Dynamics of classroom WTC : Results of a semester study

The fact that L2 willingness to communicate (WTC) can fluctuate over different time scales is no longer disputed as numerous studies have proved a dynamic rather than trait-like character of the concept (cf. MacIntyre & Legatto, 2011; MacIntyre, Burns, & Jessome, 2011; Mystkowska-Wiertelak & Pawlak, 2014). The changes in the intensity of L2 learners’ readiness to engage in communication in the classroom context have been investigated in a number of studies attempting also to establish factors capable of stimulating or hindering WTC (e.g., Cao & Philip, 2006; Pawlak & Mystkowska-Wiertelak, 2015; Peng, 2014). Many of empirical explorations of L2 learners’ WTC, although representing the ecological perspective, provide solely a snapshot of classroom reality, thus generating a question if such one-at-a-time picture frame can sufficiently accommodate diverse factors that impinge on learners’ readiness to engage in communication. The application of a longitudinal design was thus dictated by the author’s intention to gather information on WTC ebbs and flows not only during single lessons but also continuing over a number of lessons conducted throughout a semester of study, which allowed for exploring a wider spectrum of conditions that affect WTC of advanced learners of English attending speaking classes. Detailed lesson plans, interview and questionnaire data were used to interpret WTC fluctuations reported by the participants at 5-min intervals in the course of 7 lessons. Qualitative and quantitative analysis of the data revealed that the intensity of communicative behaviour depends on an intricate interplay of personal and group-related factors, with Anna Mystkowska-Wiertelak 652 a special role ascribed to the instructor whose knowledge of the group characteristics and needs coupled with didactic skills can greatly contribute to increasing WTC in the classroom.


Introduction
The study of willingness to communicate (WTC), which emerged from exploring, first, mother tongue (L1) and next, second/foreign (L2) learners/users' reticence or reluctance to communicate, has been attracting the attention of researchers for a few decades now.The early conceptualization of L2 WTC as a learner's stable characteristic or as a predisposition to initiate or avoid communication with others when given a choice (McCroskey, 1992) has been overruled following the advances made by MacIntyre, Clément, Dörnyei, and Noels (1998), who presented L2 WTC as an outcome of a joint operation of a number of distal and proximal antecedents comprising both individual tendencies as well as intergroup relations.This shift of focus to the contextual nature of WTC resulted in a whole host of studies aimed at disclosing situated, personal, and context-related variables shaping L2 WTC in the classroom setting (cf.Cao, 2011Cao, , 2013;;Cao & Philip, 2006;Kang, 2005;Mystkowska-Wiertelak & Pawlak, 2014;Weaver, 2007).The recognition of the role of context in the emergence of WTC has highlighted the dynamic character of the construct capable of fluctuating in different time scales, within a single class or a single task.Investigations into the dynamic character of WTC have differed considerably in focus and scope involving laboratory-style investigations, the prime examples being the application of the idiodynamic method by Mac-Intyre and Legatto (2011) or classroom-based studies like those conducted by Peng (2014), Pawlak and Mystkowska-Wiertelak (2015) or Pawlak, Mystkowska-Wiertelak, and Bielak (2016).Although some of the studies of classroom WTC looked into shifts in its intensity over different periods (cf.Cao, 2006Cao, , 2013;;de Saint Léger & Storch, 2009;Weaver, 2007), most of the investigations offered insight into single lessons or tasks.The present study attempts to join the two perspectives providing accounts of WTC fluctuations in individual lessons as well as those spanning the period of a semester of study.The paper starts with a brief overview of research conducted with regard to factors contributing to WTC ups and downs in the classroom.This is followed by the description of the current study, the presentation and discussion of its findings and, finally, some pedagogical recommendations as well as suggestions for future research.

Classroom WTC
Early questionnaire studies that were mainly performed in the second language context did not at first make the distinction between in-class and out-of-class WTC (cf.MacIntyre & Charos, 1996;McCroskey, 1992).The scale aimed at gauging L2 WTC consisted of items concerning various situations in which the target language (TL) might be used, including those naturally occurring in the second language context, such as talking to a stranger on the bus, a janitor or a garbage collector.Soon it became evident that L2 learners in foreign language contexts may not be able to indicate the level of readiness to engage in communication in response to items that do not reflect the conditions in which they are required to use this language.For example, in the study conducted by Asker (1998), learners of English in Hong Kong found the scale items confusing and asked whether they should "imagine" their behaviour in situations they normally do not experience.MacIntyre, Baker, Clément and Conrod (2001), exploring WTC in the immersion context, used one and the same scale to tap in-and out-of-class WTC, but instructions required a response referring to interaction either in the classroom or outside.The items were grouped around the four basic skill areas and concerned situations that were perceived as typical for immersion students.Additionally, this time no reference was made to interaction with people representing specific professions but rather with "strangers" or "friends."An important modification was made by Weaver (2005), who, trying to identify situation-dependent factors shaping learners' WTC in the classroom, used a scale consisting of items measuring both willingness to speak and willingness to write in a whole range of tasks and situations typically occurring in an L2 classroom, such as doing a role-play, writing a paragraph, or translating a sentence from one's mother tongue.Peng (2007), in turn, relied on MacIntyre et al.'s (2001) scale, but certain adaptations were made to reflect classroom realities the participants knew from everyday practice.Peng (2013) criticized Weaver's (2005) scale for the use of the word someone to denote an interlocutor, on the grounds that it did not ensure a clear reference to the nature of intended communication, since it might imply conversing with another student or a teacher, a difference that might importantly affect learners' WTC.Nevertheless, the study by Peng and Woodrow (2010) made use of selected items from Weaver (2005) to supplement the scale designed by McCroskey and Baer (1985) in order to measure L2 WTC in various tasks differentiating between three types of interlocutors: a teacher, a peer, or a group of peers.A finer-grained perspective was adopted by Khajavy, Ghonsooly, Fatemi, and Choi (2016), who used Peng and Woodrow's (2010) data collection tool to measure classroom WTC, discerning between readiness to engage in meaning-focused and form-focused activities.The foregoing discussion illustrates researchers' constant quest for data gathering tools and procedures, which, still within the quantitative paradigm, would be able to identify factors and conditions pertinent to increasing or decreasing L2 learners' readiness to engage in communication.
A considerably richer pool of data has been accrued in the course of recent mixed-methods or qualitative studies, which will now be discussed.These were undertaken with a view to, first, disclosing a wider range of factors that might impinge on classroom WTC, and secondly, revealing the mechanisms underlying the dynamic shifts in WTC intensity displayed in longer or shorter time spans.Thus, observation and stimulated recall allowed Kang (2005) to discover that the decision to start or enter into communication emerged in the presence of the feeling of security as well as excitement and responsibility.An important contribution was made by Cao and Philip (2006), whose investigation of WTC concerned three interactional situations: pair work, group work, and whole class.The scholars adopted a mixed-methods approach using observations and structured interviews as well as questionnaires to elicit data on learners' classroom WTC.Interestingly, the results did not show a correlation between selfreported WTC and the observed communicative behaviour.A much quoted study by de Saint Léger and Storch (2009) reported the changes in WTC levels in speaking classes as reported repeatedly within a semester, which appeared to be dependent on the respondents' perceptions concerning speaking activities and themselves as learners.Another step undertaken with a view to gaining even deeper insight into what increases or hinders readiness to interact in the classroom was that of Bernales (2016), who, apart from employing surveys, elicited information by means of stimulated recall to account for both speaking and reticence at specific moments during classes that were observed and video-recorded.The main focus of Bernales's study was on the ratio between planned and actual speaking behaviour, which turned out to be dynamic and dependent on learners' goals, motivation, classroom norms as well as teacher expectations.
A completely new phase of the enquiry of the dynamic character of WTC started with the innovative methodological approach advanced by MacIntyre and Legatto (2011), in which specially developed software was used to tap into minute-by-minute fluctuations of a learner's WTC in the course of a speaking activity.The idiodynamic method, as this method was termed, involves indicating levels of WTC while watching a video of one's own performance using a computer mouse, on the basis of which a graph representing WTC ups and downs is created.Then WTC fluctuations depicted in the graph are discussed with a researcher in order to identify the reasons underlying the changes.In MacIntyre and Legatto's (2011) study, which is not classroom-but laboratory-based, readiness to speak faltered due to anxiety and insufficient vocabulary resources.Similarly, Mystkowska-Wiertelak and Pawlak (2014) and Pawlak and Mystkowska-Wiertelak (2015) also report on shifts in WTC intensity, using a different data collection method, which required the respondents to indicate the level of readiness to engage in individual tasks performed in the course of a naturally occurring class they regularly attended.WTC was recorded in response to a beep heard at equal intervals of time (5 min.or 30 sec.), depending on whether an entire class or one specific task is taken into account.Disturbing as the sound must have undoubtedly been, the technique allowed the researchers to depart from a laboratory-style investigation and collect data in the real classroom, which, as was hoped, would render a wealth of information on the intricate interplay of factors and processes underpinning the development of communicative competence in a language class.The same data-gathering tool was again employed by Pawlak, Mystkowska-Wiertelak, and Bielak (2016) and Mystkowska-Wiertelak and Pawlak (2017), who investigated WTC ebbs and flows during whole classes devoted to the development of the speaking skill.The study reported here, which is a partial replication of the studies mentioned above, extends the scope of enquiry to WTC fluctuations reported during classes spanning a whole semester of a speaking course for advanced students majoring in English, with the present author being their regular teacher.The choice of a context-sensitive perspective enables getting insights into ways in which L2 WTC interacts with a whole host of other classroom-embedded factors, both situational (e.g., group dynamics, the teacher, type of task), or individual (e.g., learner preferences or learning style), and how such interactions impinge on the teaching/learning process.

Purpose of the study
The overall aim of the research project was to explore the dynamic nature of WTC among advanced learners attending a speaking class which was part of an intensive English as a foreign language programme in order to identify individual and contextual variables accounting for the ups and downs in the participants' readiness to interact in this setting.Like other empirical investigations of WTC carried out in the classroom context (e.g., Cao, 2011Cao, , 2013;;Peng, 2014), the study reported here can be said to have high ecological validity and thus its findings can be seen as indicative of real-life conditions and serve as a point of reference for classroom practices (cf.Hulstijn, 1997;Lightbown, 2000).The main aim of the study was to document fluctuations in levels of WTC during conversation classes among advanced learners of English over a period of one semester and identify variables causing WTC growth and decrease.More specifically, the study was conducted to address the following questions: 1. What is the extent to which the participants' WTC fluctuates in the course of individual speaking classes investigated for the purpose of this study?2. What is the extent to which the participants' WTC fluctuates in speaking classes in the course of the semester of study? 3. Which context-dependent and individual factors decrease and which increase the participants' WTC in the course of the speaking classes under study?

Participants
The participants were a group of 12 students (5 male and 7 female), enrolled in the final year of a three-year bachelor programme (BA) and majoring in English.
At the time the research was conducted they were on average 21.3 years of age (SD = 1.8) and their mean experience in learning English was 11.75 years (SD = 2.80), which shows that most of them started learning English while at primary school.Generally, their proficiency level could be estimated at B2/C1, according to the Common European Framework of Reference for Languages; however, it needs to be stressed that a wide range of individual variation could be observed among students as well as with respect to particular skills and language subsystems within individuals.The average grade in end-of-the-year EFL examination in the group was 3.6 on the scale from 2 (lowest) to 5 (highest) (SD = 0.71) and, as the study took part in the second semester, the end-of-semester grade in the conversation class was 4.15 (SD = 0.67).When asked to self-evaluate their ability to speak, the skill of particular relevance to the present study, the respondents indicated a slightly lower score at 4.05 (SD = 0.57).Apart from attending an intensive course in English divided into separate classes of speaking, writing, use of English, phonetics and grammar, the students were also required to attend a number of content courses in literature, history, culture, linguistics, language teaching methodology and social sciences, where English was also a means of instruction, which importantly increased their exposure to the target language and increased opportunities for communication.
The conversation classes during which fluctuations in WTC levels were analyzed were all taught by the present author, an experienced teacher of English, over the period of a semester.The topics around which the speaking activities revolved were chosen and ordered according to the syllabus designed by instructors teaching various groups at the same level, in compliance with the curriculum approved by the faculty.Focusing on the development of the speaking skill as the classes understandably were, much heed was paid to vocabulary practice and tests.For this reason, not all of the classes were included in the present analysis.Another reason for which only 7 classes out of 12 were taken into account was the fact that, given the intrusive character of the procedure, it was decided that the students had the right to receive instruction of the ordinary type, without beeps sounded every 5 minutes, so that the study did not negatively affect their learning and, eventually, grades.Due to space limitations the lesson plans will not be presented in detail; however, Table 1 presents a short description of each of the classes containing the general theme and information on the organization mode, type of activity and its approximate duration.UFO Individual or pairs: vocabulary exercises (ca.15 min); video -movie trailer (ca. 5 min); small groups: discussion on ways aliens are depicted in movies (ca. 10 min); whole class: report on pair work (ca. 5 min); pairs: preparing a report on humans that aliens would make (ca. 10 min.);whole class: feedback and summary (ca. 10 min) 6.
Addictions Whole class: video on addictions and comprehension questions (ca. 10 min); whole class: discussion on ways addictions are presented in the media (ca.15 min); pairs: making a list of things people get addicted to (ca. 5 min); small groups: questions on text read at home (ca.15 min.);pairs: discussion on students' own addictions (10 min)

Data collection and procedures
The data collection tools used in the present study included self-assessment grids that required the respondents to indicate on a scale from 1 (not at all) to 7 (very much) how willing they felt to contribute to ongoing activities in reaction to a beep they heard every 5 minutes.The grids were printed on a sheet of paper and included spaces where students provided their names and the date on which the lesson was held.Another way of gathering data were interviews conducted with one randomly appointed student after each of the lessons.Each of the students was interviewed only once and the discussion, which was recorded and transcribed, concerned the reasons for changes in WTC levels depicted in the grid as well as the interviewee's general opinions and comments on the particular lesson.Another source of information were detailed lesson plans, which included comments on the implementation of the plan and the indication of the exact stages of the class when individual sound signals were heard.The students were also required to fill a questionnaire based on the scale developed by Mystkowska-Wiertelak and Pawlak (2016) (Cronbach alpha = 0.89) to tap into the general level of classroom WTC at the beginning and end of the period the study took place.The tool consists of 16 6-point Likert-style items denoting situations that are likely to take place in the classroom in the context of advanced learners of English at a Polish university.Individual items discriminate between communication with a teacher, one peer, small group, or the whole class.Additionally, four items refer to reacting to or providing corrective feedback and one item to the use of the mother tongue.Moreover, the survey included an open-ended question that asked to enumerate factors responsible for a rise or fall of WTC levels the students experienced, first, at the beginning of the course, and next, at the time of its duration.All of the instructions in the grid and questionnaire as well as tool's items were written in the students' mother tongue (Polish) to avoid misunderstanding or misinterpretation, which, in turn, might have led to imprecise responses.For the same reason, Polish was also used during the interviews as well as in the responses to the open-ended question, as it was believed to ensure greater depth of insight and ease of expression.It needs to be explained here that while the classes lasted 90 minutes, the self-report procedure differed in length between lessons, on account of the fact that various administrative and testing actions had to be performed, thus leaving between 55 to 75 minutes to be subject to investigation.It was decided, however, that 55-minute periods of each lesson will be investigated, to enable cross-lesson comparisons.The students were asked to give a consent to take part in the research project and assured that they could withdraw at any moment without any consequences whatsoever.They were also informed that agreeing or not agreeing to participate in the present study would not affect their course grades.However, to encourage them to take part in the research and approach data gathering procedures seriously and sincerely, they were offered some additional credit in the language teaching methodology course that all of them attended.

Data analysis
The present empirical investigation of learners' WTC is a mixed-methods design where both qualitative and quantitative analysis was performed.The former concerns data collected by means of interviews and the open-ended question included in the survey, while the latter refers to the numerical data rendered by self-evaluation performed and indicated on the WTC grid and responses to Likert-type items included in the questionnaire.The qualitative enquiry involved identification and categorization of recurrent themes, whereas numerical data were used to calculate means and standard deviation values for individual beeps and entire classes, which enabled subsequent comparisons.The decision was made not to use inferential statistics with the exception of comparing the results of the administration of the in-class WTC scale.For this purpose ttests were applied to establish whether the difference was significant.The decision not to use inferential statistics with reference to WTC levels was dictated by the exploratory character of the study and the research questions posed that concerned the existence of fluctuations in WTC levels and establishing context-dependent variables causing them, rather than making between-point comparisons or comparing individual classes.The figures representing the mean levels of WTC were plotted against the activities and tasks demonstrated in the lesson plans.The interpretation of WTC ups and downs was performed with the use of the information gleaned from the detailed lesson plans and teacher notes on what the students were doing at the moments the beeps were played as well as commentary provided by students in the interviews conducted after every lesson.

Results
The analysis of the data collected by means of research instruments described above provided evidence for the dynamic character of L2 learners' WTC during single lessons as well as in the course of a semester of study.Some degree of variation was reported within the group and within individuals.Additionally, diversity was also noted in the factors that the students reported as increasing or decreasing their WTC.For the sake of clarity, the discussion will first concentrate on the comparison of WTC levels in the seven classes, which will be followed by the presentation of the results for each of them separately with respect to temporal variation for all of the students as well as the variables contributing to WTC ups and downs that were revealed in the interview and questionnaire administered at the beginning and end of the research.

WTC variation in the course of the semester
As can be seen in Table 2 and Figure 1, the self-reported levels of WTC in the seven speaking classes slightly decreased throughout the semester until measurement 7, at which time the average WTC level increased a little.When average WTC levels are compared, it is evident that the students appeared most willing to contribute to classroom interaction at the beginning of the research project and least willing at the next-to-last lesson.While the WTC means in the first four classes oscillated around 4 or neared 5, as in Lesson 1 (M = 4.91), a decrease was noted for Lesson 6, at which time the participants' level of readiness to interact slightly crossed the mid-point of the scale, remaining at M = 3.89.As evidenced later in the qualitative data, the drop might be ascribed to the overall preoccupation with end-of-the year tests and exams as well as general tension and tiredness, or perhaps the impact of the data collection procedure as such.
Table 2 The means and standard deviations for WTC levels (from 1 to 7) during the seven classes When it comes to the overall patterns of WTC fluctuations in the seven classes, the results show, as can be seen in Table 2 and Figure 1, that the participants' readiness to speak was the most stable during Lesson 6 and most diverse during Lesson 1, in the course of which the highest ratings appeared roughly in the middle of the class and remained at the same relatively high level for 4 consecutive beeps, which spans the period of 20 min.Quite a reverse trend could be observed in Lessons 2, 4 and 7, where, after initially increasing scores, a slight drop was reported with its bottom around Times 6 and 7, only to increase towards the end of the lessons.The lowest value of the respondents' readiness to contribute to the interaction was noted during Lesson 3 at Time 9 amounting to 3.17.The lesson whose onset was marked with the lowest value for learners' readiness to speak was Lesson 4 (M = 3.50); however, the initial trend was reversed until the end of the lesson, when, at Times 9 and 10, the ratings were the highest.An attempt to relate these patterns to the ways in which the classes were conducted will be made in the following section.

Variation in WTC during individual lessons and their causes
Due to space limitations the following discussion will concern only the most striking changes of WTC levels as well as most conspicuous trends evident in the selfreports produced by the participants in each of the lessons.Additionally, a more careful examination will concern the fluctuations connected to the reasons that were especially salient in the post-lesson interviews.The following analysis relies also on the researcher's own observations and notes taken during the classes.

Lesson 1: Feminism
The first speaking class during which the grids were used was devoted to feminism.Responses to the post-lesson interview and the open-ended question indicated that students were not very fond of the topic, but nevertheless their readiness to contribute to classroom speaking activities was the highest from among all of the classes under investigation (M = 4.91).The numerical information about WTC (Table 2 & Figure 1) demonstrate that the participants were generally willing to take part in the tasks and discussions.During the class WTC fluctuated depending on the organization mode and type of task.The first rating was the highest Time 1 score in the whole research reported in the present paper (M = 5.00), and was observed during a quiz intended to help students revise requisite vocabulary.Because of the competitive character of the quiz, the students remained agitated and talkative for the following 15 minutes; however, a moderate downward trend was registered at the next two beeps (M = 4.83 and M = 4.50 respectively) due to repetitiveness of the procedure and perhaps boredom, as was hinted by the interviewed student: "you know, when it went for too long we didn't feel like going on."The biggest increase of 1.0 in WTC was noted between Time 3 and Time 5, which was the highest of all of the measurements in the study (M = 5.50).Time 5 and the following three beeps were heard while the students were working in pairs discussing various questions related to the main topic of the lesson presented as cards A or B. They were also encouraged by the teacher to paraphrase the answer of their partner before proceeding to the next question, so as to make sure that they understood them correctly.This technique turned out to be most conducive to increasing students' readiness to speak.WTC levels dropped at Time 9 (M = 4.50) when students were watching a short documentary about a well-known feminist activist, at which moment they were not actually required to say anything.Activities that do not entail immediate production affect WTC levels in a negative way; however, as elucidated by the student in the subsequent interview, even if no speaking was performed, her WTC levels were initially high because, while watching, she found answers to some of the questions posed before the task: "I felt ready to speak because at first I understood everything and knew what to say."Unfortunately, failing to find answers to all of the questions towards the end of the video, she eventually started to feel less eager to talk, hence, her WTC dropped.The values of selfreported readiness to contribute to interactions slightly increased towards the end of the class when the teacher was asking questions about the film in the whole-class arrangement providing prompts and explanations, which might have positively affected the students' wish to speak.

Lesson 2: Cold case files
Much in the same vein, in this lesson too, activities that did not require immediate production did not drastically decrease the students' WTC.The class reported here revolved round a "cold case," an unsolved murder or suicide mystery.The class started with a short PowerPoint presentation including also a short audio material in which journalists interviewed a lawyer about the case.
The introductory part lasted 10 min during which time two beeps were heard.
Mean values for Time 1 and 2 were relatively high (4.18 and 3.91, respectively) which showed that the participants must have been interested and eager to contribute to classroom interaction without even knowing what tasks they would be expected to perform.Just after the second beep the students were instructed to, first, get familiar with the documents describing the case in detail from numerous perspectives, and then, discuss in pairs the most probable cause of the main character's death.The documents that the students were to inspect included, among others, a coroner's report, a newspaper article describing the victim, interrogation transcripts, documents found in the possession of the victim at the moment of his death.The students were also encouraged to watch the teacher's presentation and the film again.During this phase of the lesson the beeping sound could be heard seven times and the WTC levels indicated by the students showed an increase in comparison to the introductory part of the class reaching WTC highest level at Time 4 (M = 4.82).At Time 9, which was the last measurement during this activity, the mean WTC was slightly lower (M = 3.82), which is hardly surprising given its length and character and the fact that the students worked on their own, only occasionally asking the teacher or other students to explain new words or unknown concepts.The next stage of the class involved discussing with a partner the hypothesis formulated on the basis of the available evidence and, as indicated by Time 10 value, the students' willingness to present their interpretations of the facts was at its lowest, only to rise slightly at Time 11 when a whole-class feedback session was held.This very moment turned out to be the second occasion in the study during which a whole-class arrangement resulted in an increase of the students' WTC, contrary to the students' assertion that this way of conducting classes was most anxiety-provoking and thus hindering WTC.The increase might have been caused by the fact that at that moment students could verify their hypotheses concerning the death of the main character, presenting counterarguments to those demonstrated by others to support their point of view.The task proved to be inherently engaging and the lengthy session of the evidence analysis generated more readiness to share opinions with the teacher and the rest of the group.Interestingly, the highest levels of WTC which continued for Times 5, 6, 7 and 8 were indicated when the students were hardly expected to say anything.As clarified in the interview following the lesson, the students found the material interesting and being given an opportunity to gather information that would be useful during the discussion part of the class inspired more willingness to speak.According to the interviewee, possessing background knowledge on the topic or being provided with information on the issues to be discussed, which naturally entails the provision of useful vocabulary, is critical in provoking engagement and willingness to participate in speaking activities.In the words of the student: "All the difficult words about forensics and so on were there so I didn't need to look them up in a dictionary and I had some ideas who might have done it, killed . . . it was like in a detective movie."The impact of negative affect on learners' WTC was manifested in the student's account of the situations in which he realized that other students knew more about the case or understood better the audio material presented by the teacher.The feeling of lagging behind, in the words of the student, caused "frustration and reluctance to do anything."Additionally, he complained about being "overwhelmed by a mass of information" and the brisk tempo in which the class proceeded.

Lesson 3: Urban myths
As can be seen in Table 2 and Figure 1, Lesson 3 did not generate as much willingness to interact as the previous ones and amounted only to the level of M = 3.92.Although this lack of eagerness to speak might have been caused by some external factors, it seems plausible that the initial activity had a bearing on what happened throughout the lesson.The main topic around which the class discussion revolved was urban myths.At the beginning, the students worked in small groups and discussed questions referring to a newspaper article on people's gullibility, which they had to read at home.This stage initiated a moderate level of WTC at M = 3.75.This authentic text, despite the fact that it had been read as homework, turned out to be challenging, and conversations in groups did not proceed in a lively neither animated manner, which necessitated the teacher's prompting or participation in group discussions.A slight increase was reported for the two following times (M = 4.42 at Times 2 and 3) when students worked in pairs and answered questions provided by the teacher on separate work cards related to the main topic, at which time they could refer to their own experience and opinions.The task that was set later on immediately decreased the students' WTC (M = 3.58) as they were expected to create their own urban myths according to the criteria provided by the teacher.The student interviewed afterwards explained that the drop might have occurred because the learners found the task very challenging and requiring much creativity.This time they were not provided with any materials to rely on but were supposed to produce something almost from scratch.The WTC level remained relatively stable throughout this phase of the lesson.On completing the task the participants created new pairs and interviewed one another on the myths they produced, which resulted in the lowest WTC level (M = 3.17) during the whole class.A slight increase (M = 4.25) was indicated towards the end, when students were engaged in small group discussions concerning frightening events, which entailed a reference to students' own experiences, generally conducive to increased willingness to interact.Additional incentive came from a list of idioms expressing varying degrees of fear that the students were instructed to use in their conversations.The student interviewed after the lesson stressed the importance of "having things to communicate," which he understood as talking about his own experience or things he knew really well without the need to refer to any area of knowledge; thus, an opportunity to talk about things he found scary led to his highest WTC rating.His lowest WTC value was indicated during the activity that required designing a new myth, which, as he complained, "was hard because nothing came to [his] mind."

Lesson 4: Mysteries
Lesson 4, during which the mean WTC value amounted to 4.25 (range between 3.50 and 5.13), concerned the topic of unresolved mysteries and started with a short film shown without any introductory questions and depicting various popular conspiracy theories.The first two beeps, which sounded while the film was presented, both rendered M = 3.50, the exact midpoint on the scale used for self-report.An increase to 4.50 was reported at Time 3 when the students were providing possible explanations of conspiracy theories that they saw in a Pow-erPoint presentation.WTC values remained at a similar level at Times 4, 5 and 6 (4.25, 4.38, 4.23 respectively) during pair work involving answering questions posed by the teacher with reference to a short text on superstitions.The reported WTC level dropped at Time 7 (M = 3.88) during "Hot seat," a game-like activity, necessitating guessing words written on the blackboard behind, using the prompts offered by the rest of the group.Enjoyable as the game was, it must have generated some anxiety in weaker students, who may have felt uneasy at the necessity to perform in front of the group.In the course of the next activity, reading a text on fortune telling, WTC ratings began to grow gradually (4.13, 4.88) to reach its peak at Time 10 (M = 5.13) when the students took part in a task resembling "speed dating" in which they acted as fortune tellers themselves.Asked about the initially modest values reported for willingness to contribute to class interaction, the interviewed student commented that he was "neither willing nor unwilling" but simply waiting for the teacher to "show something interesting."In his view, the increase of WTC during the vocabulary game was only likely among students who felt confident about their lexical knowledge: "You have to know quite a few words to enjoy a game like that."

Lesson 5: UFO
Given the nature of the warm-up phase of Lesson 5, it is surprising perhaps that the opening WTC rating was relatively high, not the highest in the whole study but definitely one of the highest at M = 4.00.At that moment as well as while hearing the three consecutive beeps, the students were doing vocabulary exercises in pairs and in the whole-class arrangement, filling gaps, providing definitions and doing crossword puzzles.Such a high level of WTC testifies perhaps to the students' generally favourable disposition on that particular day, rather than their appreciation of the tasks or their willingness to interact.An alternative interpretation was that the exercises were performed in pairs in the usual sitting arrangement, thanks to which learners could work with partners they knew well and liked.Beep 5 was heard while the students were watching a trailer of a new alien invasion movie, which generated an increase in the self-reported WTC values to M = 4.40.A further upsurge was noted at Time 6 (M = 4.70) when group discussions were held on different ways of depicting aliens in science-fiction movies.Since the genre appears very popular with this audience, most of the students were highly engaged, recalling their favourite scenes in an animated way.General enthusiasm waned, as did the reported WTC level, at Time 7 when the teacher asked group representatives to outline discussions in their groups and then introduced a pair activity in which the participants were supposed to compile a report that aliens might prepare as a result of encountering humans.When this task was in progress the students' readiness to communicate remained at a relatively stable level (M = 4.70, M = 4.60, M = 4.50) but dropped (M = 4.30) during a whole-class feedback session that followed.In the opinion of the student interviewed afterwards, talking about science-fiction movies was the most conducive to WTC mainly because many of the group members had recently seen a newly released film on the topic and the memories were rather fresh.Another factor that according to that student contributed to the increase of her willingness to interact was her partner's willingness to cooperate and good atmosphere during the lesson: "[Name] was today really talkative, she spoke a lot and I didn't have to do all the talking myself."Her lowest WTC rating was associated with one of introductory vocabulary exercises that entailed speaking in front of the group.Knowing that she might be the next person to be appointed by the teacher to perform at the blackboard made her really anxious and unwilling to say anything.

Lesson 6: Superstitions
As shown in Table 2 and Figure 1, fluctuations in Lesson 6 were the smallest from among all the classes the present study concerned.The standard deviation (0.18) was the lowest, as was the mean of WTC ratings (M = 3.89).It is difficult to pinpoint one variable whose impact could be blamed for low WTC ratings, but perhaps a combination of factors contributed to such a situation, beginning with an ineffective introduction: listening to a song to fill gaps in the song lyrics.
The students were encouraged to check their answers with other students.It was evident that the students did not like the song itself and that the task was too easy.This was Time 1 and the rating amounted to 3.75.Exactly the same value appeared at Times 2, 4 and 5 when students were watching a short documentary on the impact of superstitions on people's lives and answered in pairs the teacher's questions checking the understanding.Readiness to engage in communication increased at Times 6 and 7 to the level of 3.88 when students worked in pairs and recalled omens of good or bad luck.An increase to 4.25, the highest on that particular day, at the next beep overlapped with the onset of a "call my bluff" activity and a similar rating (M = 4.13) was reported at Time 9 when the activity was in progress.The final phase of the lesson involved answering questions concerning origins of superstitions on the basis of a short text.The WTC level reported at the two measurements that took place during that period equaled 4.00.A conversation with one of the students that was conducted immediately after the lesson confirmed the teacher's assumption that the group did not like the topic.The materials used in class, including both the song, handouts with questions and texts were considered simple and unmotivating, thus the lesson design failed to generate more enthusiasm and eagerness to talk.The impact of group dynamics was visible in a comment by the interview participant, a usually quiet and shy student, who was more active during the lesson.Asked about the cause of the change, he remarked that it was due to the absence of a student who tended to dominate whole-class activities.

Lesson 7: Addictions
The last class took place at the very end of the spring semester.Not only was the mean at 4.22 higher than that in Lesson 6, so was the standard deviation (SD = 1.59) testifying to more dynamic shifts in the participants' willingness to contribute to tasks and activities taking place during the lesson.The first two beeps rendered measurements amounting to 3.83, the lesson's lowest value, and were sounded while the students were watching a YouTube clip on drug addiction (Time 1) and answering the teacher's questions related to the material they watched.The students WTC fluctuated over the three consecutive beeps above the level of 4.00, which coincided with a discussion of ways in which addicts and addictions are depicted in mass culture.A slight drop to the level of 3.88 was registered at the subsequent measurement when the class were asked to make a list of things people can get addicted to, which was followed by a small group discussion in response to a set of questions provided by the teacher and corresponding to the text the learners were supposed to prepare at home.Group work inspired more communicative behaviour and readiness to talk as became evident in mean values at Time 7 (M = 4.00), Time 8 (M = 4.67), and Time 9 (M = 4.50).The largest value of WTC ratings appeared at Time 10 in the course of the discussion in pairs on students' own addictions.This part of the class, as testified by the student who was interviewed once the lesson finished, was the most engaging because students could talk about their own experiences or share information about family members who suffered from different types of addictions.As one of the participants commented in the interview: "I talked about my brother.He is addicted to video games and we are really worried.He is allowed to play only a few hours a day and when he is not playing he is a normal kid, but he can be aggressive or depressed when playing; his mood changes so quickly."The interviewee enjoyed the video that was used at the beginning of the class to introduce the topic and, as she admitted, was nicely intrigued and waited for the lesson to unwind, which resulted in a surge of her WTC.She reported a decrease in her readiness to interact while answering questions because there were "too many of them" and some of them referred to exactly the same issues.Another aspect of the class that she particularly enjoyed was diversity of tasks and arrangements as well as a quick tempo the class proceeded.

In-class WTC scale
On two occasions, the beginning and end of the study, the respondents were asked to fill in a questionnaire based on the scale developed by Mystkowska-Wiertelak and Pawlak (2016) to tap into the general level of classroom WTC.The mean value at the first administration of the questionnaire amounted to 3.12 (SD = .43),whereas the second time the tool was used rendered a slightly higher mean of M = 3.62 (SD = .43).On both occasions the respondents' readiness to interact in the classroom reached a moderate level; however, on the second occasion the level was slightly higher and the comparison of means showed that the difference was statistically significant (t = 7.44, p < .000).The low value of standard deviation, the same at both measures, testifies to a uniform character of the sample, which is not surprising, given the similar background of the participants, almost identical aims and the fact that the items concerned classes taught by the same teacher.The students' WTC turned out to be higher at the end of the course, which can be accounted for by the growth in their general proficiency level as well as the fact that both the students within their group and the teacher had got to know each other better, thanks to which positive classroom atmosphere was created, increasing the feeling of security and reducing anxiety.
The survey included one open-ended question aimed at tapping individual factors the respondents identified as responsible for increasing and decreasing their willingness to contribute to speaking activities in the classroom.Not all of the multitude of variables that impinge on WTC fluctuations in the classroom could be taken into account: The analysis concerned only those that were mentioned most frequently by the participants.On the first occasion only 5 out of 12 respondents provided answers to the question.Only two aspects of the lesson that could impact the students' willingness to talk were mentioned: uninteresting topics and having to talk in front of the whole group.What positively impacted on the learners' WTC was an opportunity to discuss topics in pairs or small groups, preferably with a friendly, well-known interlocutor.The second time the questionnaire was administered all of the participants answered the question concerning reasons for WTC growth or decrease, and their responses were not as laconic as before.The students declared that they felt reluctant to speak if the topic of the lesson was uninteresting or required having specific background knowledge.Moreover, longer or repetitive tasks, like answering a list of questions, caused a dip in their willingness to interact.Two of the participants complained about the study procedure involving the use of the beeping sound which was distracting and annoying.Most of the students, however, mentioned one factor responsible for reluctance to speak: having to take part in whole-class discussions.This least preferred mode of arrangement provoked most tension and anxiety related not as much to the presence of the teacher as to the fact that all other students were listening.Among factors contributing to the increase of their WTC, the students mentioned a partner who is willing to interact him/herself, preferably a good friend of theirs or someone they like.Positive classroom atmosphere enhanced by the teacher's supportive behavior together with engaging topics that referred to common rather than specialist knowledge as well as variety of tasks were mentioned as most conducive to interaction in the classroom.

Discussion
The aim of the present section is to offer a discussion of the findings of this longitudinal study performed during regularly-scheduled speaking classes taught to English majors by one teacher by addressing the research questions, the first of which referred to the extent to which the participants' WTC fluctuates in the course of individual speaking classes.As can be seen in the analysis of the self-ratings marked on the WTC grids, the students' readiness to speak during all the 7 classes reported here exhibited a considerable degree of fluctuation.WTC ups and downs were reported not only in the course of single lessons but also during specific tasks and activities, as was also established in the course of previous studies into classroom WTC (e.g., Cao & Philips, 2006;Cao, 2011Cao, , 2013;;Mystkowska-Wiertelak & Pawlak, 2014;Pawlak & Mystkowska-Wiertelak, 2015;Pawlak, Mystkowska-Wiertelak, & Bielak, 2016).The magnitude of these fluctuations differed between single lessons but also between individual students, and even if these differences were not conspicuous, as evident in the low values of standard deviation, the impact of individual characteristics cannot be excluded; however, it needs to be noted here that discussing WTC fluctuations of individual students was not the focus of the present study.Some explanation, however, pertinent to the present research question is also tied to the other research questions and thus will be provided in what follows.
The differences in the levels of students' WTC between the speaking classes they attended were addressed by question two.The means for WTC selfratings made on a 7-point scale oscillated around 4 in all of the seven classes under investigation, with 5 classes' ratings slightly above and 2 slightly below this point.This largely uniform trend shows only a mild degree of variation between classes despite the fact that each of the lessons referred to a different topic.However, it needs to be taken into account that each of the classes involved a combination of pair, group and whole-class groupings and that interaction was each time triggered by written text or visual stimuli.The fact that lesson scenarios differed considerably, varying the tempo, sequencing, arrangement and so on, as demonstrated in the results section, turned out not to be decisive in causing greater shifts in WTC levels.The factor that was more influential in this respect was the fact that all of the classes were designed and conducted by the same teacher representing a specific teaching style, thanks to which a consistent environment was created rendering, in turn, a relatively stable WTC ratings over the whole semester.
Another condition whose impact might be held responsible for moderate semester fluctuations is the group composition with its dynamics and power distributions.This aspect is also tied to the third question the present investigation addressed, namely, which context-dependent and individual factors decrease and which increase the participants' WTC in the course of the speaking classes under study.It needs to be stressed here that not all of the context-related and especially individual characteristics lend themselves to an orderly and precise analysis as the characterization of ways in which they operate should take into account the influence of other factors exerted on them.An attempt to separate in the analysis the impact of specific factors seemed futile at times as their effects can only be seen in conjunction with the operation of one or more other variables.Sketchy as the present discussion has to be due to space limitations, it will concentrate on the influences whose impact was perceived as most significant.Moreover, it seems warranted at this juncture to note that there appeared a certain degree of mismatch between the opinions and perceptions disclosed by the students in interviews and the open question and observations made the researcher -the group's regular teacher, as was the case with discussions held not with their regular partners during which they appeared more engaged and lively, despite declaring that their WTC increases with their best friends.Talking to a person one has limited contact with perhaps turned out inherently more interesting that conversing with someone whose point of view and preferences we know really well.
Only some out of the many factors and influences shaping learners' WTC in the classroom setting will become part of the present discussion.The criterion that was used for selecting them was the frequency with which the participants referred to a particular variable.Hence, the present considerations will comprise a reference to class-arrangement modes, interlocutor, topics and tasks as well as stage of the class.As to the first of these, the way the students were supposed to perform a particular task, in groups, pairs, or whole-class, was perceived as the crucial condition in boosting or hindering their readiness to interact.As has been mentioned above, pair work was unanimously considered as the most convenient arrangement.It seems that working with one person at a time provided a feeling of security, which affected willingness to speak in a positive way.By contrast, the need to speak in front of all others in the whole-class mode was perceived as most unpleasant and daunting because of its facethreatening nature, which was exacerbated by antagonistic group dynamics in the class.Directly related to the role of the factor discussed so far is the impact of the interlocutor on the speaker's wish to sustain communication.The interviewed students expressed a preference for working with familiar people, which might be reminiscent of the tension within the group as well as their apprehension that more proficient speakers might dominate discussions thus diminishing their own practice opportunities.The sentiments might have also originated from unfavourable reactions some of them experienced in other classes as well.On the other hand, more proficient learners preferred to cooperate with their regular partners representing a similar level because it ensured an interesting exchange of ideas without the need to stumble over language problems.As was noted by some of the participants and also corroborated by the researcher's observations, highest WTC levels were generated if all/both group members were equally engaged and willing to complete a task.It appeared that the interlocutor's involvement and general positive disposition exerted a positive impact on all of the participants working in a particular organization mode.
Another issue that was frequently evoked by the participants was the topic of a lesson.The respondents agreed that interesting topics enhanced their willingness to speak but were not able to specify what makes a particular theme interesting to them.In the course of interviews it turned out that "interest" boils down to concepts and ideas that students are familiar with, and possess requisite vocabulary and knowledge about.Although the impact of diverse individual preferences was also visible, generally, the topics they considered interesting were those that referred to pop culture, and mass media, were related to everyday existence, and allowed them to draw upon personal experience.By contrast, learners' WTC dropped when boring topics were made the leading theme of a lesson.These included discussions concerning complicated, sophisticated subjects that could not be dealt with without solid factual background, the knowledge of related concepts and the use of advanced language, all of which might impede communication among most of speakers.On the other hand, relying on written sources deprives interaction of its authentic character, since all answers are included in the text everyone has read.
The task the students were supposed to perform also importantly affected their reported readiness to initiate or sustain communication.Closely related to the organization mode, the nature of the task immediately translated into the learners' involvement, resulting in either an increase or a decrease of WTC.Repetitive activities, for example, those involving answering a list of questions in pairs or finding answers while analyzing written material rarely instigated enthusiasm and WTC levels dropped.In turn, game-like activities were preferred by more proficient and less anxious students whose willingness to participate increased.Interestingly, receptive activities, such as watching a video, that did not necessitate immediate production did not adversely affect learners' WTC ratings.Conversely, they either increased or remained at roughly the same level.What was stressed in the responses to the open-ended question included in the survey was the students' appreciation of frequent changes of tasks in the course of one class, which increased the tempo of the procedure, added variety and necessitated changes of partners and arrangements as well as shifts between production and reception.
Apart from Lesson 1, all lessons' self-ratings were lower at the beginning and grew towards the middle.As noted by some of respondents, this was because they were "waiting for the lesson to unwind," not knowing what to expect as each of the classes followed a different scenario.A downtrend was observed towards the end of the classes in all but two of them; however, it needs to be borne in mind that these were not the actual lesson closings, following the initial decision to compare identical time spans in each of the classes under investigation.
Generally, it might be assumed that with time boredom and tiredness set in and, as a result, WTC levels fall.A comment appears in order here that not all of the above-mentioned factors played an equally prominent role in each of the classes; moreover, their impact was not even across group members.What is more, it is difficult, if not impossible, to establish the impact of various factor parameters on the shifts in WTC levels.

Conclusions and limitations
As demonstrated above, classroom-based investigations of willingness to initiate or sustain communication enable gaining a deeper insight into its fluctuations and factors underlying them.The departure from laboratory-style procedures, like those applied by MacIntyre and Legatto (2011), allowed for including a wider range of variables in the investigation and shed light on the ways these variables mutually interact.Moreover, joining the role of the teacher and researcher created opportunities for, on the one hand, shaping the conditions that were investigated and, on the other, conducting observations and enquiring students about their behaviour and opinions almost at the moment they occurred.
Although the study has, to some extent, contributed to the understanding of the changes that willingness to contribute to speaking activities undergoes in the course of language classes in the instructional setting and factors underpinning these changes, it is not free from limitations that need to be addressed here.The first line of criticism can be leveled against the procedure that required the participants to react to a beeping sound at 5-minute intervals throughout the whole lesson unit.Disruptive as it was, which was also confirmed by the students, it offered an opportunity to tap into fluctuations on a moment-by-moment basis (for a discussion of the procedure flaws and merits see Mystkowska-Wiertelak & Pawlak, 2017).Another weakness of the investigation procedure that deserves attention is the fact that in-class WTC surveys were not administered after every class and only one student was interviewed each time.However, the decision was taken not to disrupt the usual lesson schedule, not to use up lesson time, otherwise compromised, neither require students to be engaged in data collection instead of enjoying breaks.Other logistical problems concerned students having to attend other classes and the teacher conducting others.
Naturally, the present discussion did not take into account all of the variables that impact on in-class WTC, offering just a glimpse at the intricate interplay of situated WTC antecedents.The analysis of the data would have undoubtedly benefitted from juxtaposing the identified trends and regularities against individual differences variables such as personality, learning styles or language learning strategies, the more so that, even if some remain stable, certain aspects might be developed or transformed thanks to teacher intervention.Undoubtedly, the study of communication-related variables deserves further attention, especially in the context they naturally occur before sound pedagogical recommendations are offered.Nevertheless, some tentative suggestions seem in order here, especially those concerning organization modes and group composition.It appears that willingness to perform communicative activities grows in the presence of individuals that students like and trust.The most nerve-racking arrangement proved to be the one with the biggest audience -whole-class.However, contrary to students' declarations, when the activities performed in this mode necessitated genuine exchange of information, they turned out effective in increasing WTC levels.That is why teachers should pay special attention to the ways in which classroom interaction is organized and students are grouped.Moreover, it seems that warm-up or introductory activities can shape the mood for most or all of the class, starting either an upward or a downward trend.Hence, a lively, invigorating, funny, or intriguing opening activity can make students more willing to work during the class.What also appears critical is the opportunity on the part of the students to prepare for the speaking phase, hopefully, in the course of engaging in listening or reading activities, during which requisite vocabulary is provided or revised and factual information is offered.

Table 1
Topics and tasks performed in individual classes and their duration