly engaged, notably when doing the first task, and revealed many of the traits associated with flow, such as losing track of time and a high level of enjoyment.The creation and examination of tasks simi lar in design to task 1 may merit interesting results in the area of motivational design in elearning environments.Lastly, as tasks were created mainly to help students with spoken production skills, research into the benefits of using a voice microblog to create tasks aimed at improving students' spoken interaction skills is an area in need of exploration.Voice microblogging is an exciting new area of study, which opens a new window into the future of mobile assisted language learning.The researcher hopes that the findings in this study may serve as a reference for future researchers interested in voice microblogging and in particular, the social and technological affordances that it offers.

acknowledgements
The author would like to thank the students for sharing their opinions on using Bubbly as a blended learning tool for practicing their spoken production skills and Gary Cook for agreeing to use the application in his class and helping the author to design speaking tasks to be used on Bubbly.

Figure 4 .
Figure 4. Confidence, motivation and ability cycle within a voice microblogging social networking system

Table 1 :
Results of factor Analysis for All Subjects(N = 33)