How To Give Effective Feedback to Language Learners?: An Example of Vygotskian Responsive Assistance

  • Previously, I wrote a piece critical of what many EFL teachers might think is good advice for giving student feedback. I mentioned in that post that a more effective method would be Dynamic Assessment. This post then is a follow-up to that post. Here I detail what exactly DA is and then provide an extend example with data of what it looks like in action.

Introduction

Since “the sociocultural turn”, the field of second language acquisition has seen a shift in the way many educationalists and linguists view the dialogic nature of the teacher-student interaction. The turn from traditional initiate, respond, evaluate forms of teaching has been replaced with a Vygotskian form of responsive assistance, in some cases called instructional conversations (IC). The core these IC communications, according to Meskill & Anthony (2010) is the dynamic relationship between two people and how they “recognize and respond appropriately to the myriad of teachable moments” (loc. 515).

In order to recognize and respond, much more than attending to the forms and function of a language in the classroom is necessary. Our sociocultural theory of education must also be ecological in viewing itself as situated in a particular environment, with its own unique and variable affordances.

Technological advancements has made online or computer mediated communication (CMC) and language teaching much more available, both to students and teachers who might not have many reasons to engage with the new technology. The purpose of this essay then, is an attempt at understanding how effective Vygotskian IC can facilitate second language (L2). To do so, we will present useful concepts and a specific example of learning between an English tutor and a university-aged Korean L1 English-language learner and how the ecology of synchronous online computer mediated communication (SCMC) affects language learning. The specific research questions of this study then, are as follows:

Q1: How many IC feedback interactions facilitated L2 learning?

Q2: In what ways did time and mode affordances positively or negatively impact IC in SCMC?

To begin, first we will examine the relevant literature related to IC and SCMC ecology. This will be followed by the qualitative methodology of the current study and the observed results of the tutoring sessions. This report will conclude with a discussion which will synthesize the information in the results and point towards potential future research options.

 Useful Concepts

Instruction Conversation (IC)

Since Vygostky’s formulation of the ZDP, educational theorist have devoted much effort into describing or examining how other-assistance becomes self-assistance and finally automatization in learning. Tharp & Gallimore (1988) identified a difference between early stages of assistance, when a learner may need more help, than at near-self assistance stages. Early help is focused and elaborate, with the intention of getting the learner to see the “component steps (p. 55) of the task. At later stages the help is abbreviated and more general. Tharp & Gallimore (1988) describe this assistance as “responsive assistance” (RA; p. 53) and establish the idea that other-help is contingent on the exact needs of the learner and how near self-assistance they are.

From the RA of Tharp & Gallimore (1988), Meskill & Anthony (2010) describe their Vygostkian assistance in CMC language education as instructional conversation (IC). The focus on the communicative act, instead of just assistance, is the recognition of the “human activity” (loc. 527) within teaching. Pedagogically, the instructional aspect of their IC is the conscious effort the teacher puts into formulating their conversational turns in order to guide the student towards comprehension and ultimately automatization of the language forms and functions. For Meskill & Anthony (2010) there are two key factors that make IC instructional. First is the intention to instruct or teach and second is the constant updating and awareness of the student’s current ability and potential.

For Meskill & Anthony (2010) there are several forms of technique which teachers can take advantage of in IC, including modeling and saturating correct usage, pointing out incorrect use of form or vocabulary and the level of implicit or explicitness in the IC. They do not, however, suggest that IC should move from the most implicit to explicit, only that instructional moves should be tailored to the needs of the student. Other researchers have explicitly ordered their assistance from the most implicit to explicit. In their 13 point scale, Aljaafreh & Lantolf (1994) argue that the communicative framing of student and teacher is itself an implicit form of assistance, where the student may be more careful or more likely to examine their own production just because an experienced other is present. Lantolf & Poehner (2011) describe a more simple scale created by an elementary school teacher. The main idea behind both scales is the movement from very broad to very specific assistance, following Vygotsky’s concept of automatization. For example, the teacher in Lantolf & Poehner (2011) begins their scale with a pause and then moves to repeating the whole phrase that a student has produced, not specifying where the error occurred. In both scales, the teacher moves down the scale, becoming increasingly more specific and narrow in their assistance, until the student makes a correction, or the teacher ultimately explains the answer and why.

The move from implicit to explicit is theoretically important. As a learner requires other-assistance, it is possible to see microgentic growth if the learner goes on to require less and less other-assistance for the same type of error. Lantolf & Poehner (2011) document the case of a young boy who initially required a high level of assistance in gender agreement in Spanish, but in later IC moves, required a less explicit amount. Three days later, the boy successfully produced the grammatical forms without other-assistance. The authors argue that this is evidence of growth in language learning.

Affordances and IC

Establishing that learning occurs, however, is difficult to show even in formal IC scales when we consider that students may not move up or down the scale in ways that clearly show growth as defined in Lantolf & Poehner (2011). van Lier (2000) notes that in the scientific study of learning, researchers must be careful to examine not just the pedagogical processes and contexts, but must also “show the emergence of learning” and “the location of learning opportunities” (p. 250), or what the environment surrounding learning affords the learners.

Lamy & Hampel (2007) describe these language learning affordances in terms of how learning is mediated. In their description of computer-mediated language learning, they identified three areas that afford learners opportunities for learning: participant interactions, the tasks they perform and the technology used. Face-to-face (f2f) language learning is both similar and very different to computer mediated communication (CMC) and within CMC environments, there are various affordances related to the type of technology used.

Important affordances in CMC environments are related to time and mode. Lamy & Hampel (2007) include a list of the various CMC environments, such as chatting, e-mail, forums or video-conferencing; along with the shared and different affordances. Time affordances relate to how quickly each interlocutor turn-takes in the conversation and is split into synchronous and asynchronous formats. Chatting affords synchronous communication, as interloctors wait for each other to turn take in real-time, much like f2f oral communication. Unlike f2f environments however, chatting affords specific limitations related to how and when each participant ends or begins their turn.

Lai & Zhao (2006) for example, examined noticing in f2f and text-based chatting environments. They identified two affordances in online chatting that appear to facilitate noticing, longer wait times between responses and the permanency of written text. The authors suggest that between these two ecological affordances, learners were more able to process incoming information and re-analysze it due to the written format. Additionally, the authors note that the lack of f2f paralinguistic cues may have been beneficial in that the learners were not distracted and spending processing power on interpreting them.

Mode affordances are described as the semiotic resources afforded in a given environment (Lamy & Hampel, 2007). Importantly, communication is rarely, if ever, just linguistic, but almost always multimodal (Kress, 2015). What is sometimes called the para-linguistic features of communication, themselves afford important opportunities in communication which speakers and hearers use. Kress (2015) notes that the multimodality of communication offers “complementarity” (p. 56) in the work required in a communicative act. What one mode (such as language) lacks, another (gesture) may be able to fill easily. As a combined whole, multimodal communication is the entirety of the semiotic act.

Together, time and mode create the “semotic budget” (van Lier, 2000) that interlocutors have in any situated communicative act. This budget necessarily both provides and constrains possible interactions. Vetter & Chanier (2006) found that language learners in their study were more likely to interact not as a feature of their language ability, but depending on which modality of communication they preferred, text or spoken. An advanced second language speaker may struggle or interact less in a written mode if they are less able or familiar.

An Extended Example of Vygotksian Responsive Assistance and IC

Background

This example examines the interaction between one student and their tutor, who is also me. The tutor is a 28 year old native English teacher in a traditional South Korean elementary school. I have been teaching for about six years, though not often through CMC, such as chat-tutoring or skype.

This was my first attempt at formal uses of IC and feedback, but it has been part of my teaching repertoire in some fashion since I was in teacher-training. Prior to this tutoring, my IC theory was related to the idea of simply not giving the answer away to the student. A primary example of this pre-theory IC would be spelling. If students asked how to spell a word, I regularly just repeats the word back to the student and makes them hypothesize and notice the sounds. As the student makes guesses the teacher would move on or correct them. This study represents the first attempt at a SCT-grounded attempt at IC.

The student in this example is a 20 year old undergraduate English language major. Her current English level is somewhere between an intermediate-mid to intermediate-high user of English. Additionally, she is beginning to learn German. The student has lived her entire life in South Korea where she attended public school until she began university, when she spent one year abroad in Australia. She is very familiar with the use of online chatting programs in her L1 (Korean), and reports using them in English as well. On a scale of 1 to 8 where 1 is “no ability” and 8 is “fluent”, she self-reports at a 7 for reading online chat like Kakao, and a 6 for writing in the same environment.

Instructional Conversation

Meskill & Anthony’s (2010) approach to IC is the guiding principle of this example and as such it was the goal of this study to “tailor” (loc. 536) the interaction between tutor and student to both improve the student’s ability to use English and also maintain a functional conversation. As such, there was an attempt to both promote use and correct usage in the student’s English through several tactics mentioned by Meskill & Anthony (2010) such as saturation, corralling or linguistic traps. Alongside the promotion of specific usage, the tutor attempted to correct linguistic errors through a synthesis of IC and RA.

IC in this study was formalized in a similar fashion to the dynamic assessment of Lantoft & Poehner (2010) where an IC chart was constructed which allowed the tutor to move from the most implicit instructional moves to the most explicit in order to promote use and correct errors.

da-scaleTable 1: IC chart used in the current study. Adapted from Aljaafreh & Lantolf (1994) and Lantolf & Poenher (2011).

Data Collection

The information documented by the IC chart visualizes each possible instance of IC and allows the researcher to see how and when exactly the student needed assistance, if at all, from the tutor. As the student required less and less assistances in IC interactions, this was seen as evidence of growth or learning. Over the course of the tutoring sessions, if the student showed less and less need for the tutor to assist in their error correction or promoting the use of specific grammatical or lexical forms, or began self-regulating this would show up in the IC chart. Additionally, no longer making the same mistakes can be evidence as well of growth.

A simple example of microgenetic growth might be seen in Example 1 below. The grammatical form in question is the use of determinate and indeterminate articles in English, a difficult grammatical form to use for Korean L2 English learners, but also a difficult form to teach. Over the course of two sessions, the student required less assistance from the tutor to arrive at the correct form. In the first tutoring session, the student was describing one of six pictures of a car moving away from a motorcycle accident. Because there was only one car, the tutor felt the correct form should be “the”.

Excerpt 1

First error Second error
tutor
ok, either would be correct 🙂
a car ran away?
Student
i mean to flee
tutor
right 🙂  “A car ran away”?
Student
erase a
tutor
ok erase “a” and…
Student
hint hint
tutor
hehe
__ car ran away?
Student
theee
tutor
great 🙂
student
a motorcycle guy was talking with a woman so he didnt see a car after a corner
so he crushed a car
and a car guy got angry so he fought with a motorcycle guy
finally he required money to a motorcycle guy
thats why a motorcycle guy got frustrated
~the end~
tutor
Ok, are there any changes you want to make?
Student
so he fought with the motorcycle guy
finally he required money to the motorcycle guy
thats why the motorcycle guy got frustrated
tutor
good job with “a” and “the”

The student required a very specific form of mediation, linguistic gap filling, in order to understand the error the tutor was attempting to correct. Later, the student was summarizing the pictures as a complete whole and by simply asking the student if they wanted to change anything, implying that perhaps they should, the student made changes to her article use.

These IC interactions were mediated by the affordances of the environment in which they took place. In order to track the evolution of the effect of these affordances, every tutoring session was followed by a very structured journal which was provided by my own professor. The journal outlined the important affordances of time and mode related to synchronous online communication and allowed the tutor to self-examine their ability to respond to the student’s educational needs.

As mentioned in Lai & Zhao (2006), online chatting affords specific affordances. These can have both positive and negative effects in learning. For example, in the first tutoring session, keeping the conversational flow orderly was difficult in the beginning, as can be seen in table 2.

In table 2, I attempted to get the student to go line-by-line through the picture descriptions which she had already made. At first, the student fails to understand the directions completely, but instead summarizes the story herself (an activity that would be done later). Both the tutor and the student produce long, multi-sentence entries, which took minutest to compose, even though they were miscommunicating. In f2f interactions, interlocutors are likely to interrupt such miscommunications with something like, “what was that?” or “excuse me”. However, in the SCMC chat program used, the only thing that could be seen by each participant was a simple sign that indicated that the other was typing. Additionally, the tutor attempts to clarify and then move on with the task before getting explicit confirmation, something that is common in SCMC.

Excerpt 2:

tutor
you will tell a story about it in a little while
so you can hang on for right now
Student
finally he required money to a motorcycle guy
thats why a motorcycle guy got frustrated
~the end~
tutor
haha, well, let’s keep that story in mind, but first, I am going to copy and paste your descriptions of each picture and I want you to re-read them and tell me if there is an error or not
Student
about the plot?
tutor
the picture description, the first thing you wrote about
A
i can see a man and a woman is fighting with a motorcycle guy
also i can see some crowds around there
are*
Student
my first guessing was wrong
a second one with whole story is the one i wanted to say
tutor
Right, your second story is more clear 🙂
but I want you to re-read your sentences for each picture, and just see if there are any language errors
starting with A

Results

Learning promotion in IC

Throughout the three tutoring sessions there were 18 instances of error correction and 4 instances of use promotion. Additionally, there were 3 test IC items designed to illicit what Poehner (2007) describes as transcendence, or the ability of the student to use forms mastered in one context and accurately apply them in a separate, unrelated, context.

Of these 25 instances of IC, four areas of potential for learning were identified. These include errors related to: subject-verb agreement, article use, lexical use and preposition use. There were other, one-off, errors such as a misplaced conjunction or pronoun use that happened only once in all three sessions. These single instances do not represent good opportunities to document learning, though it’s possible they are representative. For our purposes here, the four areas identified will be the focus of learning documentation.

Table 2: The error types, divided by tutoring session and the level of IC mediation required.

1 2 3 Learning?
Subject-verb
agreement
1st – 0
2nd – 3
3rd – 0
4th – 3
No errors No errors Across session – yes
Articles 1st – 3
2nd – 4
3rd – 0
4th – 0
1st – 2 1st – 4
2nd – 4
Within session 1 – yes
Vocabulary No errors 1st – 5
2nd – 5
3rd – 4
4 – 0
1st – 0
2nd – 2
Across sessions – yes

Within session 2 – yes

Prepositions No errors No errors 1st – 2
2nd – 2
no

Evidence of microgentic growth can be described and analyzed in both intra-session and inter-session terms, or as both occurring within individual tutoring sessions and between them. This helps understand, for example, why the student appears to go back and forth in their ability with subject-verb agreement in session 1. When looking across sessions, it becomes more clear that her ability to correctly conjugate English verbs is somewhere between self-regulation and automatization.

Article use is difficult to determine, due to the highly contextual and variable situations that articles are used in English. However, it can be seen in session 1 that the student moved from high-levels of mediation to much more implicit levels as the session progressed.

Excerpt 3

Second instance of article IC Third Instance of article IC
tutor
“and a motorcycle guy frustrated”
Student
is frustrated
and erase a aagain
tutor
Ok good…
tutor
a car ran away?
Student
i mean to flee
tutor
right 🙂  “A car ran away”?
Student
erase a
tutor
ok erase “a” and…
Student
hint hint
tutor
hehe
__ car ran away?
Student
theee
tutor
a motorcycle guy is replying to a woman’s question
“you ok?” “yup”
Student
erase every a again
the
tutor
where are you putting “the”?
Student
a places

However, perhaps more than demonstrating learning of article use, this demonstrates learning the teacher-student and task expectations. In the second instance of article correction, the student doesn’t respond to the teachers attempts to draw attention to the article but instead focuses on the meaning of her writing. When the tutor moves to level 2 mediation, repeating the whole phrase questioningly, the student responds with “I mean to flee”. In later mediations the student does not make this same error, but attends more focusedly on her grammar when the tutor engages her.

The same analysis seems valid for the vocabulary items in IC mediation. Similar to article use, lexical variability is wide and hard to show that one instance of using a wrong word is applicable to another. However, in the second and third sessions, a task toolkit was provided to help the student focus on the kinds of vocabulary words that the student should use. In the second session, the first instance of lexical correction, I used one of the most explicit IC moves in the data, I pointed out the specific nature and area of the error. Later, as the student becomes familiar with using the task toolkit, less explicit levels are used.

Excerpt 4

2nd use promotion 3rd use promotion 4th use promotion
tutor
finally, their island broke into two pieces 
Student
good job
tutor
can you think of another way to say that?
Student
since they fought on the island, it broke into two pieces
tutor
good, yeah
it broke into two pieces?
Student
hint
tutor
it broke into two pieces?
your writing is good, but is there another way to write it?
Student
instead of broke?
then i will use crashed into
tutor
Is there a word in the toolkit you might use?
Student
divided
tutor
what happens when the figures fall into the ocean?
before the monster
saves them
Student
they are going crazy
almost died
tutor
they are going crazy,
they _____, but then the red figure saves them
Student
sank 
tutor
anything else in that last sentence?
” They fell into the ocean and kind red monster saved them”
Student
rescued? 

In the second instance, the tutor specifically calls the students attention to the task toolkit, reminding her that part of the exercise is to use those forms. With that in mind, the 3rd and 4th instances are perhaps indicative of her ability to attend to the task, more than lexical problems.

Negative and positive impact of affordances on IC

Similar to Lai & Zhao (2006), this study found that the impact of the synchronous nature of SCMC afforded the participants two key features related to the permanency of written text and the ability to wait longer between turns compared to f2f interactions. Additionally, the interaction between the affordances of time and mode fundamentally influenced the way I approached the sessions.

Because of the longer wait between turns (time affordance) I had the ability to use the level 1 mediation “pause” in various ways. Beyond just waiting, which wasn’t effective, I frequently typed “hmm…”, a written analogue to the f2f oral response. Additionally, the chat technology afforded the participants the ability to know when the other partner was typing with moving ellipsis at the bottom of the chat. This could be manipulated by the participants to communicate that they were attempting to write something, but perhaps paused mid-way. For myself, this was used at times to signal to the student that I was trying to think of a way to begin error correction.

It is possible that these level 1 pauses contributed to frustration on the part of the participants. For myself, the extended waiting times were often difficult to bear. The student as well, at times seemed to give up on the “pauses” and went straight to responses like “I don’t’ know” or “give me a hint”, as excerpt 5 from the third session shows.

Excerpt 5

tutor
ah right
with others?
Student
yep
tutor
is their something missing there? with others?
Student
no idea with that
tutor
with ___ others
Student
the

The ability to bold, italicize or create blanks to fill written text in online settings is another affordance that the tutor used and the student seemed to take advantage of in the tutoring sessions. It is possible to do some of these written enhancements in f2f settings, however, they take more time to do and things like bolding or italicizing are difficult to make clear to students via handwriting. Excerpt 5 also demonstrates an example of blank filling which immediately draws the students attention to the area needing correction and is easy to produce in SCMC environments.

Finally, for the IC, being able to copy and paste previous dialogue or usage was very important for helping the student see exactly what linguistic environment the error in question occurred in. Additionally, the permanency of the written text in SCMC allowed the participants more mental processing to evaluate their written messages.

Discussion

IC and Learning

Lantolf & Poenher (2011) take the less explicit mediation by the elementary school teacher in their study over time to be evidence that the students exhibited microgentic growth, or that they are in a sense learning. The results from this study also show similar, albeit complicated, data that would support a similar claim. The results from the IC mediations related to subject-verb agreement seem to be solid evidence that the student at least moved towards self-assistance and possible automatization.

Through the use of dynamic assessment in excerpt 6, the student was also able to easily and quickly identify misuse and correct use of subject-verb agreement in writing unrelated to anything to do with the tutoring sessions. Demonstrating a type of transcendence in the Poehnerian style (Poehner, 2007).

Excerpt 6

tutor
“No one expect results like that on the test.”
Student
expects
tutor
nice!
tutor
“the girl twists the lemon forcefully.”
Student
the sentence seems ok
tutor
great ^^

However, another interpretation is available. While the student is demonstrating microgenetic growth, it seems that the movement from explicit to implicit may be a feature of the collaborative construction of the learning frame. Or in other words, the student was learning what the teacher expected her to accomplish in each task through trial and error. This is possible most evident in the examples of lexical use. The key moment in which the student seems to become aware of the need to use the toolkit, is when mediation becomes much less explicit. The student exhibits learning, but possibly what she has learned is not so much related to language ability and use, but instead the task design and expectations.

Article error correction also seems to be vague in terms of what exactly was being learned. In part, this is my fault, as some of the article “errors” weren’t really errors in the way we might suspect. Instead errors, such as in excerpt 1, demonstrate a preference that the tutor had for picture description. It becomes reasonable to suggest then that the student isn’t exactly learning how to use English articles, but instead, how the tutor thinks they should be used. Of course every teaching environment works in a similar way, with the students learning the biases of their teacher or mentor, but in this case, it becomes difficult to see and the varying levels of IC mediation seem to support, what exactly the student is learning.

In terms of SCT and IC, either result is favorable for theory. As the results still suggest that the student is moving through their ZDP and that the instructional moves are intentional motivated to correct a language behavior. However, in terms of language education, this result could be disconcerting and it is worthwhile to examine closer the difference between learning to interact in a specific context and task, and learning language through IC.

IC and Affordances

Similar to previous research related to SCMC interactions (Lai & Zhao, 2006) this study found positive influences from the SCMC environment. The ability to manipulate text in various ways was effective and efficient in helping the student focus their attention on important areas. Additionally, the longer time between turns as compared to f2f interactions and the permanency of the written form is seen as both improving the quality of the student interaction, but also the IC performed by the tutor.

The specific ecology of the tutoring sessions plays an important role in the vague results obtained. The tutoring sessions afforded the student and tutor a situation much like the first day of class, with the teacher trying not so much to teach content, but the “how things work here”. Many teachers do this by giving the students simplified, but similar, tasks to what they will be expected to perform throughout the year. The teacher in these situations isn’t so interested in the student’s ability to perform the task well; so much as they want the students to get accustomed to the task itself.

The SCMC environment of the tutoring in this study afforded the student a somewhat careful introduction into how the other sessions would be conducted, however this was not explicitly stated to the student and it wasn’t really the tutor’s intention to first teach the student how to do the tasks and then conduct language teaching. Instead, they were the same thing. The extra processing required to do both tasks is possibly a confounding factor to determining exactly what kind of growth the student demonstrated.

Conclusion

Both participants, myself and the student, did demonstrate learning of some kind. This study of the relationship between IC and learning and how the SCMC environment mediates it found that both participants benefited from the use of online, written language and the ability to process information for longer periods of time while still engaging in synchronous conversation. However, the type of learning was found to be at best vague. Whether the student was really demonstrating growth as previous research has suggested, or just learning the tasks is not entirely clear from the data in this example.

It is suggested then that future research attempting to demonstrate learning or growth clarify how the movement from explicit to implicit assistance is evidence of what type of learning. Performing IC in well-established SCMC learning situations, whose tasks are familiar to both teacher and learner is an important requirement for determining the type of learning exhibited.

References 

Aljaafreh, A., and Lantolf, J. P. (1994). Negative feedback as regulation and second language learning in the zone of proximal development. The Modern Language Journal, 78(4), 465-483.

Johnson, K. (2006). The sociocultural turn and its challenges for second language teacher education. TESOL Quarterly, 40(1), 235.

Kress, G. (2015). Semiotic work: Applied Linguistics and a social semiotic account of Multimodality. AILA Review, 28(1), 49-71.

Lai, C., and Zhao, Y. (2006). Noticing and text-based chat. Language Learning & Technology, 10(3), 102-120.

Lamy, M. N., and Hampel, R. (2007). Online communication in language learning and teaching. Basingstoke: Palgrave Macmillan.

Lantolf, J. P., and Poehner, M. E. (2011). Dynamic assessment in the classroom: Vygotskian praxis for second language development. Language Teaching Research, 15(1), 11-33.

Meskill, C. and Anthony, N. (2010). Teaching languages online (2nd ed.). Retrieved from http://read.amazon.com

Poehner, M. E. (2007). Beyond the test: L2 dynamic assessment and the transcendence of mediated learning. The Modern Language Journal, 91(3), 323-340.

Tharp, R., and Gallimore, R. (1988). Rousing minds to life. Cambridge: Cambridge University Press.

Van Lier, L. (2000). From input to affordance: Social-interactive learning from an ecological perspective. Sociocultural theory and second language learning, 245.

Vetter, A., & Chanier, T. (2006). Supporting oral production for professional purposes in synchronous communication with heterogenous learners. ReCALL,18(01), 5-23.

Advertisements

One thought on “How To Give Effective Feedback to Language Learners?: An Example of Vygotskian Responsive Assistance

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s