The Effects of Dynamic Assessment and Gender on EFL Learners ' Oral Narrative Task Performance

The present study is an endeavor to explore the potential of dynamic assessment (DA) as a way of scaffolding English as a foreign language (EFL) learners’ oral production. Although the literature on dynamic assessment is rich with studies focused on language components and skills such as vocabulary, reading, and writing, very few studies have attempted to examine the role of dynamic assessment in fostering realtime oral production. It is assumed that assessment of learners’ oral production need not focus solely on evaluation; rather, learners need to determine their weaknesses and try to compensate for them with the help of a more knowledgeable peer. Therefore, the present study also focused on evaluating the role of DA in Iranian EFL learners’ oral narrative task performances in the classroom. Furthermore, an attempt was made to explore the role of gender in the production of mediated narratives. For this purpose, a total of 60 students from two private universities were assigned to two intact classes: One class served as the experimental group, which was exposed to teachers’ mediation in learners’ narrative productions, and the other group functioned as the control group and received no such mediation for their oral narrative language production. The overall speech quality of learners’ production from each narrative over the treatment period provided the required data. The results were analyzed through independent samples t-tests, which revealed significant effects of both group and gender differences. The experimental learners, as expected, were able to progress and cultivate their oral production abilities after having been exposed to DA. Regarding gender, male participants were more successful in terms of enhanced second language oral production. The study has significant implications for the integration of DA into learners’ oral production, suggesting that EFL narrative production is a skill that is highly anxiety-provoking for learners.


Introduction
Dynamic Assessment (DA) has attracted a great degree of attention in both the educational field and the language assessment literature over the past few decades (e.g., Elliot, 2003;Leung, 2007;Poehner, 2008;Sternberg & Grigorenko, 2002).According to Baek and Kim (2003), dynamic assessment refers to an active "interactive exploration of the learner's learning and thinking processes, and aim to investigate a learner's strategies for learning and ways in which these strategies may be extended or enhanced" (p.109).DA is based on a view of assessment and teaching in which teaching or feedback is integrated with assessment or testing.In spite of the fact that DA can be advantageous in classroom practices until recently only a handful of teachers have integrated it into their instructional practice and activities.This limited use of DA has been explained with regard to two significant issues (Lidz, 2009;Poehner, 2011).Primarily, instruments for DA have been mostly designed by psychologists, which make the use of such instruments difficult and challenging in classrooms due to the difficulties in making a transfer from the clinical environment to the classroom.In addition, substantial differences exist between teachers and psychologists with respect to language and background factors, which make the instruments less clearly understandable and useable for teachers.Secondly, the majority of DA instruments are developed for one individual child, while teachers need to consider their whole classroom and the environment in which they teach.In order to perceive and understand the full potential of DA for teaching practices, it is crucial to critically evaluate its implementation in classrooms with regard to students' different language skills and abilities.
The effectiveness of DA, in the present study, was gauged against the backdrop of EFL learners' development of oral narrative skills through the concept of 'mediation'.A student's confusion in choosing an appropriate repertoire for the classroom context can be encouraged by the teacher.This 'mediation' can be fostered by an understanding of Rogoff's (1997) concept of interpersonal planes of analysis, which can also aid in planning for these mediating actions (Fleer, 2002)."Using an interpersonal plane for program planning helps us think about not just the content of an interaction, but to also actively plan the type of interaction or mediation that best fits with the child/children."(Fleer, 2002, p. 114).Knowledge of students' likely preferred strategies for participation that will be demonstrated through DA is useful for the teacher, as it can help in a better plan for interaction and engagement with the activities in the classroom, and can expand children's repertoires.
According to Abrams, "a narrative is a story, either in prose or verse, involving events, characters, and what the characters say and do" (1993, p. 123).Teachers who are familiar with the narrative genre categorize narratives as the type of speech production that includes beginning, middle, and end components.A practitioner might also describe narratives as a conceptual instrument applied to make sense of the world and experience (Burnur, 1988).In order to indicate connection while recalling a story, narrators undertake the cognitive operation of sequencing events in the order of beginning, middle and end (Gergen & Gergan, 1986).Gee (1991), nevertheless, asserted that narrative is supposed to link past, present, and future in the light of having linear properties.Together with the fact that narratives often link endings to new beginnings, and therefore give them circular properties, this concept results in narratives having both linear and circular perspectives.Therefore, oral narratives pose several cognitive real-time language-processing challenges for language learners, which need to be scaffolded by appropriate teaching and evaluation methods that take into consideration the fact that narratives are also used for assessment purposes.The present study was, therefore, an attempt to look into the potential of DA in fostering enhanced oral narrative skills in EFL learners.Furthermore, the effect of gender was explored.

Review of Literature Dynamic Assessment
In language teaching research, an approach to DA that combines activities of assessment and teaching on a classroom level scale has been formulated.In order for this activity to be reliable, valid, and appropriate in everyday classroom activities, van der Veen and Poland (2012) recommended a framework that integrates the standardized and interactionist types of DA, as depicted in Figure 1.This framework constitutes a unique model, considering the fact that the majority of the available approaches and instruments for DA implement either a standardized (also known as interventionist; e.g., Budoff & Friedman, 1964) or an interactionist framework (e.g., Feuerstein, Falik, Rand, & Feuerstein, 2003, cited in Lantolf, 2009).The purpose of the standardized DA is to base mediation on a stable and unchangeable array of guidelines, clues, feedback, and teaching that is presented to the children as they progress through a test task.Teaching is recognized as a movement from implicit to explicit; as a result, the teacher obtains opinions and insight regarding the level of a child's progress in the task (Lantolf, 2009).Thus, this standardized type of DA is based upon a sandwich format, or testtrain-test design, where the child is initially administered a static pretest, followed by some type of short-term, standardized, and prespecified teaching about the abilities or knowledge that has been measured.Lastly, the child is once more tested on a posttest in order to understand the degree to which the child can utilize the intervention (i.e., teaching).In contrast to the standardized DA, Interactionist DA is specified by interaction and collaboration between the teacher and the child contextualized within meaningful classroom tasks, with the purpose of determining the degree of the child's development.In this type of DA, mediation is not standardized or stable; instead, it is discussed with the individual, which implies that it is progressively and consistently adapted and changed according to the learner's progress and responses (Lantolf, 2009).
There are different models of DA.The Standardized model fully integrates assessment and instruction so that one doesn't exist apart from the other.The Interventionist model has 2 subdivisions: "cake" format and "sandwich" format (Poehner, 2008).In sandwich format, a mediation phase is sandwiched between the pre and posttest phases.In cake format, the intervention is embedded in the administration of the test itself in the form of predetermined hints.Different studies have investigated the effect of DA on different skills and sub-skills, each according to one model of mediation of DA.While several studies have been conducted to investigate DA in foreign language learning, (Ableeva, 2007;Lantolf & Poehner;2004;Poenher, 2008;Sternberg & Grigorenko, 2002), it appears that little research has been carried out to examine the effects of dynamic assessment on grammar and the comparison of different models.
In an attempt to integrate both approaches to DA, a sandwich format (or test-intervention-test design) was adopted.This format presents the children with a standardized pretest first, followed by an interactionist intervention.Finally, learners are tested again on a posttest (Sternberg & Grigorenko, 2002).The intervention (or teaching) is not recognized as a short-term, task-focused type of standardized scaffolding.Rather, it is lengthened to a teaching interval of several weeks, due to the fact that the development of complicated capabilities, such as language, requires more than merely a single task; instead, tasks are acquired and internalized over time.In fact, DA, on the whole, can be traced back to Vygotsky (1978) who emphasized the role of the social context as a facilitator of the learning process (Kozulin & Garb, 2002).The concepts of mediation, the zone of proximal development (ZPD), contingency, and scaffolding are considered the benchmarks of DA.Vygotsky's theory of learning highlights the notion of mediation since it can teach learners effective ways to use cognitive and metacognitive strategies, such as in a problem-solving activity.DA has attracted several research endeavors (e.g., Leung, 2007;Poehner & van Compernolle, 2011;Rea-Dickins, 2006;Tzuriel, 2011) and has also been implemented in classroom-based assessment (Ableeva & Lantolf, 2011;Sternberg & Grigorenko 2002).In DA, teaching and testing are integrated into a cooperative activity which addresses the involvement of learners' cognitive and metacognitive processes (Ableeva & Lantolf, 2011).Empirical studies (e.g., Gass, 1997;Khezrlou, Ellis & Sadeghi, 2017;Lidz, 2002;Swain, 2001) have demonstrated that learners turn into co-constructors of meaning in collaborative joint activities in which knowledge and meaning can be argued and mediated.These activities must be context-bound.
In a study by Ableeva (2008), the effect of DA on listening comprehension was investigated.The results showed that DA had a positive effect on learners' listening skills after the treatment.Ableeva and Lantolf (2011), in another longitudinal study, demonstrated the potential of applying DA in fostering mental processes in listening comprehension in French.This study also showed significantly positive findings.There have been other research studies on testing related to collaborative interactions in language skills such as speaking (Fulcher, 1996;Swain, 2001); however, scant attention has been paid to such interactions in the form of oral narrative tasks.Furthermore, the results of previous studies have been inconclusive regarding the role of gender in the role of DA, with some studies showing significant differences (e.g., Hashemnezhad & Fattollahzadeh, 2015;Alavi & Taghizadeh, 2014), and others showing non-significant differences between male and female learners (e.g., Frisby & Braden, 1992;Ramazanpour, Nourdad & Nouri, 2016).However, as far as the researchers of the present study are concerned, there has been no previous study examining the role of DA in narrative production by both male and female learners.In the following section, the significance of narrative tasks in terms of cognitive processing and interactional value is discussed.

Narratives in L2 Learning
Narrative availability across various contexts, cultures, and times is an undeniable fact.Narratives have always been used as a means of conveying culturally significant information (Reilly et al., 2004).In other words, storytelling is universal and timeless (Juqueira, 2010).Human beings have an innate storytelling ability, which is believed to be a semiotic skill; narratives make it easy for people to make sense of the world, as well as their own personal experiences (Bruner, 1988;Johnstone, 2001).Bruner (1988) affirms that narratives are of great significance not only for the sake of interactional functions, but because of narratives' ability to empower knowledge, memory, and the organization of discourse.Cross-cultural differences are possible while telling a story; in other words, narrative approaches, as well as methods of relating them, are considered to be culturally diverse, and many still remain undiscovered (Junqueira, 2010).
Regarding second language learning, Junqueira (2010) believes that storytelling has moldered away; in spite of the fact that telling and hearing stories are counted among the first tasks language learners commonly tackle.It is indeed an uphill struggle for language learners to devote themselves to efforts to understand one another's stories, particularly when the language learners are from different cultures.
There have been many studies done on narratives, and the results suggest that narratives reflect both cultural and linguistic influences, even if there is an underlying narrative schema (Tannen, 1993;Boudreau, 2008).
There is no concordance between the definitions suggested by linguists for a piece of narrative, in that they believe there is no clear-cut definition for a narrative (Celinska, 2004).According to Bloom (2003), there are plenty of definitions proposed for this particular act.These definitions correspond with different research paradigms, methodologies, and purposes; in light of this, Gazella and Stockman (2003) claimed that "narrative is a word in danger of being overused" (p.65).Labov (1997) advocated for a more effectual definition, stating that "narrative is a method of recapitulating past experience by matching a verbal sequence of clauses to the sequence of events which actually occurred" (pp.359-360).Notwithstanding Labov's point of view regarding the definition of narrative, Joshntone (2001) decided to apply the word "narrative" to refer to a sequence of events that happened in the past, and "story" to refer to the act of recounting those events.In distinguishing between narrative and story, Le, Coelho, Mazeiko, and Grafman (2011) asserted that a narrative is the text of a story, and the researchers emphasized that narratives do not exist by themselves, as narratives and stories are not separable.As stated by Georgakopoulou and Goustsos (1997), "story" is a term used by ordinary people to refer to "narratives," as "narrative" is considered to be specialists' term.By and large, there are many researchers who believe narratives to be a series of actions in the past (McCabe, & Bliss, 2003;Schneider, Hayward, & Vis Dubé, 2006).
Identifying "certain formal characteristics" is another factor in the process of defining narratives (Stadler, & Ward, 2010;Toth, 2011).Concerning the definitions of narrative mentioned above, there are two key elements to take into account, according to Stadler and Ward (2010): in fictional texts assumed to be narratives, the events recounted are time and place-bound--occurring in the past and at a particular place.
For Stadler and Ward (2010), however, narratives are stories that embrace some extent of transformation.Burner (1991) expands on this, saying that "stories achieve their meanings by explicating deviations from the ordinary in a comprehensible form" (P.47).
In oral narrative tasks, research has investigated which aspects of the characteristics of tasks have the most prominent impact.Narrative performance can be highly influenced by the contextual support provided in the tasks, such as the use of pictures, picture books, memory, and films.Different contextual support provided by these factors and prior knowledge of the topic can affect the complexity of the produced narrative (Fiestas & Pena, 2004).Research studies with bilingual participants have mainly used pictures, picture books, and film retellings, since these techniques do not require learners to imagine events and as a result eliminate much of the cognitive load of the task (Berman, 1995).A very wellknown prompt is a picture book developed by Mayer (1969) that is used in different cross-linguistic studies with bi-and trilingual children (e.g., Cenoz, 2001;Kellerman, 2001;Pearson, 2002;McKinnie & Priestly, 2004).The benefits of using Mayer's book for the elicitation of narrative discourse are the accessibility of monolingual corpora in several languages, as well as the analytical model.
The present study, too, employed picture-cued oral narrative tasks for the purpose of contextualizing oral language production.Participants' oral production was scaffolded through the use of dynamic assessment in order to observe the effectiveness of this approach in fostering one of the most difficult skills for EFL learners.In the present study, the following research questions were addressed: 1. Does a dynamic assessment of EFL learners' narrative performances enhance their second-language oral production? 2. Is there any statistically significant difference between male and female participants with regard to their scaffolded oral productions?

Method
Participants A total of 60 EFL students from two private universities in Tehran participated in the study.Majoring in subjects in the Arts and Humanities, 34 were freshmen students and 26 were sophomores, ranging in age from 19 to 23.In order to evaluate the effect of gender, participants included both male (N = 26) and female (N = 34) learners.None of the participants had ever been to an Englishspeaking country and all consented to take part in this study.Learners' level of proficiency was determined to be lowintermediate (M = 77.64,SD = 12.89) according to the Preliminary English Test (PET).The experiment was conducted in their General English courses where one whole class was assigned to be the experimental group (N = 30) and the other class served as the control group (N = 30).

Instrument Oral Narrative Task
All participants were involved in performing narrative tasks using six-frame cartoons (see Appendix).A total of four narrative tasks were used in the present study, all of which were adopted from Heaton (1997).All the narrative tasks were completed in each testing session.The narratives included: 'A surprise', 'The chase', 'Waiting for a bus', and 'John and his boxes'.All of these cartoons are similar in terms of the structure since they include six frames and a clear punch line.Each story has a different topic.The orally narrated stories were audio-recorded and subsequently transcribed.Rater reliability was estimated by comparing the percentage of agreement between the author and a Ph.D. student on a sample of 2 oral productions from each task.The percentage of agreement ranged from 85.9% to 94.7% for the oral narrative tasks.

Procedure
The 60 participants in the two classes were assigned to two groups, an experimental group, and a control group.The students in these classes were learning English as a foreign language.The first task, namely 'A surprise' was administered in the first session and acted as the pre-test.The following tasks served as the post-tests.The scores were added up and then divided by the total number of tasks for the final post-test score.In the control group, subjects' oral skills were practiced using the traditional method, with no teacher mediation after the pre-test.In other words, although the control group learners carried out the same oral narrative tasks, they did not receive any mediation.In the experimental group, on the other hand, a dynamic approach was used in the class after the first narrative task (the pre-test).DA in the experimental group consisted of mediation between the teacher and the subjects, such as the provision of clues, descriptions, and clarifications, suggestions, prompts, and more crucially guiding questions by the teacher.Put differently, this study was based on the pre-testmediation-posttest design, called the sandwich model, of DA.In this format, the procedure included a mediated learning phase and a post-test.The mediation program was aimed at supporting students in their development of narrative production.The results of the pretest were used to indicate difficulties that learners faced.The mediation program provided participants with several (3 -6) sessions, each approximately 20 minutes long.The mediator offered feedback, gave explanations, and mainly focused on giving hints about answers while not revealing the answer.After sessions of the mediation program, participants were given a static posttest.Subjects' performance on the posttest was compared to that of the pre-test in order to determine how much improvement was made.

Scoring
Global speech quality.The oral narrative tasks instructed students to use appropriate language for telling the story behind the pictures and to do their best to produce a story in such a way that would be accurate, fluent, and native, as if in the absence of the cartoons.Two independent raters were asked to rate the oral productions for global quality.The raters were trained to provide a general rating for the quality of the speech with respect to five criteria, using any numerical value they considered proper.The raters had the opportunity to refer to the specified criteria while rating the transcribed L2 productions.The five criteria were selected on the basis of a scaling procedure (Blok, 1986), including a very weak speech (10th percentile), a weak speech (25th percentile), an average speech (50th percentile, which had a score of 100), a good speech (75th percentile), and a very good speech (90th percentile).
Raters were asked to focus on content and flow of ideas, and to note language errors only when they blocked the understanding of the story.

Results
Firstly, a Kolmogorov-Smirnov test was carried out to ensure the normality of the scores obtained from the rated oral narratives.The results show that the narrative scores were normally distributed.The analysis of covariance (ANCOVA) was conducted to provide an answer to the first research question of the study.Put differently, the reason behind the utilization of ANCOVA was to consider the pretest scores as a covariate in order to see subjects' development from the pretest to the posttest.The major assumption of ANCOVA is the homogeneity of variances; as a result, Levene's test was applied.Table 2 shows the results of the homogeneity of variances.The homogeneity of variance assumption (F = 5.59, p = .021,p = .02)was violated, as the p-value for Levene's test (p = 0.21) was less than .05,and F observed was 5.59, with (58, 54.63) degrees of freedom.Therefore, an independent samples t-test was utilized to compare the mean differences between the pre-test and post-test in oral conferencing and collaborative writing groups.
Having obtained the assumptions of an independent samples t-test as a parametric test (i.e., the normality of data), the next step was to conduct the t-tests.First, the results of descriptive statistics are exhibited in Table 3.As the mean and standard deviation scores in Table 3 demonstrate, there were differences between the experimental (M = 11.73,SD = 2.31) and control (M = 10.20,SD = 2.98) group learners' performance in the narrative post-tests on the whole.These mean scores show the superiority of the experimental group over the control group.However, in order to get more accurate and reliable results, an independent samples t-test was conducted The results of this test are displayed in Table 4.The results of an independent samples t-test show a statistically significant difference (t (58) = 2.22, p= .03) between the experimental and control groups in the post-test.The descriptive statistics, too, point to the same finding, showing that learners in the experimental group (M = 11.73,SD = 2.31) outperformed those in the control group (M = 10.20,SD = 2.98).The results are better clarified in Figure 2.
In an attempt to identify differences between the male and female learners' performances in the narrative tasks, an independent samples t-test was run.First, groups' descriptive statistics are reported in Table 5.Table 5 shows that there are differences between the male (M = 13.25,SD = 2.59) and female (M = 8.50, SD = 2.11) participants' performance in oral narrative tasks.The results of t-tests are indicated in Table 6.
The results of the independent samples t-test showed statistically significant differences (t (58) = -4.75,p = .000)between male and female subjects regarding their speech productions.The results show that male students scored higher than female students in their oral narratives.

Discussion
The present study intended to examine the potential of dynamic assessment in enhancing EFL learners' oral production skills.For this purpose, both male and female participants took part in this study and were exposed to mediation conducted by the teacher for five sessions.The results of data analysis revealed a significant effect for group differences, as the DA group significantly outperformed the control group.Furthermore, male participants were found to be score higher in their speeches than their female peers.This finding is consistent with Nicolopoulou's (1997) emphasis on the value of a sociocultural approach in the evaluation of storytelling's place in learners' formations of reality, as well as in the development of identity at both a personal and a larger societal level.The application of both linguistic and sociocultural models in the investigation of learners' narratives presents opportunities for understanding the complex elements of narratives, as in this study.
The role of learner-centered scaffolding teaching in enhanced narrative performance is significant when one considers the meaning-negotiation purpose of narratives.As regards the meaning-making capacity of narratives, the process of moving from a beginning to an end by setting off a chain reaction should be an essential ingredient."Unity" and "integration of thought" are two inevitable corollaries of the narrative.Based on conjecture, it can be suggested that deploying narrative in the classroom would facilitate students' use of "unity of concepts" and "subjects" (Karabinar, 2012).It is presumed that the meaning-based nature of narratives in the present study enabled learners to better adapt to the provision of scaffolding by the teacher, instead of when faced with a static non-interactionist method of assessment.
The capabilities essential for production and comprehension of narratives constitute well-established linguistic and cognitive skills (Klecan-Aker & Gill, 2005;Klecan-Aker & Colson, 2009), due to the fact that the storyteller is expected to use a combination of mental, semantic, and pragmatic knowledge types (Le, Coelho, Mazeiko, & Grafman, 2011).More specifically, the complicated nature of narratives results from the correspondence between the narrative's structure, meaning, lexical items, and relevant context.This complex structure of narratives makes it essential for learners to be supported by a mediator, whose purpose is helping learners overcome the challenges of narratives, as was the case in the present study.
The role of DA in narration can also be significant in terms of contextual elements.In addition to the language used in narratives, contextual components also play a significant role in telling a story.For instance, the context where the narrative occurs and the common ground shared by the storyteller and the listener of the message obviously impact the narrator's selection of lexical items and the amount of information to transfer (Khezrlou, 2018;Jorgensen & Togher, 2009;Ukrainetz & Gillam, 2009).The context in which the story is told determines the contextualization of the narrative speech (i.e., learners personify the objectives of the speech and the audience, outcomes of the talk, and other elements) For instance, Curenton and Justice (2004) argue that a contextualized narrative can transfer meaning through the use of hints from context, common ground, and extralinguistic tools such as body language and facial expressions.This was especially the case in the present study, which helped the teacher understand learners' difficulties and provide the required assistance.However, it needs to be noted that this approach might not be effective in some environments where narrators must make use of decontextualized information to deliver meaning when they narrate stories.The integration of a contextualized narrative framework through the presentation of pictures to learners, as well as the provision of DA, are assumed to have worked in combination to lead to the positive results in the experimental group.
With regard to the gender effect, the results showed that males were inclined to receive higher scores in the production of narrative speech.This finding corroborates the results of previous studies referring to the gender differences in language production (Khezrlou, 2012;Lorenzo-Dus & Bou-Franch, 2003;Okamoto, 2002).Moreover, the findings of the present study are in line with those of Okamoto (2002) in the sense that gender should not be separated as an independent factor for the investigation of language use; rather, other factors should be recognized as relevant at the same time.Previous studies have suggested that female learners may show possible dominance in their second language-learning process (Burstall, 1975;Boyle, 1987).With regard to these issues and considering that Iranian schools are not coeducational, the examination of gender-based differences between learners using DA techniques and its benefits for male learners could be valuable; however, there is a need for more research to shed light on this issue.

Conclusion
Assessment is a relative concept and one individual performance is not a clear-cut index of real ability.As McNamara (2004) puts it, "assessment is a process of gathering information about test takers from observed performance" (p.766).He further holds that this process includes observing testees under assessment conditions and making inferences about their underlying abilities in order to make statements about their probable performance in a nonassessment situation.Thus, the active role taken by the examinee during testing is a core difference between DA and non-DA (NDA) assessment types.As Vygotsky (1998) states it, the results of conventional assessment models look like an empty medical diagnosis in which a doctors' job is merely restating the patients ' known problem in scientific terms (p.205).Vygotsky (1998, p. 205) believes this kind of diagnosis is unacceptable and must be replaced with a kind of "True Diagnosis," something that includes "an explanation, prediction and scientific basis for practical description."DA attempts to achieve this; therefore, the conventional and traditional terms applied in NDA procedures such as generalizability do not apply in DA.This approach, as shown in the present study, proved effective in terms of oral skill development and is therefore suggested to be implemented in speaking classes.The study, however, was limited in a number of ways that warrant further research in this area.Firstly, the study was limited to the examination of oral production skills devoid of any attention given to the learners' attitudes, anxiety levels, working memory, and other individual variables.Further research is necessary to carry out replications of this study and examine these issues.Another issue is the limited number of treatment sessions, which could limit the results of this study and calls for further longitudinal research studies.

Figure 2 .
Figure 2. Experimental and control group 1

Table 1 .
One-Sample Kolmogorov-Smirnov Test for Oral Narrative Tasks

Table 3 .
Descriptive Statistics of Oral Narrative across Groups

Table 5 .
Descriptive Statistics of Gender Differences in Narrative Tasks

Table 6 .
T-test Results of Gender Differences in Narrative Tasks