The Assessment of Speaking Skills at the Tertiary Level

Since “English has become the lingua franca for academic interaction of learners and academics” (Koo, 2009, p. 77), the development of the EFL learners’ oral performance proficiency constitutes the central interest of current English language teaching methodologists and practitioners. The importance of speaking as a productive skill has been echoed in the literature. Indeed, it is viewed as a crucial “part of the curriculum in language teaching ... and ...an important object of assessment as well” (Luoma, 2004, p. 1). Thus, the prime aim of this study is to explore the prevailing conceptions and actual practices of the assessment of EFL learners’ speaking skills at the tertiary level. The respondents of the current research were 20 instructors who taught at the Higher Institute of languages in Gabes and at the Faculty of Arts and Humanities in Sfax, Tunisia. To collect the necessary data, a questionnaire survey was utilized. Then, the obtained data were analyzed using SPSS package. The findings revealed that the teachers’ conceptions of assessment are directed towards the development of the learners’ speaking skills. Despite the existence of a number of hardships, the teachers’ classroom teaching practices revealed a compete reliance on authentic, ongoing, organized and thoughtful oral language assessment procedures which were meant to sustain and boost the learners’ oral skill achievements.


Introduction
In the 20th century, assessing the learners' oral production is the central concern of various scholars (Kang, 2013;Celce-Murcia, 2013;Louma, 2004).It has become viewed as "an essentially interactive process, in which the teacher can find out whether what has been taught has been learned, and if not, to do something about it" (William, 2007(William, , p. 1054)).According to Shepard (2000), classroom assessment refers to the "kind of assessment that can be used as part of instruction to support and enhance learning" (p.4).In this respect, the central foci of current EFL instruction practices of assessment tend to go beyond the traditional standardized forms of testing the learners' competencies to a more reformed communicatively-oriented assessment paradigms.What keeps the issue of assessment luring for researchers is its influence by several factors such as the instructors' pedagogic and methodological ways, their major beliefs and prior knowledge, and the main purposes that urge them conduct assessment.
Thus, it is vital to study the teachers' conceptions of speaking assessment in order to explore the wide range of employed assessment strategies which might help their learners to demonstrate their speaking proficiency.

Objectives
The present research paper seeks to unveil different aspects related to the assessment of speaking skills.It tries to display the interconnection between performance assessment and speaking skill development.

The Importance of Speaking Skills
With the advent of communicative language teaching, research on the development of the learners' level of speaking proficiency has gained ground among current educational practitioners and linguists (Nakamura, 1993).The mastery of speaking skills is viewed as a necessity for most EFL and ESL learners (Richards, 2008, p. 19;Florez, 1999).In an attempt to define the concept of speaking, Richards (2006) argues for the "natural language use" which occurs "when a speaker engages in meaningful interaction and maintains comprehensible and ongoing communication despite limitations in his or her communicative competence" (p.14).Richards' definition of speaking echoes the interactional aspect of language that is used to fuel social intercourse (Yule, 1989, p. 169), to promote emotional intelligence (Bora, 2012), to fulfill certain goals or to convey particular ideas, intentions, desires, preferences, or opinions.In the same line, Zaremba (2006), Celce-Murcia (2001) and Kayi (2006) stress the crucial role of speaking in maintaining oral communication.Seen from this perspective, speaking helps strengthen interpersonal ties especially when it is used appropriately in social interactions.According to Bygate (1987), it is viewed as "the vehicle par excellence of social solidarity, of social ranking, of professional advancement and business" (Bygate, 1987, p. vii).
The intricacy of speaking is related to a set of competencies which can be determined by the "…the ability to fill time with talk…the ability to talk in coherent, reasoned and semantically dense sentences… the ability to have appropriate things to say in a wide range of contexts"; and the ability to "be creative and imaginative…in language use" (Richards, 1990, p. 75).
A part from the interactional aspect of speaking, other scholars stress the use of language for transactional purposes (Nunan, 1991, p. 42;Basturkmen, 2002, p. 26).According to Richards (1990), the transactional uses of language refer to situations "in which language is being used primarily for communicating information" (p.54) such as in broadcasting news or asking someone for directions on the street.The difference between the interactional and transactional purposes of speaking is clarified by Thornbury & Slade (2007) as follows: "primarily interactional language is primarily listener-oriented, whereas primarily transactional language is primarily message-oriented" (p.20).
Since speaking is considered as "the most basic means of human communication" (Celce-Murcia, 2003, p. 103), attempts to link both the interactional and transactional usages have been argued for by Brown and Yule (1983, p. 3) and Kingen (2000, p. 218).According to Kingen (2000), speaking helps fulfill the following twelve functions: 1).Personal -expressing personal feelings, opinions, beliefs and ideas.
3).Narrative -creating and telling stories or chronologically sequenced events.4).Instructive -giving instructions or providing directions designed to produce an outcome.5).Questioning -asking questions to obtain information.6).Comparative -comparing two or more objects, people, ideas, or opinions to make judgments about them.7).Imaginative -expressing mental images of people, places, events, and objects.8).Predictive-predicting possible future events.9).Interpretative -exploring meanings, creating hypothetical deductions, and considering inferences.10).Persuasive -changing others' opinions, attitudes, or points of view, or influencing the behavior of others in some way.11).Explanatory -explaining, clarifying, and supporting ideas and opinions.12).Informative -sharing information with others.
Besides, various researchers have investigated the particularities of speaking as distinct from other macro English skills.They accentuate on the characteristics of oral discourse such as the use of formulaic expressions, hesitation devices, repetition and rephrasing (Hughes, 2002, p. 76); the use of contractions, reductions or elisions; the use of appropriate pronunciation, idioms, stress, rhythm or intonation (Brown, 1994).Along with Brown' ideas, Harmer (2007) indicates that: "Speakers have a great range of expressive possibilities at their command.Apart from the actual words they use, they can vary their intonation and stress which helps them to show which part of what they are saying is most important.By varying the pitch and intonation in their voice they can clearly convey their attitude to what they are saying, too; they can indicate interest or lack of it, for example, and they can show whether they wish to be taken seriously" (p.53).
Other oral discourse features are manifested through the accompanied use of body gestures as expressive language (El Menoufy, 1997, p. 10;Harmer, 2007, p. 53); the management of turn taking (Cornbleet & Carter, 2001, p. 27;Harmer, 2007, p. 53), the resort to a particular type of speech (formal vs. informal) depending on the setting; the possibility of being involved with more than one interlocutor; the use of a range of communication strategy, the need to take into account time constraints and language functions.Thus, the various parameters that interfere in the process of delivering an output render speaking "the most demanding of the four skills" (Bailey & Savage, 1994, p. 7).
In language teaching, the process of speaking assessment is considered as very contentious and intricate (O'Sullivan, 2006).For this reason, a number of scoring instruments are presented for the evaluation of students' oral performances such as the use of checklists and the two dominant types of rubrics which are the analytic and the holistic ones.

The Crucial Role of Performance Assessment
Within the realm of CLT, oral performance assessment "arises spontaneously from the naturally occurring classroom environment and leads the teacher to a judgment about an individual student's level of development" (Stiggins & Bridgeford, 1985, p. 273).Consequently, communicative teaching and communicative assessment has become interdependent (Bachman, 1990).According to Stiggins & Bridgeford (1985), performance assessment is manifested through "the observation and rating of student behavior and products in contexts where students actually demonstrate proficiency" (p.273).As such, it plays an essential role in providing "feedback on performance to improve and accelerate learning" (Nicol & Milligian, 2006, p. 64).Therefore, various researches stress the utility of performance assessment in making decisions about the students' progress, and in obtaining information about their own learning process (Jones & Tanner, 2008;Murray, 2006;Nitko & Brookhard, 2007;Stiggins, 2008).According to the National Research Council (1996), assessment and learning "are two sides of the same coin" (p.5).Consequently, assessment as learning emanates from the idea that learning involves the students into an active and interactive process of cognitive restructuring (Earl & Katz, 2006).For this reason, its implementation relies on a number of interactive classroom activities for the production of a consistent and meaningful output (Luoma, 2004;O' Malley & Pierce, 1996).Hymes (1974, in Buck, 2001) maintains that the assessment of oral skills is ensured when learners practice producing the target language appropriately and meaningfully in a variety of social situations (p.83).In this respect, assessors can draw upon the use of: authentic, or real-life situations and activities that require communication and that are relevant to the lives of the learners-role-plays, games, interviews, problem solving activities, and the like-.Teaching communicatively also means that students learn about the language -its rules and exceptions-which are essential if students are to accurately and efficiently use the language.In other words, a communicative approach concentrates on developing the learner's ability to communicate effectively and views grammar study as just one of the vehicles that can be used to promote communicative competence.(Flaitz, 2000, p. 4) In performance-based assessment, the classroom activities incite learners to reveal their fluent speech delivery as they focus on the spontaneous and meaningful use of language, promote communication, emphasize the resort to communication strategies, manage unpredictable language output and be relevant to the context of language use (Richards, 2006, pp. 13-14).The choice of speaking activities intersects with the rating scales in order to define the kind of oral language skills or abilities that the tasks have to elicit (Luoma, 2004).
In this way, assessment is "an ongoing process that involves the student and teacher in making judgements about the student's progress in language using by non-conventional strategies" (Heurta-Macias, 1995, cited in Richards & Renandya, 2002, p. 337).Indeed, it serves to improve instruction, enhance students' performance and develop their autonomy (Bostwick & Gakuen, 1995;Cohen & Hill, 2000).

The use of Rubrics as Assessment Tools
In performance-based assessment, the assessment of the learners' speaking abilities can be measured through the use of rubrics.According to Arter & McTighe (2001), rubrics measure learners' performance in a way that goes beyond the information gathered by the traditional standardized testing.They help maintain consistency in the assessment process.In the same vein, Lazear (1998) argues for the positive and formative use of assessment through the use of rubrics.The latter become to be viewed as instructional "teaching tools that support student learning" (Andrade, 2003).By enhancing the learners' self -efficacy (Zimmerman & Martinez-Pons, 1990), rubrics enable learners to develop their higher order thinking skills (Huba & Freed, 2000;Andrade, 2001).In this respect, they boost their accountability for their own learning.Flynn & Flynn (2004) argue that "with the use of a rubric, students are held accountable for their work" (p.5).Therefore, rubrics "blur the distinction between instruction and assessment" as "they serve the purposes of learning as well as of evaluation and accountability" (Andrade, 2003).

Analytic Rubrics
Analytic rubrics give teachers the opportunity to deconstruct the learners' performance into separate and different dimensions.Identifying which criteria to be assessed helps "test developers to be clearer about what they are requiring of test takers and raters, and to think through the consequence of such requirements" (McNamara, 1996, p. 45).Then, the total score is obtained by gathering the several individual scores (Moskal, 2000).By covering various aspects of oral performance, they depict speaking assessment on its "multidimensional level" (Mertler, 2001).As such, their implementation in performance-based assessment is relatively time-consuming (Thornbury, 2005).However, analytic rubrics offer detailed, focused and precise assessment.They keep the instructors' and the learners' attention focused on the pre-established aspects that have to be assessed (Arter & McTighe, 2001).
According to Nitko (2001), analytic rubrics help students get specific feedback on their produced output.Accordingly, they help keep track of the learners' speaking proficiency development (Andrade, 2005;Moskal, 2000), provide information about their own areas of strengths, weaknesses and learning needs (Moskal, 2000;Andrade, 2000), incite learners to ameliorate their future oral performances and promote students' learning (Arter & McTighe, 2001).
What is more important about analytic rubrics is that they provide consistent scoring and offer high fairness, validity and reliability in conducting speaking assessment (McNamara, 1996;Thornbury, 2005).Therefore, the use of rubrics "adds to the quality of assessment" (Jonsson & Svingby, 2007, p. 131) and is typically implemented for formative purposes.

Holistic Rubrics
Holistic rubrics are scoring guidelines upon which raters base their general and overall impression of the level or quality of the learners' speaking proficiency.As such, they depict assessment on its "unidimensional level" (Mertler, 2001) as raters do not evaluate each separate criteria aside (Nitko, 2001).Instead, they yield a single grade that is used to measure the learners' oral language skills as a whole.As no detailed feedback is sought, Thornbury (2005) describes the use of holistic rubrics as "quick" and convenient in informal testing formats (p.127).Besides, evaluators find holistic rubrics as feasible and practical with a large number of students.Their time saving feature can also be derived sometimes from their generic nature as they can be applied to a range of classroom tasks.The disadvantage of holistic rubrics is that they are not suited for formative assessment as they do not provide specific feedback on areas that need improvements in students' oral productions.

The Study of Temporal Variables
The assessment of the learners' oral performance can be carried out by measuring the learners' speaking fluency through the examination of a number of temporal variables that define the learners' speaking fluency (Kormos & Dennes, 2004;Jamatlou, 2011;Zellner, 1994).The temporal characteristics of speech comprise an analysis of speech rate, articulation rate, total number of silent pause per minute, mean length of silent pause or length of runs, total number of pause filler per minute, total number of disfluencies per minute.1) Speech rate: measured in words per minute or syllables per minute.
2) Pauses: measured by the average length of pauses (filled and unfilled) and by the percentage of pause time in relation to the total time used by the speaker.
3) Length of utterances: measured by mean length of utterances (MLU) and by the mean length of fluent runs in between pauses.4) Retracing: measured by the number of retracings per 100 words, either without modification (simple repetition) or with repetition (self-correction).
Concerning the disfluency markers, assessors might look at the use of filled pauses, drawls, repetitions and false starts (Grosjean, 1980;Raupach, 1980).Each will be clarified as follows: 1) Unfilled pauses refer to "a period in time in which no acoustic signal occurs for at least 200/270msec" (Hargrove & Mc. Garr, 1994, p. 109).
2) Drawls refer to instances where speakers articulate certain syllables which contain vowels in a prolonged way.This lengthened syllable "can be used as time-gaining devices for the planning of subsequent speech units" (Faerch & Kasper, 1983, p. 215).
3) Repetitions "involve repeats of one or two words followed by the completion of an utterance usually occur after the first or second function word of a constituent beginning an utterance to allow the speaker time to plan for the next part of the utterance-usually a content word" (Fathman, 1980, p. 84).
A part from the quantitative analysis of temporal variables, there are four types of qualitative fluency measures (Osborne, 2008).They include: 1).Propositional content: measured by the number of information units per minute, per 100 words or per utterance.An additional measure of "granularity" (Noyau et al., 2005) indicates how much detail a speaker provides, by looking at the number of micro-events in relation to the number of macro-statements.
2).Syntactic density: measured by the number of syntactic units per 100 words.The number of syntactic units per utterance can also be calculated, to give a measure of "condensation" 3).Vocabulary range: measured by Vocd (Malvern & Richards 1997) and by the proportion of words used which fall outside the first 2000 word frequency band.4).Accuracy: measured by the rate of errors (errors per minute or per 100 words).
Most studies report that when learners' proficiency in a language increases, their speech rate increases, their silent pause rate decreases, and the duration of their silent pauses decreases (Cucchiarini, Van Doremalen,& Strik, 2010;Trofimovich & Baker, 2007).
The audio -collected data are recorded, analyzed using a software program which represents the digitalized speech waveforms.Traditionally the T-Unit was employed in the analysis of speech data.It represents one main clause with all subordinate clauses attached to it" (Hunt, 1965. p. 20)."Although this measure has been widely used for the analysis of second language writing and spoken production, it has also been criticized against other more reliable measures such as a sentence bases analysis" (cf.Bardovi-Harlig, 1992).Consequently, other alternative forms of speech analysis units have been employed in the analysis of the learners' output such as Foster et al.'s Speech Analysis Unit (2000) (referred to as AS Unit).

The Research Method
The present research study seeks to provide a qualitative examination of the university teachers' conceptions concerning the assessment of their learners' speaking performance.To this end, it describes the instructors' current perceptions of assessment and explores the perceived impact of their assessment methodology on the learners' speaking skills.

Research Participants
The contacted participants for this research were enrolled at the Higher Institute of Languages in Gabes and at the Faculty of Letters and Humanities in Sfax, Tunisia.The selection of the previously mentioned two universities was especially motivated by personal reasons.As I was a former scholar at those two universities, having access to data collection was feasible and easy.Since the major objective of the study was to investigate various aspects of speaking assessment, I narrowed the scope to include only instructors who taught 'oral skill' sessions.The total number of the participants included in this study is twenty.More information about their gender, age and the years spent teaching oral skills will be revealed in details when presenting the research findings.

Data Collection Instrument
The present study relied on the use of a questionnaire as the main research instrument.It included 10 questions that were classified into two major parts.The first part elicited demographic information about the participants including their gender, age and their professional teaching experience in conducting oral skill sessions.The second and final part comprised information about the teachers' conceptions and their actual practices of speaking assessment.In this second sub-part, the teacher informants were encouraged to reflect on the motives that urged them to conduct speaking assessment, the different classroom tasks that they had frequently employed, the type of assessment that they relied on, the major source of speaking assessment, the diverse measures that they drew upon and the challenges that they faced.
In order to ensure the reliability of the used research instrument, The Cronbach's Alpha was calculated using Statistical Package for Social Science (SPSS) software and the result revealed α=0.80, indicating a high level of internal consistency.

Data Collection Procedure
Prior to the administration of the questionnaire survey, the researcher got a list of the teachers' names and the subjects they taught (from the administrative staff) so that she could get in touch directly with the instructors of "oral skill".The process of delivering the questionnaire lasted approximately two months for the researcher had to contact each teacher during his/ her working days.The contacted teachers were informed of the nature of the research.The ones who voluntarily approved to be included in this study filled in the questionnaire and their names were kept anonymous for confidential reasons.

Data Analysis Method
The informants' responses to the questionnaire questions were run into SPSS.The analysis yielded statistical results that helped in the interpretation and discussion of the obtained information.The each question of the questionnaire survey gave teachers the opportunity to express themselves freely about what they wanted to add.Thus, such answers were also very helpful in presenting a descriptive and qualitative analysis of the research findings.

Research Findings and Discussion
The analysis of the questionnaire findings reveal some demographic information about the informants' profile and a clarification of the teachers' beliefs, attitudes and classroom practices of speaking assessment.Each will be clarified as follows.

Displaying Demographic Information about the Informants' Profile
The first part of the questionnaire seeks to display demographic information about the informants' profile by drawing upon their gender, age, and their experience with conducting oral skills sessions.It has been demonstrated through the analysis of the questionnaire findings that 60% of informants are provided by female teachers while the rest, 40%, constitutes the percentage of male teachers participating in the present research.
Then, the following question of the questionnaire depicts data pertaining to the informants' age.The results reveal that the instructors' age can be classified into three main categories; a category that represents those who are less than 30 years old, another category that shows those who are aged between 31 and 40 while the final category that indicates those who are aged more than 40.Based on the questionnaire findings, the majority of informants belong to the second category with a percentage that reaches 55%, a weaker percentage goes for the ones who are aged more than 40 (30%) while the weakest percentage represents the informants who are less than 30 years.Therefore, the research finding highlights a number of informants of mixed age groups with an overall majority for a relatively young group of EFL university teachers.Concerning the teachers' professional teaching experience of oral skills, the results demonstrate that 65% of informants have had more than five years dealing with oral skills sessions while only 35% of informants have had a relatively short experience with conducting oral skill sessions.Thus, the research finding depicts an overwhelming group of informants whose teaching experience of oral skills is relatively long.

Investigating the Teachers' Beliefs, Attitudes and Classroom Practices of Speaking Assessment
After gaining some demographic information about the participants of this study, the second part of the questionnaire seeks to reveal other aspects about the informants' deeper knowledge and actual classroom practices of speaking assessment.
The first sub question of the second part reveals the major motives that urge teachers to conduct oral assessment.
As it is indicated in the table below, none of the informants opt for the assessment of speaking for testing the efficiency of their own teaching pedagogy.Another weak percentage of informants (15%) resort to the assessment of the learners' output for the sake of providing their learners with grades.The two highest percentages indicate the rate of informants who utilize oral assessment in order to improve students' learning (40%) and more importantly to support them enhance their verbal performances in the future (45%).Based on the findings indicated above, the majority of EFL teachers accord a great importance to their students' learning process and progress.As such, the Tunisian teachers' interest in using assessment to excel their learners' speaking capabilities and boost high levels of learning echoes similar results obtained from other researches (Sparks, 2005;Black & William, 1998;Murray, 2006;Vardar, 2010;Bennet & Gitomer, 2009).
Knowing the teachers' real motives behind carrying out assessment helps in understanding the diverse types of classroom tasks that are actually being under use.The findings related to the second sub question of the second part reveal that the teachers' responses are in favor of the use of oral presentations, debates and role plays as the major speaking tasks.The latter highlights a complete reliance on interactive, communicative and fluencyenhancing activities as a platform that facilitates for educators the continuous assessment of the learners' output.Consequently, the reliance on productive tasks urges learners into active involvement and communicative interactions with one another.In the same line, Hatcher & Bringle (1997) argue for the effective use of classroom activities for the purpose of assessment.According to them, the classroom activities have to "(a) link experience to learning objectives, (b) are guided, (c) occur regularly, (d) allow feedback and assessment, and (e) include the clarification of values" (p.157).Therefore, The nature of the covered classroom activities revolve around the implementation of oral performance based tasks which stir the learners' motivation, supports their metacognitive skill development (Stiggins, 2002), enhance their critical thinking (Facione, 2011)and thereby increase their communicative competence.
Through the implementation of such activities, the teacher informants want to create opportunities for the English language to be "used for the purpose of communication, in a particular situation and for a particular purpose" (Hymes, 1974) notably that their students are considered as foreign language learners due to their limited exposure to English outside the classroom (Zhang, 2009).In this way, they value "not what a person knows about the language, nor how grammatically correct they are, but whether they can actually use it to communicate in the target language use situation" (Hymes, 1974, cited in Buck, 2001, p. 83).Therefore, the classroom assessment process at the tertiary level reveals the teachers' authentic practices of assessment as learners are evaluated based on their oral performances on a number of communicative tasks.In the vein, Bachman and Palmer (2010) stress the authenticity of assessment activities when they reflect real life situations.As it is stated by Bachman (1990), there should be an interconnection and interdependence of communicative teaching and communicative assessment.
As opposed to the findings reached by Li (1998) who argues that teachers face serious troubles in assessing their learners' speaking skills based on authentic interaction, this research study highlights that within the current CLT approach, the teachers demonstrate their skillful abilities in handling classroom assessment based on authentic, meaningful, relevant and natural-like contexts for language use.Therefore, the careful choice of oral performance tasks mirrors Stiggins ( 2008)'s ideas that assessment does help teachers make instructional and educational decisions.Through the classroom management of appropriate and relevant tasks, the quality of assessment is guaranteed.
In the third sub question of the second part, the informants are asked to reflect on which types of speaking assessment they resort to for the evaluation of their learners' verbal output.As it is depicted in the table above, the findings reveal that the overall majority of teachers (70%) show their inclination and interest in the implementation of both types of assessment: the formal and the informal one.The informants' choices can be justified by the merits that they gain when applying each type.
As far as the informal oral assessment is concerned, the teachers keep track of the learners' level of speaking proficiency and progress through their participations in oral classroom activities and through their own collaborative interactions with their peers.For this reason, DiRanna et al. ( 2008) argue that "ongoing and systematic use of formative assessment and formative feedback is critical to reduce existing learning gaps among student population".Consequently, teacher informants in this research integrate assessment within their actual teaching practices.A similar finding is also echoed by a number of researchers who regard assessment as part and parcel with instruction (Segers, Dochy, & Cascallar, 2003;Cohen & Hill, 2000).The importance of conducting ongoing classroom speaking assessment challenges the traditional standardized forms of testing and is viewed instead as a monitor of learners' performance (Airasian, 1997;Surgenor, 2011).Through frequent assessment practices, teachers scaffold their students in their learning process.Indeed, they provide them with "the systematic evaluative appraisal of an individual's ability and performance in a particular environment or context" (Payne, 2003).As such, teachers succeed in helping their learners notice their strengths, remedy their weaknesses and be effective assessors.
Concerning the formal oral assessment, the informants invite their learners to sit for an oral test within the semester and an oral exam by the end of the semester.The oral exam is administered before a number of juries who determine the students' level of speaking proficiency.In this case, it should be noted that the implementation of formal assessment is basically imposed by the requirements of the examination system whose objective is to provide a final score or a grade to the assesses' oral performance.As such, the outcome of formal assessment is meant to improve the quality of assessment decisions.Consequently, the assessment of the learners' output for the sake of assigning a grade or a final score is especially motivated by accountability and administrative reasons.Thus, the existence of formal oral assessment has a "gate-keeping function" (Davies et al., 1999) as it can be classified within the high stakes testing upon which the test takers' success or future study plans depend.
Therefore, the assessment of speaking in a number of Tunisian universities has both summative and formative aspects.In Brief, it should be noted that the main objective of informal oral assessment is to support and sustain the learners' speaking development while the aim of formal assessment is to articulate the learners' oral proficiency into grades.
In the third sub question of the second part, information about the major source of oral assessment are elicited.
The findings indicate that a number of teachers consider themselves as the sole providers of assessment, others believe in the potentials of their learners in assessing themselves and their peers efficiently, while another group of teachers admit that the assessment of speaking is a mutual responsibility of both the teacher and their learners.Based on the statistics mentioned in the table above, 45% of informants consider that the assessment process is a shared responsibility between them and their students.Other 35% of informants believe that classroom assessment should emanate from learners and only 20% of informants admit that they are the ones who ensure classroom assessment.Based on the achieved findings, the majority of teachers collaborate with their learners in the construction of classroom assessment during the oral expression sessions.A similar finding in reported in the works of Jandris (2001) who argues that "the heart of assessment is a continuing process in which the teacher, in collaboration with the student, uses information to guide the next steps in learning" (p.4).In the present study, the informants tend to provide their learners with opportunities to conduct assessment by themselves.They are also determined to develop effective thinkers who are able not only to assess their own oral productions but also their peers' output.In this way, their efficient assessment practices transform learners into expert assessors (Earl & LeMahieu, 1997, pp. 149-167).Through the encouragement of self assessment and peer assessment, they give their learners insights about how they are capable of enhancing their performance (Mory, 1992) and how they can be "cognitively, motivationally, and behaviorally active participants in their own learning" (Zimmerman, 2001, p. 5).
Seen from this perspective, the assessment of speaking skills is viewed as "an ongoing process that involves the student and teacher in making judgments about the student's progress in language using by non-conventional strategies" (Huerta-Macías, 1995, cited in Richards & Renandya, 2002, p.337).Based on the responses of the teacher informants, the Tunisian EFL learners are fully involved into the process of oral classroom assessment through the enactment of self assessment and peer assessment.For a number of researchers, the reliability of the learners' engagement within the assessment process can reach that of their tutors (Bachman & Palmer, 1989;Blanche, 1990).
Therefore, the Tunisian students' participation in the evaluation of their oral performance is in total contradiction with other research studies which argue for the passivity of learners in the assessment process and the lack of their classroom interactions (Jackson, 2002).
Placing special emphasis on the development of autonomous learners as assessors does not entail the relegation the teachers' roles but rather emanates from thoughtful decisions taken by teachers in their attempt to make the main objectives of the Communicative Language Teaching approach go along with their assessment practices.When implementing speaking assessment tasks, the informants' roles can be summarized, following Brown's terms, into prompters, participants and feedback providers (2001, pp. 347-348).In fact, the informants regard themselves as creators of efficient grounds upon which assessment occurs in a smooth, supportive and authentic way.Within this process, they act as the learners' "co-partner and guide in a common enterprise" (Dewey, 1964, p. 10).Thus, the assessment of speaking at the Tunisian tertiary level has a collaborative aspect as it accentuates on the cooperative construction of classroom evaluations that are shared among teachers and learners.
Concerning the measures that are used for the assessment of the learners' speaking abilities, the results pertinent to the fifth sub question of the second part reveal that teachers rely on a variety of techniques which are the use of verbal feedback, rubrics and technology-based instruments.As it is depicted in the table above, the teachers' distribution along the various measures differ.A very weak percentage of teachers acknowledge the use of technology mediated assessment for the measurement of the learners' speaking.Examples of technology-based assessment, teachers report to have used digital chats, recordings of the learners' speaking performance, and video scripts commentary.Indeed, such tasks "represent authentic language use" (Jamieson, 2005, p. 233) and provide "an excellent capability to trace test takers' language development thus enabling researchers to better understand how aspects of the construct evolve across different ability levels" (Dhalhoub-Deville, 2001).Other researchers argue for the use of e-portfolios as effective multimedia tools for the assessment of the learners' oral language proficiency (Cepik & Yastibas, 2013).According to Gray (2008), the resort to e -portfolios can be defined as "a purposeful aggregation of digital items-ideas, evidence, reflections, feedback, etc., which presents a selected audience with evidence of a person's learning and/or ability" (p.7).Besides, Brown (2008) indicates that the resort to Facebook and Twitter can lay the grounds for oral speaking assessment to be held.
Despite the availability of various technology based tools for the assessment of speaking, only a weak proportion of teachers make use of them.This might be due to the difficulties and the hardships encountered when assessing the learners' oral skills through technology (Levy & Stockwell, 2006).
Another group of informants (35%) believes in the use of verbal feedback as an assessment strategy that is employed to help learners' monitor their oral language progress.Through the implementation of sufficient practice and opportunities to speak in the classroom, the reliance on oral feedback strengthens classroom interactions, develops communicative strategies and sustains thereby informal oral language assessment.The frequent contact between EFL teachers and their learners renders the assessment of speaking an ongoing process that thrives through the use of English as a medium of communication.Similarily, MacGregor (2000) stresses the ongoing, frequent and continuous nature of speaking assessment.
In addition to the use of technology-based assessment and the reliance on verbal feedback, the majority of teacher (55%) rely on the use of rubrics particularly analytic rubrics for the detailed assessment of their learners' output.Despite the varying attitudes other teachers at the university level may hold towards the use of rubrics (Reddy & Andrade, 2010), all the informants who report to have used rubrics maintain that they rely mostly on analytic rubrics in an attempt to provide adequate assessment results.Therefore, the teachers' equipment with measurement instruments for the assessment of their learners' oral proficiency is very necessary (Nagata, 1995).
In a previous research, Li (1998) maintains that the Korean English teachers do not give importance to issues of validity and reliability when performing speaking assessment.However, the present paper reveals the university teachers' concern with fixing some dimensions of oral language upon which the learners' speaking performances are assessed and in trying to make their students' familiar with the set of criteria determined in advance.In order words, the use of rubrics is very pertinent in assessing the "process, performance and progress" (Duhaney, Salend & Whittaker, 2001, p. 9).
Because of the complexity of speaking as a productive skill (Richards & Renandya, 2002, p. 201), educational practitioners and researchers see the need to deconstruct speaking into a set of essential criteria in order to provide an adequate assessment of the learners' output.
Based on the results of the questionnaire, the informants differ in identifying the criteria upon which the learners' oral performances are judged.Some informants mention that they include in their analytic rubrics the following criteria: 1) Fluency-accuracy-organization of ideas-interaction-vocabulary use.
The variety in terms of the criteria that are chosen for speaking assessment reflects a similar confusion among linguistics in their attempt to identify oral proficiency scoring categories.While Brown (2004) argues for grammar, vocabulary, comprehension, fluency, pronunciation, and task (pp.172-173), Ben Knight (1992) stresses the insertion of the following assessment criteria: grammar (range and accuracy), vocabulary (range and correctness), pronunciation (individual sound, stress and rhythm, intonation, linking, elision and assimilation), fluency (the use of hesitation markers and temporal variables), conversational Skill (ensuring topic development, taking initiatives, being consistent, keeping conversation going), sociolinguistic skill (appropriate use of register and style, use of cultural references), non-verbal (maintaining eye contact, relaxed body posture, appropriate use of gestures and facial expressions) and content (consistency in terms of arguments and relevance).
Despite the informants' variation in terms of which particular criteria have to be mentioned, they seem to have a consensus regarding the necessity of providing specific feedback targeting a pre-definite set of oral language dimensions in order to help learners identify their strength and enable them to remedy and deal with their weaknesses.
Despite the existence of a number of achievements pertaining to the authentic, thoughtful, organized and ongoing oral language assessment procedures, there are a number of hardships that still have to be resolved.In the same vein, Zhang & Burry-Stock (2003) argue that problems pertaining to assessment are overwhelmingly present in every performance assessment.
In the present study, some of these problems comprise the following: 1) The ongoing assessment might keep learners on their nerves as they are constantly under control from both the teacher and their peers.In this way, the learners' level of anxiety can be increased, which might influence their oral performances negatively.Along with the current study, Brown (2001) argues that anxiety makes learners perceive their oral productions as incorrect, unintelligent or ambiguous.
2) The learners' weak knowledge or unfamiliarity with some topics influences their performance and the way they are being assessed.
3) The amount of support given to learners through assessment varies from one learner to another.In other words, the performance conditions have an impact on the learners' output (Nation & Newton, 2009).
4) The hardships pertaining to the assessment of the learners' speaking performance though the use of technology comprises the difficulty in handling or in implementing the use of multimedia or digital devices.
Thus, the problems that can have a negative impact pertaining to speaking performance assessment are related to affective or psychological factors, knowledge factors, performance factors and technical reasons.

Conclusion
To conclude, the assessment of speaking has received much consideration at the tertiary level.Its implementation serves in improving the students' learning process and in boosting their speaking potentials.In addition to their resort to analytic rubrics, the teachers relied on a number of interactive, authentic, communicative activities that guaranteed an ongoing sustainment of the learners' speaking proficiency.The creation of meaningful opportunities for oral language evaluation to be held during the English sessions incites both teachers and learners alike to assume active roles in the assessment process.Despite the existence of some problems related to affective/ psychological, knowledge, performance and technical reasons, teachers strive to implement both types of classroom assessment which revolve around the use of formative and summative use of speaking assessment.

Implications
The present research has a number of implications.It gives a clear picture about what teachers think of speaking skill assessment and how they deal with that in practice.It explores how classroom assessment is used to monitor students' speaking skill and to support teachers' instructional and teaching methods in order to guarantee maximal learner communicative growth especially that assessment is regarded "as a positive tool for learning and an interconnected part of teaching and learning" (Berry, 2011, p. 89).In this way, the present study tries to provide an account of speaking assessment procedures in a number of Tunisian universities.Besides, the challenges depicted in this research might compel other EFL teachers or scholars in general to address imminent methodological issues encountered when conducting assessment in order to help EFL practitioners ensure more efficient assessment practices.Last but not the least, this research tries to contribute to the existing knowledge on speaking assessment in the world in general and in Tunisia in particular.

Limitations
The present study comprises a number of limitations.The latter includes the reliance on a single research instrument.Therefore, the resort to triangulation helps in strengthening the research findings and conclusion.Among other noticeable weaknesses is the reliance on a single perspective which is that provided by teachers.So, taking into consideration the learners' conceptions might also help in arriving at a more comprehensive overview of classroom speaking assessment.The previously mentioned weaknesses are hoped to serve as an impetus for other researches to be carried out. Poor knowledge about certain topics.
 Perplexing and incomprehensible input.
 Difficulty of applying multimedia or digital devices to the process of speaking assessment.

Table 1 .
Information about the informants' gender

Table 2 .
Information about the informants' age groups

Table 3 .
Information about the informants' teaching experience of oral skills

Table 4 .
The informants' major reasons for conducting speaking assessment

Table 5 .
The different types of speaking assessment used by teachers

Table 6 .
The major source of oral classroom assessment

Table 7 .
The measures used in the assessment of the learners' speaking performance