The Effect of EQ and Gender on Oral Cloze Performance of Simultaneous Interpreting Students

Introduction

Interpreting as a kind of oral mode of transferring meaning between two languages was defined as “presenting in the target language, the exact meaning of what is uttered in the source language either simultaneously or consecutively, preserving the tone of the speaker” (Mahmoodzadeh, 1992, p. 231). 

Simultaneous interpreting (SI) is considered as one of the most difficult linguistic skills by researchers investigating brain and language (Grosjean, 2011). Part of the difficulty of SI can, according to Chernov (2004, p. 14), be attributed to the simultaneous nature of the task, i.e. “the concurrent nature of SI speech perception and TL speech production and the need to start the translation process before the SL utterance is completed.” [emphasis in original]. This is what Gonzalés, Vasquez and Mikkelson (2012) referred to as the non-linear feature of SI since “the interpreter does not first grasp the meaning of the SL message and retrieve the target language (TL) equivalent elsewhere in the brain” (p. 855). Rather, there is a kind of overlap between the two processes of listening and speaking but “human brain is incapable of performing two complicated tasks at the same time” (Viaggio, 1988, p. 399).

Another aspect of the extreme conditions of SI is time constraints or the external control over the rate of SI. Unlike the translator who can freely determine his own rate of information processing, “the amount of information processed by a simultaneous interpreter, depending partly on the pace set by the speaker, is limited by human psychological factors and the capacity limitations of the human brain” (Chernov, 2004, p. 16). The fact that in SI the source text is always delivered at a faster rate than the interpreted target text can be explained by the difference between improvised discourse versus recited texts. The point is elaborated by Chernov (2004, p. 18):

In SI the interpreter always produces her communication spontaneously and on the spur of the moment, “accessing” the sense (content, message) of her communication from the SL discourse. Speakers, in contrast, very often do not formulate their communication on the podium, but simply recite a pre-prepared text.

This recited nature of ST creates additional difficulty in SI since a hesitancy pause which is the result of searching for the intended word by the speaker is removed from recited texts. Consequently, the text is uttered at a higher speech rate. Moreover, the lack of redundancy in recited speech and the more elaborate syntactic structures of a prepared text are considered as other sources of difficulty in SI by Chernov (2004, p. 18). Furthermore, according to Chmiel (2008, p. 261)  SI “is a challenging task in terms of cognitive processing” which “is performed under stress and time constraints, which leads to fatigue”. Therefore, to reduce the difficulties of this mentally challenging activity it is “necessary to ensure special working conditions for SI, with at least two interpreters present in the booth” (Chmiel, 2008, p. 261). Based on the rule mentioned by The International Association of Conference Interpreters (AIIC) “an interpreter shall not, as a general rule, work alone in a simultaneous interpretation booth without the availability of a colleague to relieve her or him should the need arise” (AIIC, 2015).

Another source of difficulty in SI can be probable comprehension problems as a result of accented source speech. The empirical study conducted by Lin, Chang and Kuo (2013) approved the decrease in accuracy with phonemic and prosodic deviations. Harris elaborated the point in this way:

The interpreter, besides speaking his language, must learn to hear and understand them as pronounced by others: by people talking various dialects of them and by some who speaks them abnormally because of poor education or because they are nervous or because languages are not native to them. (1981, p. 159)

The combinations of the above-mentioned specific features of SI are considered as various sources of pressure on the interpreter which in turn give birth to another peculiarity of SI, i.e., stressfulness.

The aforementioned complexities of simultaneous interpreting turn it into a frustrating task for interpreting trainee students. Inasmuch as some students with acceptable proficiency in English, cannot show the desired achievement in this course. Many students claim that their inability to perform the task adequately is due to the stressful nature of simultaneous interpreting which is amplified for some interpreting students. This group blames their personal characteristics as the main reason for their lack of success in interpreting. This view acts as an obstacle in the way of training interpreting students. But, achievement is the outcome of instruction, and as mentioned by Osokoya (1998) it is the end product of a learning experience. Accordingly, a great deal of attention should be devoted to the teaching method and material. Put in other words, the individual differences should be the focus of attention in addition to the development of skills required for interpreting.

Emotional Intelligence and Simultaneous Interpreting

The term emotional quotient (EQ) or emotional intelligence was introduced by Mayer and Salovey (1997). It refers to the ability to monitor one’s own and other’s feelings and emotions, to differentiate among them and use this information to guide one’s thinking and actions. For Bar-On (1997), emotional intelligence is a collection of non-cognitive capabilities, competencies, and skills that influence one’s ability to succeed in order to manage environmental demands and pressure.

The study of emotions is not a new field of inquiry. However, it has remained under-researched in Interpreting Studies. Although it is known that individual differences between interpreters is one of the possible reasons of their differing performance in terms of quality (Schweda Nicholson, 2005). In the same vein, Hubscher-Davidson (2013) highlighted the study of non-cognitive psychological processes in Translation and Interpreting Studies. She stated that studying the emotional intelligence of translators and interpreters can help us to have a better understanding of these two kinds of transforming activities. Hubscher-Davidson (2013) claimed that “exploring the emotional intelligence of translators and interpreters could add something to our understanding of their working process. In turn, this information could have important repercussions on –and feed into- the ways translators and interpreters are currently being trained” (p. 333).

This emphasis in literature for the importance of considering interpreters’ EQ in training was a source of inspiration for the researchers. Moreover, the association between SI and EQ was created on the basis of various potential links and connections between the two.

First of all, literature abounds with various studies that revealed the impact of stress on the quality of interpreting (Seleskovitch, 1978; Henderson, 1987; Moser-Mercer, 2008; Timarova & Salaets, 2011) on the one hand. A large number of studies demonstrated that EQ can significantly moderate responses to stress (Mikolajczak & Luminet 2008; Mikolajczak, Roy, Luminet, Fillée, & de Timary, 2007) on the other hand. Hence, it seems plausible to suppose interpreting trainees with higher levels of EQ can have a better performance in SI.

Secondly, a study conducted by Dewaele, Petrides, and Furnham (2008) showed that people with high EQ have the lowest level of communicative anxiety. Since one of the requirements of interpreting is the ability to speak publicly, here once again there seems to be a link between high levels of EQ and better performance in SI interpreting.

Thirdly, research indicated that higher EQ individuals are more prone to use problem-focused rather than avoidance coping strategies (Petrides, Pita, & Kokkinaki, 2007). It sounds logical to assume some kind of relation between EQ and the ability to interpret simultaneously. There are some coping tactics in SI introduced by Gile (1995), which should be used by interpreters to deal with problems the selections of which are not at random. However, it seems that the ability of problem-solving can have an effect on the choice of the appropriate tactic. Moreover, Mikolajczak et al. (2007) found that high EQ helps individuals to appraise situations as challenging rather than threatening.

Furthermore, the effect of EQ on team performance and group cohesiveness was shown by Quoidbach and Hansenne (2009).  In simultaneous interpreting, there should be a kind of cooperation between active and passive interpreters. Consequently, it seems those interpreter pairs with higher level of EQ may be better candidates for SI, due to their better performance in team work.

Generally speaking, among the various skills required for interpreting such as tact, judgment, sense of humor, calm nerves, listed by the AIIC (2017), one can find the trace of link with personality. Therefore, it sounds reasonable to suppose a kind of relationship between EQ and the ability to interpret.

Simultaneous Interpreting Exercises

One of the debates with regard to interpreter aptitude is the issue of nature versus nurture (or born vs. made). Most scholars believe in the role of training in interpreting (Mackintosh, 1999; Kalina, 2000; Timarová & Salaets, 2011). Lambert (1992, p. 264) clearly stated that “there is no such a thing as a born interpreter”, rather there are some skills required for success as an interpreter-trainee or working interpreter. These various skills highlight the importance of thoughtfully designed interpreting programme which according to Lambert (1992, p. 264) can “optimize chances for the growth of both the students and the profession itself”.

Lambert (1992) distinguished two types of interpreting exercises. The first group is selection tests that “serve as eliminatory tests in that trainee-interpreters may be discouraged from moving on to the next stage if they are unable to master graded skills” (Lambert, 1992, p. 264). A stated by Lambert (1992), these kinds of tests are used to “determine which candidates are most likely to become successful interpreters once they enroll in the training programme” (p. 265). The selection tests include: “shadowing, cloze, sight translation/sight interpretation, memory test (Wechsler) and interview” (Lambert, 1991, p. 578).

The second group is training tests that are used to develop interpreting skills of candidates who have proven to be successful would-be interpreters. Lambert (1992, p. 264) introduced twelve pedagogical techniques to be used in simultaneous training namely “listening and recall; shadowing; dual-task training or parallel processing; paraphrasing; abstracting or telescoping; clozing; sight translation; sight interpreting; lagging; anticipating; processing digits, names, acronyms; ear preference and hemispheric processing”.

As can be seen, the exercises of shadowing, cloze, sight translation and sight interpretation are common exercises in both groups. The inclusion of these exercises as both selection exercises and training exercises can be indicative of the important role of these exercises in training interpreters.

The cloze technique was chosen as an instrument in the present study to investigate the plausible relation between EQ, gender and SI performance. 

The Importance of Oral Cloze in SI

The cloze procedure is a method of test construction which consists of deleting words from prose and asking students to fill in the blanks. As declared by Oller and Conrad (1971, p. 183):

The term ‘cloze’ was used with the notion of Gestalt ‘closure’ in mind, referring to the natural human psychological tendency to fill in gaps in patterns. The restoration of words deleted from a selection of prose in order for the passage to make sense is a special use of this ability to complete broken patterns.

According to Lambert (1992), the cloze test is suitable for interpreting and can be used both as a tool for selecting interpreters as well as an effective interpreter training technique, since it “serves as a measure of integrative skills, namely the anticipatory processes involved when native speakers of a language engage in normal conversation, including listening as well as speaking” (p. 226). The idea of employing oral cloze for interpreter training was introduced by John Long in a study to select conference interpreters (Gerver, Longley, Long, & Lambert, 1984) on the “premise that interpreters normally work from spoken discourse and because oral cloze effectively paces the listener in the same manner as a conference speaker does” (Lambert, 1992, p. 226). Oral cloze technique as asserted by Anaya and Lopez (1990, p. 651) can be used to reinforce various kinds of interpreting skills such as “how to anticipate word meaning, how to handle the syntaxes of any specific text, how to improvise vocabulary, and to improve his memory, concentration and creativity”.

The cloze test, as stated by Lambert (1992, p. 226) can be devised in various forms. The simplest way is choosing a passage in the students’ mother-tongue and asking them to fill in the blanks (visual/aural mode). This version of a cloze exercise is useful for developing students’ public speaking ability and reading ahead in anticipation. The second category of mother tongue-to mother tongue cloze requires students to shadow the text in addition to fill in the blanks.

Unlike the previous cases, the second group of cloze procedure is devised taking translation factor into account. Students listen to a passage and are asked to write down the deleted words either in the SL or in the TL (aural/written mode). Moreover, ‘language-switch cloze exercises’ can be used where in addition to filling the missing words, a simultaneous interpretation of the passage should be offered (aural/oral mode).

Another variation of cloze developed by Timarová and Ungoed-Thomas (2009, p. 230) is called personalized cloze where students listen “to a short piece of text in which the speaker provided some basic details about himself. Students repeated the text verbatim in the foreign language and replaced all personal information-name, age, nationality, etc –with information about themselves”.  Hence, the combinations of the shadowing and paraphrasing exercises are used in this kind of cloze. Pöchhacker (2011) also introduced another new form of cloze. The SynCloze combines anticipation and vocabulary exercises. Students are required to listen to an aural cloze and fill in the gaps by giving as many synonyms as possible at the sound of the beep.

Purpose of the Study

The researchers of the present study set out to investigate the association between emotional intelligence, gender, and oral cloze performance in SI.

To this end, the following research questions were posed and investigated in this study:

1- Does EQ (both total scores and subscales scores) predict the performance of trainee students on oral cloze test significantly?

2- Are there any differences between males and females regarding their scores on SI oral cloze test?

Methodology

Participants

For the purpose of this study, 70 participants majoring in English Translation from Shahid Bahonar University of Kerman were selected. They were all in their 8^th semester and had enrolled in the course ‘Oral Interpretation’ taught by the researcher and had no experience of interpreting before. All participants had Persian as their A-language and English as B-language. They sat for the oral clause exam as one of the simultaneous interpreting aptitude tests. The demographic information of the subjects (i.e. gender and age) is provided in Tables 1 to 3.

Table 1. Participants’ Gender

Table 2. Participants’ Age

Table 3. Participants’ Descriptive Statistics (Gender, age)

Instruments

The study used the emotional intelligence questionnaire and an oral cloze test as its instruments for data collection. In what follows, a detailed explanation of each instrument is offered.

Emotional Intelligence Test

To evaluate students’ emotional intelligence, a Persian version of the emotional quotient inventory (EQ-I) developed by Samouie (2003) was used. The original version of this self-report instrument designed by Bar-On included 133 items with Likert-type responses, later revisions applied to the test by Bar- On (1997) himself reduced its size to a considerable degree, so that the modified version of the test comprised only 117 questions. Due to the fact that some questions were found irrelevant to the Iranian context, Samouei (2003) developed a modified and translated version of the questionnaire which encompassed only 90 questions. He found that this test did have sufficient reliability and validity indices in the Iranian context. Based on her findings, the reported Cronbach’s alpha gained for the questionnaire was 0.93, and the reliability index gained through odd-even, split-half method was 0.88. In another study, Dehshiri (2003) found the reliability index of 0.77 for the questionnaire. With respect to its wide use in the Iranian context, this 90-item version of the questionnaire was employed as the principal means of gathering data in the current study.

It's worth noting that each of the items on the questionnaire is related to one of the 5 composite scales that comprise 15 subscales of emotional intelligence. All the questions were scored on a five-point Likert-type scale ranging from 'strongly agree' to 'strongly disagree'.

Reliability Analysis

The internal consistency reliability of the questionnaire employed in this study was investigated running Cronbach alpha. Table 4 presents the total and subscales of alphas for the EQ questionnaire. Evidently, all the great majority of alphas are above .7 which indicated a high internal consistency in all the scales and their subscales. As regards the low alphas (i.e. alphas for problem solving, reality testing, responsibility and empathy), the alphas could still be considered acceptable since firstly the number of items in these subscales is just 6, and secondly they are not that much different from the alphas reported in similar studies in which the reliability of this instrument has been checked (Samouei, 2003).

Table 4. EQ Questionnaire’s Reliability Statistics

After computing alphas, item-total correlations as a measure of item discrimination (ID) were computed for all the items of the questionnaires. The statistics indicated that most of these values are above 0.3, showing a high correlation between the item score and total scale score, hence high item discrimination. Regarding the few low IDs, the column titled “alpha if item deleted” should be considered, which shows that when these low-ID items are removed from the questionnaire, the alphas are not going to improve much (beyond the current alphas). Therefore, it is concluded that all the items of the EQ questionnaire used in this study are relatively functioning well, and the questionnaire is of acceptable items and internal consistency reliability.

Oral Cloze Test

To devise the oral cloze passage, the research followed Gerver et al.’s (1984) framework. An oral passage of approximately 407 words in length was chosen from “VOA Special”. The initial and final sections of the oral passage (about 45 words) were left intact to provide context and warm-up for the trainees. Every 10^th word was then deleted from the rest of the passages. The Nuendo (5.0.1) software was used to create an oral cloze from the VOA Special English passages chosen. The software helped to delete words from the oral passage and create a beep instead. The students were asked to listen to the passage and write down, as quickly as possible, the missing words signaled by a beep tone.

Data Collection Procedure

During the data collection phase, the emotional intelligence questionnaire was administered. Students had their time to answer the questions. The oral cloze test was administered in a separate session. To answer the cloze test, students had to write down 34 words signaled by a beep. The exact word method was used to assess their performance in this test.

Data Analysis and Findings

RQ1: Do EQ scores (both total score and subscales scores) predict the performance of SI trainee students on oral cloze test significantly?

Answering this question required employing multiple regression for EQ scores (both total score and subscales scores) as the predictor variables and oral cloze scores as the predicted variables. However, before running multiple regression, it is necessary to see which predictor variables are significantly correlated with the predicted variables.

Table 5 presents the descriptive statistics of EQ (both total scale & subscales) and oral cloze scores, and Table 6 presents the Kolmogorov-Smirnov (KS) test results on the normality of the EQ (both total scale & subscales) and oral cloze scores.

Table 5. Descriptive Statistics of EQ & Oral Cloze Scores

As the results of KS test in Table 6 show, only the self-actualization, and oral cloze scores are not normally distributed (p < .05), but since the KS alpha is just a little below .05, and the sample size is this study is large enough (n = 70), Pearson Product Moment correlation (Pearson r) as a parametric test could be considered robust against this violation of normality. All in all, it was decided that the data were normal enough to allow for running Pearson r as a parametric test.

Table 6. One-Sample Kolmogorov-Smirnov Test

Table 7 presents the Pearson r for the EQ (both total scale & subscales) and oral cloze scores. Evidently, only Stress Tolerance and Flexibility are significantly and positively correlated with oral cloze scores (p < .05). Therefore, only these variables were included in multiple regression analysis, and the rest of the variables were excluded from the analysis.

Table 7. Correlations

To run multiple regression analysis, one assumption which needs to be met is the normality of standard residuals which was checked by drawing the histogram. Evidently, the histogram shows an almost normal bell-shaped distribution of standardized residuals. Moreover, the Normal Probability Plot of the regression standardized residuals indicates that the points have lain in a relatively straight diagonal line from bottom left to top right without too many deviations; therefore, the assumption of normality of the regression standardized residuals is met. Finally, the scatterplot of the standardized residuals indicates that there are only a few negligible outliers which have lain outside the rectangular cluster of the data in the center. Moreover, since there is a clear or systematic pattern to the residuals (e.g. curvilinear or higher on one side than the other) with relatively few deviations from a centralized rectangle, it is assumed that there is no violation of homoscedasticity.

Multiple regression also assumes no multicollinearity, which is investigated by considering the correlation among the independent or predictor variables. The result indicates that the correlations among the independent (i.e. predictor) variables are not above 0.7, which demonstrates that the multicollinearity assumption is met. After examining the assumptions of multiple regression, two multiple regressions were run: one with SI aptitude total score as the predicted variable, and the other with oral cloze scores as the predicted variable.

After checking the assumptions of multiple regression, the actual regression tables are investigated as follows. It should be first noted that since there was not any logic or justification to specify an order for the inclusion of the predictor variables in the model, simultaneous rather than hierarchical multiple regression analysis was run.

Table 8 presents the Adjusted R Squares for multiple regressions, which is .095 for oral cloze scores as the predicted variable. This means that:

9% of the variance in oral cloze scores is explained by the 2 predictor variables (i.e. Flexibility, Stress Tolerance), which could be considered as a small effect size according to Cohen’s guidelines for effect size.

Table 8. Model Summary^b

Table 9 indicates that the results are statistically significant (p < .05), suggesting that the regression model overall (i.e. including all predictor variables) predicts oral cloze scores significantly well.

Table 9. ANOVA^a

Table 10 presents the beta coefficients for the predictor variables. While we know from the ANOVA results before that all predictor variables together can significantly predict oral cloze scores, according to Table 10, the t-tests indicate whether each predictor variable individually is making a significant contribution to the model or not. If the t-test associated with a b-value is significant, then the predictor is making a significant contribution to the model holding the other predictors constant. As the results of both regressions indicate in Table 10, individually none of the predictor variables is making a significant contribution to the model; therefore, it is concluded that:

Oral cloze scores are only predicted positively by the 2 predictor variables together but not separately (i.e. Flexibility, Stress Tolerance) with small effect size according to Cohen’s guidelines for effect size (adjusted r square = .09).

Table 10. Beta Coefficients^a   for the Predictor Variables

RQ2: Are there any differences between males and females regarding their scores on oral cloze test?

Answering this question required dividing the participants into males and based on their oral cloze scores. After doing so, the descriptive statistics of the males and females were computed based on oral cloze scores (Table 11).

Table 11. Descriptive Statistics Based on Oral Cloze Scores

Table 12 presents the Levene test results on homogeneity of variances as an assumption of MANOVA. As the results indicate, the test results are not significant (p > .05), hence meeting the assumption.

Table 12. Levene's Test of Equality of Error Variances^a

Table 13 presents the multivariate tests results comparing the dependent variable - oral cloze scores. Evidently, the results are not significant (p > .05) for oral cloze scores across male and female groups.

Table 13. Multivariate Tests^a

Table 14 presents the univariate tests results comparing males and females in terms of the dependent variable - oral cloze scores. Again, the results are not significant (p > .05).

In sum, the multivariate and univariate results indicate that the null hypothesis to this research question is supported. That is to say, there is no difference between males and females regarding their scores on oral cloze test.

Table 14. Tests of Between-Subjects Effects

Discussion and Conclusion

The present study aimed at finding the possible relationship between SI student trainees’ performance on oral cloze test, their emotional intelligence and their gender.

The results of the first research question demonstrated that no significant relationship was found between total EQ and students’ performance on oral cloze test. However, the findings revealed that 9 percent of the variation in students’ performance can be explained by taking the subscales of Flexibility and Stress Tolerance together into account. In other words, the prediction power of the two subscales of EQ is significant but minimal. However, this minimal significant effect of the two predictor variables can have an important pedagogical implication for interpreter training.

The importance of flexibility is due to the need to cope with different speakers, subject matters, and situations in different interpreting settings, since as stated by AIIC (2017, para. 6) “not all meetings are the same, nor do they demand the same aptitudes from interpreters”. Moreover, Roeseler (2017, p.74) believed that conference interpreting is an inconstant activity by nature which demands interpreters to “think quickly and flexibly”. In addition, a study by Timarova and Salaets (2011, p. 31) proved that “self-selected interpreting students are cognitively more flexible and are less negatively affected by anxiety”. Accordingly, it seems students with a higher sense of Flexibility would be better candidates for simultaneous interpreting. Hence, paying attention to students feeling of flexibility and helping them to develop it can facilitate interpreter training process.

Stress Tolerance has been found as another positive predictor. As has been aforementioned in the literature, simultaneous interpreting is an inherently stressful activity due to the existence of various stressors. Previous researches pointed out that individual differences in coping with stress, are one of the influential factors that distinguish interpreters based on their level of success. The study conducted by Jiménez and Pianzo (2001) revealed that the ability to cope with stress is one of the important requirements for interpreters. In the same vein, Kurz (2003, p. 55) showed that “conference interpreters have better situation-dependent control over their feelings of anxiety and manage to label their anxiety in a positive way”.

The second research question examined the effect of gender differences on oral cloze score. The results indicated that there was no difference between males and females in their performance on oral cloze test. In this line, previous research suggested no significant difference between male and female student interpreters with regard to the quality of SI (Hasanshahi & Shahrokhi, 2016). The pedagogical implication of this question is that both male and female students have an equal chance of acquiring oral cloze ability as one of skills required for interpreting. In other words, students’ performance on oral cloze as one type of SI aptitude test is not pre-determined by their gender. However, gender-differences have been reported in interpreting from a different perspective by few studies. Research, for instance, reported gender imbalance in conference interpreting with a preponderance of women in the profession (Ryan, 2015). Another study revealed significant differences between female and male simultaneous interpreters’ approaches to prosody (Cecot, 2001), women have also been found to have a longer EVS than men (Collard & Defrancq, 2016), women applied self-correction as a strategy in SI more than men (Ziobro-Strzępek, 2014). The findings of these studies do not seem to contradict the results of the present study; rather they may indicate distinguishing professional interpreters from trainee interpreters while investigating the role of gender. Put in other words, studies conducted on professional interpreters showed some gender-bound differences, whereas studies concerning novice trainee interpreters did not demonstrate such gender-related differences. This can be an interesting finding in interpreting studies, which suggests that both male and female students have an equal chance of acquiring the skills required for interpreting (in this case oral cloze skill). In other words, students’ performance on oral cloze test is not pre-determined by their gender.  But the gender-bound differences found in professional interpreters can appear as a result of social conditions of the profession, individual achievements developed through experience or other possible factors that should be recognized.

As a general conclusion, the findings of the study suggested that interpreter training programs should incorporate the development of personality traits such as EQ in their curriculum. In other words, individual differences can be one of the influential ingredients for a successful training of would-be- interpreters. Hence, acquiring interpreting skills alone may not guarantee successful achievement in SI. Moreover, the role of these factors to reduce trainees’ stress can accelerate the process of interpreter training for both parties, i.e. trainers and trainees.