The Effects of Audio Comedy Test on Listening Comprehension Skills of EFL Learners

This study aims to investigate the effects of audio comedy on English listening comprehension test results of English as a Foreign Language (EFL) learners. The English listening test comprises four sections; each section has a seven-minute comedy audio mode and 13 questions, and participants listen to the four sections successively. This study was conducted with 117 sophomore, junior and senior students at Sebelas Maret University in Indonesia. Two-way Analysis of Variance (ANOVA) was adopted to distinguish between male and female students on the four successive test sections measuring English listening comprehension skills. The findings are: 1. Successive practice tests will stimulate English listening skills; and 2. Successive practice tests will improve students’ English listening skills. The comedy audio mode creates a low-stress English listening atmosphere and reduces the learner’s anxiety.

comprehension tests (Sharpe, 2013). Pan (2017) explains four fundamental guidelines for listening macro-skills: meaning, function, context, and form. Meaning refers to understanding of the gist, main idea, and details of the discourse which are explicitly or implicitly stated. Function is related to the understanding of the core speech act, macro speech act, and micro speech acts which are directly or indirectly performed. Context describes the situation and the background knowledge of the discourse. Form involves recognition of the linguistic or phonological form of the speech. The topic of audio listening may revolve around personal life, social and cultural issues, or popular knowledge, with said topic being relevant to, and informing, the design of related tests.
Listening comprehension is important in language learning (İnceçay & Koçoğlu, 2017;Hsu, Hwang, & Chang, 2014). Listening skills have the function of allowing individuals to understand the content of language and imitate it based on incoming sounds. By imitating these sounds, learners can investigate and express what they have learned. Carreira (2014) states that "the fluency in listening lies in understanding that while most listening behaviour is not visible, listening actually does involve real mental processes that can be regulated by the listener" (p. 281). Learners can arrange the psychological process of listening that makes them capable of communication. Hsu, Hwang, Chang and Chang (2013) state that listening is the first step of English learning, followed by reading and writing. Rost (2014) explains the three aspects of listening fluency as "ease with continuous listening and recall of main ideas, an ability to recognize specific words and phrases in rapid speech, and smoothness of listening behaviour in face-to-face interaction" (p. 282). In relation to these three aspects, Carreira (2014) states that listening fluency depends on three frameworks, namely: top-down, bottom-up, and interactive fluency. Top-down fluency means listeners have the understanding to listen comprehensively to long, authentic input conversations. Listeners possessing top-down fluency have high receptivity, develop new ideas, understand tasks, and are familiar with strategies to use when they encounter problems in listening situations. Acquiring relevant background knowledge and participating in meaningful activities during and after listening can support top-down listening fluency. In particular, listeners will easily understand a listening topic if it relates to their educational background. The understanding of content during and after listening guides listeners as they perform subsequent actions, commands, and tasks. Bottom-up fluency, in contrast, means the ability to decode speech automatically. This also implies that listeners can express their thoughts effectively. Their native language provides a default template in their mind that allows them to decode the language. Interactive fluency is the interactive and cooperative aspect of language learning. Listeners who are interactively fluent are capable of speaking and expressing their understanding via conversations. Shahani and Tahriri (2015) investigated 45 Iranian female high school students using quasi-experimental design. The finding of the study shows that the experimental group outperformed the control group. The similar finding also stated by Ghasemboland and Nafissi (2012)

Method
The effects of audio comedy test on listening comprehension skills of EFL learners is focused on the learner's understanding about the comedy conversation in English. All questions of the test are arranged based on the content of the audio comedy.

Experimental Design and Hypothesis
This study adopts a time series design to investigate the effects of using a repetitive comedy audio mode on the English listening comprehension skills of EFL learners. Participants were exposed to audio comedy, after which they completed a listening comprehension test. The test was conducted in four sections, with each section having different content in terms of the specific audio comedy and the questions. All sections and questions were played once only, successively. The details of the listening comprehension test procedures are shown in Table 1. Note. AC means audio comedy; m means mean score.
The hypotheses tested in this study are: (a) Students' performance on successive sections of the practice test will exhibit a significant difference; (b) There will be no significant difference between male and female students in terms of English listening skills; and (c) Students' English listening skills will gradually improve with exposure to audio comedy.

Participants
This study was conducted with undergraduate students in Indonesia, where the students learn English as a second language. Total 117 students were participated. The details of the participants are shown in  Note. ab means the parallel class. were similar. Participants listened to the audio comedy for seven minutes and answered questions based on what they heard. Each correct answer was scored as one, while wrong answers were scored as zero.

Instruments and Procedures
The valid questions and distribution based on listening macro-skills (Pan, 2017) are shown in Table 3. Section one had three invalid items: q3, q9, and q12; section two had three invalid items: q4, q5, and q10; section three had two invalid items: q2 and q8; and section four had three invalid items: q4, q9, and q10. The total number of invalid items was 11 questions or 21.15% of the total. The valid items (78.85%) were checked using an item difficulty index (P-value) to verify whether all items were acceptable or not. The P-value was between .20 and .80, and the mean P-value was .47. It has been suggested that P-value should ideally be between .20 and .80 (Kline, 1993). The results indicated all items were acceptable and we could then proceed to the next level of analysis: a reliability test and two-way ANOVA. The result of the reliability test was acceptable n = 117 (r = .92, p < .05) and the results of the two-way ANOVA are shown in the Analysis and Discussion sections.

Analysis
The mean score on section one was .33. The mean score climbed up over .40 on section two and section three, then fell to .36 on section four. Scores can thus be categorized into two groups: section one and four are the "low" group, while section two and section three are "high" group. The standard deviation (SD) of each section was quite stable, between .17 to .19, revealing that no dramatic movement happened. The details of the mean and standard deviation are shown in Table 4. Based on the descriptive statistics, the mean scores of each section indicate improvement of the students' listening comprehension skills. The lowest mean listening comprehension score was in section one.
Sections two and three showed higher mean scores, and section four showed the second-lowest mean.
The mean scores on sections two through four were all higher than the mean score on section one. The SD from the four different sections remained stable, which means the English listening comprehension

Successive Practice Test
The two-way ANOVA was used to analyze whether there was any difference between genders, as well as to analyze whether there was any difference among the four sections of the test. The scores on the four sections were analyzed as follows. First, the average numbers of correct answers were grouped into four sections. Second, the average numbers of correct answers were grouped into two groups, corresponding to female and male students. The results of the two-way ANOVA are shown in Table 5. English listening comprehension skills. This is supported by the p-value of each section (see Table 4) which shows there was no significant difference between male and female students. The total score of the gender comparison test also shows there was no significant difference between male and female students (see Table 5).

The Improvement of English Listening Comprehension Skills
The graph Figure 1 shows that the students' English listening comprehension skills slightly improved over the four test sections, beginning with .33 in section one and rising to .40 and .45 in sections two and three, respectively, before dropping down to .36 in section four. The trend of the mean scores shows similar performance for male and female students; if the mean score of the male students increased, the mean score of the female students also increased, and if the mean score of the male students decreased, the mean score of the female students also decreased. Table 6 describes the multiple comparison of each section using the Bonferroni method. There are three pairs which exhibited significant differences when compared: section 1 versus section 2; section 1 versus section 3; and section 3 versus section 4 (see Table 6).

Discussion
The scores on the listening comprehension test could be categorized as low scores because the overall mean was only 38 per cent correct answers. It can be seen that the scores ranged from the mean of test 1 (M = .33) to the mean of test 4 (M = .45) (see Table 4) and the grand mean was .38. The low scores on the listening tests indicate that participants either were not familiar with the audio topic of the travel agency, or have poor listening comprehension skills. All participants were students in the field of education, and the audio comedy topic involved a travel agency. While the students' study background is not related to the audio topic per se, some elements of the audio content are also discussed in the field of education. The listening comprehension test in this study was designed to cover four listening macro-skills corresponding to meaning, function, context, and form (Pan, 2017). Most participants were weak when it came to answering questions based on context and meaning macro-skills. It bears noting that the context macro-skills are related to background knowledge; participants may have misunderstood the conversation and questions in this part due to slang and colloquialisms used in the conversation, or other unfamiliar words. Wallwork (2014) argues that comedy can be used to improve English listening skills, but it will be more difficult for learners when non-standard English or slang is used to convey the humor.
In addition to the background knowledge, the range of subject matter the listening materials cover (i.e., whether they pertain to local or global topics) can be considered as an obstacle to success on the English listening comprehension test. Becker (2016) states that questions containing local information are easiest to answer, compared to questions targeting preferences and pragmatic inferences. Pan (2017) stated that topics for English listening comprehension tests could be about personal life, social and cultural issues, or popular knowledge which is relevant to the design of the tests. The comedy dialogues used in this study featured actors speaking with British accents, which may have been an obstacle to participants' listening comprehension. Table 5 shows the calculation between test sections and gender, and the results can be summarized as follows: (a) scores on the sections within the test were significantly different (F = 8.19, p < .05); (b) no significant difference was found in the test results between genders (F = 1.51, p > .05); (c) the intercept between test sections and gender was not significantly different (F = 1.03, p > .05). The most important findings revealed by this study are findings (a) and (b), which are discussed below.

Successive Practice Test Sections
Students' scores on sections within the test showed significant difference (F = 8.19, p < .05), indicating that repeated exposure to audio comedy successfully improved participants' listening skills, as measured by the test. In other words, the audio mode helped participants to improve their English listening skills. This finding is supported by İnceçay and Koçoğlu (2017) who found that using listening tests using audio mode is a familiar phenomenon for students, and there will be no problems with doing so as long as the students focus on the instructions and questions. Moreover, audio mode can reduce the students' stress compared to visual mode. Audio mode reduces meaning redundancy  (Ozdemir, Izmirli, & Sahin-Izmirl, 2016) and in line with this principle, it has been suggested adding text on screen which duplicates spoke audio contributes to poorer learning (Samur, 2012). Moreover, listening with audio mode fosters the improvement of English learners' listening skills. Although the participants have poor listening skills overall, as mentioned at the beginning of the Discussion section, they can still improve their listening skills via repeating listening. As stated by Deá k (2014) "language processing becomes faster with practice, even in childhood and adolescence" (p. 286).
A format involving learners listening repeatedly to English helps listeners to familiarize themselves with the test format. All sections of the test used in the present study relied on multiple choice format, with four options for each answer. The format helped test takers to predict the questions because the textbook was distributed before the listening test started. Participants could also have the textbook containing the answer options open while the audio played. This gave the participants the opportunity to quickly read over the answer options, which helped them to predict the questions and made it easier for them to choose the best answers. The assumption is consistent with Koyama, Sun and Ockey (2016) and Hemmati and Ghaderi (2014), who found that the ability to preview questions only, answer options only, or both questions and answer options helped listening comprehension test takers, compared to test takers who did not have the ability to preview in these ways.

Comparison of Test between Genders
Comparing the test results between genders did not reveal any significant difference (F = 1.51, p > .05).
The mean scores of male and female students were not significantly different, although male students showed slightly quicker responses on section 2. The mean scores on the section, third and fourth sections reverted back to both genders having the same mean values (Figure 1). This finding is consistent with Malagoli and Usai (2018), who found that male students have a better working memory than female students, but there is no significant difference between male and female students in executive function, when executive function is defined as a mixed ability construct composed of inhibition and working memory. The findings of Malagoli and Usai's (2018) study with 227 students aged 14-to 19 (148 female) supported previous research on both cognitive and neurological domains and were in accordance with what we know about knowledge development. Bohanek and Fivush (2010) likewise found that female students have higher levels of cognitive processing narrative skill than male students, but when the narrative itself is treated as a controlling variable between male and female students, there is no significant difference. Cox and Davies (2012) tested the English speaking capability of 179 EFL learners using automatic speech recognition (ASR) scored elicited oral response (OER), and their findings also showed no significant difference between male and female students. Cox and Davies did not analyze listening skills in particular, but English listening skills have been shown to be essential in language learning (Hsu, Hwang, & Chang, 2014;Carreira, 2014 them down into six components: memory, cognitive, compensation, metacognitive, affective, and social.
They found that male and female of undergraduate students used similar learning language strategies, and only the compensation component showed a significant difference between male and female students. This supports the assertion that there is no need to provide special learning strategies to members of each gender to improve language learning.

The Improvement of English Listening Comprehension Skills
The results of the test of students' English listening skills show a slightly upward trend if we look at the mean scores on the four successive sections, with a gradual increase from the first to the third sections, but a subsequent decrease on the fourth section ( Figure 1). In visual terms, learners' listening skills appeared to move up and down like a zigzag line. The trend can be illustrative for English educators in that they show English listening skills will not always necessarily show a steady increase, but may periodically decline back to previous levels, if only temporarily. This is consistent with Zafar (2016) who stated that remedial teaching strategies could be considered to improve students learning when errors or misuses of English language vocabulary were noted. The same proposition was advanced by Hung (2017) who found that remedial teaching given prior to summative tests helped students who had problems mastering learning materials. Additional information in the form of the post hoc comparison test showed that students' performance on section three (which had the highest median score) showed a significant difference compared with their performance on both section one and section four (see Table   6). This finding is consistent with Schmidt, Houwer and Rothermund (2016) who observed a learning acquisition curve that showed lower performance at the beginning and at the ending of a practice phase.

Conclusions and Implications
This study empirically examined the effects of what we call comedy audio mode on the English listening comprehension skills of EFL learners who are undergraduate university students. The results showed a significant difference in terms of students' test scores among four test sections, but no significant difference between male and female students. Exposure to successive practice tests appeared to allow the English learners to become increasingly familiar with the test format, put them at ease, and may have even helped them to identify the best answer for each question. The listening topic chosen can influence test takers' levels of anxiety and stress. Comedy dialogues were chosen with a view to reduce the level of tension in the English learning atmosphere and to motivate English learners to follow along with the audio content.
The audio mode of listening comprehension reduces students' stress during listening training and helps them to grasp the content of conversations while focusing on the questions and answer options.
Students could focus on questions and the audio dialogue without being distracted by pictures, running texts, or video images, and this freedom from distraction could help them to choose the best answers.