PROCESO DE ELABORACION DE UNA TEJA O LADRILLO

The fifth and last module focused on invitations, excuses, and leaving and taking messages. The allocated time for this module was six hours. The final goal of the unit was to enable students to talk about their activities and organize an

invitation. Students were expected to invite people, ask follow-up questions (where to meet, what time, what to bring, etc.), and accept or refuse the invitation (provide an excuse). Special attention was paid to politeness as well as face expressions and body language to communicate feelings.

As with all the previous modules, vocabulary and grammar structures (modal

would, present continuous for future) were introduced. Students practiced asking

people out as well as accepting or refusing invitations. To see language used in different contexts, a video featuring a surprise party, and a funny video about a little girl taking a message were presented. Listening activities were also completed to allow students practice taking messages. To practice language in context, students took turns to take and leave phone messages.

Finally, to organize their role-plays, students formed groups of four or five. A divergent role-play model (Christopher & Smith, 1990) was suggested by the

teacher, who provided general guidelines and asked students to prepare a sketch about the organization of a party. Students were again encouraged to use non-verbal cues (facial and body expressions) as well as words and pitch to express feelings,

Ana Isabel Andrade Chacón 67

such as surprise, happiness, boredom, etc. For approximately one hour, each group was in charge of developing a dialogue in which they would invite their peers to an event. Then a confirmation or an excuse needed to be provided. General

instructions were given by the teacher orally and in writing, while students‟ roles were assigned within the group (see Appendix N).

The next class (two hours) was devoted to role-play rehearsal and presentations. Special attention was given to the use of non-verbal cues (facial expressions, body language) in addition to the use of language. Students were evaluated individually using the speaking rubric.

After the completion of the five modules and their corresponding role-plays, students were given a questionnaire to express their opinions regarding the

intervention. As previously mentioned, this final questionnaire contained questions that allowed students to provide their insight about the applied technique and whether role-play had influenced their speaking proficiency during the course. Additionally, the students‟ speaking skill was measured through a post-test which was later compared to the results of the pre-test to determine whether students had improved their speaking proficiency. Also, the scores from each role-play were analyzed to assess students‟ progress and possible improvement related to the intervention.

Ana Isabel Andrade Chacón 68 CHAPTER IV

DATA ANALYSIS AND RESULTS

The present chapter presents the results obtained after the investigation. It includes an analysis of the data collected through the various instruments employed to provide answers to the investigation questions; that is, to what degree is role-play related to the enhancement of the speaking skill of level 2 students in an EFL

classroom at the University of Azuay? and How do level 2 EFL students perceive the implementation of role-play in the classroom setting?

The first question will be answered through the analysis of quantitative data collected using speaking pre-tests and post-tests to measure the effect of role play activities on the students‟ speaking performance. Data from the tests were analyzed with the Wilcoxon signed-rank test using SPSS statistics to determine if the means of the two tests were significantly different from one another. The use of non-parametric statistics (Wilcoxon‟s test) is recommended for analyzing data from continuous

variables or ordinal variables (Mackey & Gass, 2011). Examples of the former are exam scores while Likert scales are examples of the latter. Also, this test is

recommended when the analyzed data come from related groups; that is, the same subjects are assessed in different times during the study. Additionally, when data do not conform to a normal distribution, a Wilcoxon signed-rank test has been

suggested (Laerd Statistics, 2013; Larson-Hall, 2016). Data from this research study met all these conditions, since exam scores from the same participants were

obtained at different times, and the collected data did not respond to a normal distribution (see Appendix O); thus, the use of Wilcoxon‟s test was considered appropriate. Additionally, it has been tried in studies with very small sample sizes.

Ana Isabel Andrade Chacón 69

The outcomes of these studies have shown that the Wilcoxon test provides more feasible results than the t-test with such small samples as 5 participants (de Winter, 2013).

The test is conducted by calculating the differences between the first and the second sets of data and ranking the absolute value of these differences from the lowest to the highest so that the lowest difference receives the rank of 1, the next receives 2, and so forth. The test is based on numeric differences between the results from the first and the second set of data (in this case, between the speaking pre-test and the post-test). A negative difference indicates that the second set of scores were higher than the first, while a positive difference shows a lower result. Depending on the null or alternative hypothesis, these could be interpreted as an increase or a decrease, respectively (McDonald, 2014). For the present study, a negative difference between the scores of the pre-test and the post-test was indicative of an increase in the students‟ speaking performance.

In the following section, the quantitative results will be explained through descriptive statistical tables, rank tables, and test statistical tables. The descriptive statistics show calculations such as mean, minimum, and maximum scores, as well as quartile statistics. The ranks table of the Wilcoxon test shows the number of participants who achieved a higher (as indicated by the term positive rank) or lower score (as indicated by the term negative rank) between the two tests. In cases in which there was no difference between the two tests scores, the term ties was used to group these results. The test statistical tables indicate whether the differences in scores were actually statistically significant according to the p value, also known as “critical value” or the probability of obtaining the observed results. According to

Ana Isabel Andrade Chacón 70

Mackey and Gass (2011), a 0.05 level of significance (5% probability) is an acceptable level for second language research.

To answer the second research question, qualitative data collected through two questionnaires that students completed at different times during the intervention to assess their feelings and points of view regarding role-play, were analyzed and triangulated with the teacher‟s journal to provide a view of the students‟ opinions regarding role-play.