Gender and Learners’ Beliefs in Corrective Feedback in Learning Chinese as a Foreign Language: A Network Questionnaire Survey Report

This network questionnaire survey aims to confirm whether gender differences exist in learners’ beliefs in corrective feedback (CF) in learning Chinese as a foreign language (LCFL). Ninety-seven (53 male and 44 female) university learners of Chinese from 33 countries participated in the survey. The statistical results indicated that there were little significant differences in male and female learners’ beliefs in CF in LCFL except in two psychological effects in which male learners seemed more emotional or grateful to teachers’ correction; that culture did not influence male and female learners’ beliefs in CF; that self-correction and teachers’ correction are more welcomed than peers’; that immediate correction in specific contexts was preferred; that teachers were supposed to tell the learners where the mistake was before correcting it; that it was not confirmed whether direct or indirect correction was better; that learners’ preference of frequency of CF was between sometimes or often; and that negative effects of CF were not found. These findings may contribute to the Chinese teachers’ understanding of learners as individuals and to correction methods teachers adopt while confronting learners’ errors in teaching Chinese as a foreign language (TCFL) classes.


Introduction
The difference between male and female is something that everybody knows and nobody knows (Money, 1987, 13). A long history of research has shown that females are superior on many measures of memory, verbal tasks without considering the types of memory, like episodic memory, and the quality of the speech (Halpern, 2012, 115-123). Numerous indicators of gender differences in second language acquisition (SLA) or foreign language learning (FLL) have been presented by scholars and researchers (Aliakbari & Tazik, 2011;Ehrlich, 1997;Kobayashi, 2002;Tercanlioglu, 2004;Zoghi et al., 2013). And the relationship between learners' beliefs and the preference of CF has been studied (Ding & Zhao, 2019;Zu & Ma, 2015). Results display that teachers may not be aware of the students preferred CF types in written errors (Alshahrani & Storch, 2013). The preference of CF type is closely related to anxiety. For example, recast and metalinguistic feedback types were better rated by the students who reported higher levels of anxiety in oral communication classes (Martin & Valdivia, 2017).
However, research on gender and learners' beliefs in CF in LCFL is rarely reported. Do gender differences exist in learners' beliefs in CF in LCFL? This is a crucial issue in LCFL or TCFL that the author tries to explore to contribute to LCFL or to SLA.

Gender in SLA
Gender's role in SLA or FLL has been a controversial issue for researchers, educators, and scholars for over half a century. Earlier research in gender differences in SLA or FLL can be divided into two main streams: One features performance-related differences, learning strategy differences and language learning as being a particularly female activity; The other manifests more social constructionist views of language and gender (Feery, 2008). Gender differences are studied from the perspectives of sociolinguistics, ethnology, pedagogy, linguistics, psycholinguistics, and cognitive linguistics, etc.
Much evidence seems to have shown that females tend to have better verbal abilities than males, males are better than girls in mathematics. Bećirović's (2017) findings indicate that female students are more successful at learning English as a foreign language than male students at each group/grade level. But females' advantage in verbal abilities is likely to be small and depends on the type of verbal ability that is measured (Halpern, 2012, 119). And their language proficiency is a genetically-influenced ability interacting with environmental factors such as motivation, orientation, education, and learner strategies that still mediate between endowment and acquiring language proficiency at an adult stage (van der Slik et al., 2015). Hyde et al. (2008) research in several states in the United States indicates that there was virtually no difference in the average math scores for girls and boys from Grades 2 through 11. But, as the authors noted, the items in the tests required lower level thinking rather than higher level thinking. Besides, races play a dominant role in the significance of sex differences in mathematics at advanced level. For White students, predominance of boys to girls among the highest scorers was obvious, but it was vague for Asian/ Pacific Islander students (Halpern, 2012, 146). Karlak & Medve's (2016) research reveals that female learners were significantly more motivated than males in certain motivational dimensions, but male learners of English achieved statistically significantly higher results than female peers in standardized evaluation. However, in German no statistically significant gender differences were found, irrespective of the measure of communicative competence.
Besides, males and females are proven to be different in cognitive styles or psychological differentiation (Witkin et al., 1962;Kao et al., 2017). Evidence suggests that girls were better at recognizing faces than boys (Rehman & Herlitz, 2006;Sasson et al, 2010). Regarding the interaction of age with gender, females consistently outperformed males in the reported strategy use, in all grade levels and categories except for the social (Platsidou & Sipitanou, 2015). Females were found to be more motivated than male language learners, and have more positive attitudes toward speakers of the target language (Gardner and Lambert, 1972;Muchnick and Wolfe, 1982). On the other hand, in the research on the gender differences in affective variable in FLL, females showed higher anxiety than males though they earned higher scores on exams (Bell & McCallum, 2012). But it was also reported that boys, but not girls, used negative attributes for self-description (Martínez-Marín, 2019). Kriegbaum et el's (2018) study found no significant moderator effects for gender. Males and females exhibited different advantages and disadvantages in learning online within the personal learning environment (Wu & Cheng, 2019).

CF and Language Learner Beliefs in SLA or FLL
CF refers to "any indication to the learners that their use of language is incorrect, including various responses that learners receive" (Lightbown and Spada, 1999, 171-172). In a narrow sense, CF was usually categorized into six types (Lyster & Ranta, 1997), including explicit correction, recasts, clarification requests, metalinguistic feedback, elicitation, and repetition. But in a broad sense, it takes two forms: explicit and implicit or direct and indirect Karim & Nassaji, 2020;Kim et al., 2020;Lyster et al., 2013).
Language learner beliefs in SLA or FLL were initially proposed and then were theoretically put forward by Horwitz (1985;1988) for the well-known Beliefs About Language Learning Inventory (BALLI). The current research on language learner beliefs in SLA or FLL often revolves around the following eight questions based on the early five ones (Ellis, 2009;Hendrickson, 1978;Zhu & Wang, 2019): 1. Is it necessary for the teacher to correct learners' errors? 2. What errors should be corrected? Errors in spelling or grammar or pronunciation or reading aloud or structure, or any other errors? 3. Who is to correct the error? It is the teacher or the student himself or herself or the peers who should correct the errors? 4. When to correct? Should errors be corrected immediately or afterwards? 5. How to correct? Should errors be corrected explicitly or implicitly (or directly or indirectly)? Or by using any or some of the six types mentioned above? 6. Which type of CF is the most effective? 7. How often is to correct? Always or often or sometimes or seldom correct? 8. What effects of CF produces on learners? Positive or negative effects? So far, the research on the gender and learners' beliefs in CF, especially in CF in LCFL is rarely reported. Therefore, it is of vital importance to conduct this research to fill the gap in this aspect.
This article is to compare male and female learners' beliefs in CF in LCFL with the aim to determine whether gender influences learners' beliefs in CF in LCFL to eliminate gender-based stereotypes in language learning. Moreover, it tries to uncover what male and female learners perceive about CF in LCFL and how they expect the teacher to correct their errors. It is hoped that the results of this study have some potential implications for teachers to improve teaching pedagogy in TCFL and considering learners as individuals instead of simply different gender groups.

Research Questions and Hypotheses
This article explores the relation between gender and learners' beliefs in CF in LCFL. The questions the author raises are: 1. Do male and female learners perceive CF differently in Chinese classes? 2. Do male and female learners anticipate different types of CF from their Chinese teachers?
3. Do cultural factors influence male and female learners' beliefs in CF in LCFL since they are from 33 different cultural backgrounds?
To answer the above questions, we pose the following hypotheses:

Methodology
In this network questionnaire survey, the quantitative research was applied to discern whether gaps exist between male and female learners, hoping to find some constructive clues for LCFL or TCFL. A 5-point Likert scale was mainly used to measure students' beliefs in CF in LCFL.

Participants
Ninety-seven international students, including 53 males, 44 females, participated in this on-line survey. The participants had studied Chinese for 5 months to 5 years, a great majority of them were studying in different universities in Anhui Province in China. They were from 33 countries in Asia, Africa, Europe, island country of Southwest Pacific Ocean and South America.

Instrument
The network questionnaire survey on learners' beliefs in CF in LCFL is made up of three sections with 56 items in total. Section 1 was about demographic data, items 1-4 in this section were: To know more about how and when to correct in in oral and written classes, Section 3 was designed, which included 5 questions. Different from those above, the choices in the section were more detailed and concrete. Choices in items 52-55 were: 1=immediately correct; 2=correct after I finish speaking; 3=correct me personally after class; 4=correct together with other peer' errors; 5=not correct. The choices for item 56 were: 1=correct together with peers' errors; 2=The teacher underlines the wrong places, then the students correct them by themselves; 3=give the correct forms directly; 4=correct face to face; 5=The teacher provides the related network link addresses, and then the students search for the relevant information themselves; 6=not to correct; 7=other methods.
To guarantee the quality of the questionnaire survey, the items of the same category were maximally separated; affirmative and negative items were alternated; and the same issue was asked occasionally twice in different ways to avoid the participants' inertial selection.
Wen juan xing (https://www.wjx.cn/), an online questionnaire survey platform in China was used in distributing, collecting and analyzing the questionnaire survey. The statistical technique employed for comparison ier.ideasspread.org International Educational Research Vol. 3, No. 4; and analysis in this platform was a paid on-line SPSS software through using one-way anova. It was used to calculate the mean of the male and the female learners' choices, and finally test if the observed data supported the hypotheses.
Gender was considered as the single independent variable, and all the other 52 items in Sections 2 and 3 as dependent variables.

Data Collection and Analysis
Learners were invited to complete the survey at we-chat without time limitation. After their submission, the data were automatically collected and calculated at wen juan xing. By using one-way anova, the results of learners' beliefs in CF in LCFL were obtained.
The following is about the results of one-way anova about male and female learners' beliefs in CF in LCFL in Section 2 (see Tables 1) Results indicate that statistically significant differences were not found between gender and perception of committing verbal errors in Chinese classes and eight issues (p>0.05) except in items 26 and 27(* p<0.05 ; ** p<0.01). ier.ideasspread.org International Educational Research Vol. 3, No. 4; The mean in each item in Tables 1 demonstrates that male and female learners showed little difference in perception of committing language errors. Concisely, both male and female learners agreed or quite agreed to the normality to commit errors in the process of Chinese learning and necessity to correct errors (whether to correct). They showed little difference in what to correct, too. Specifically, they agreed to or strongly agreed to error correction in oral Chinese practice class. As to whether every mistake or common/habitual or grammar mistakes should be corrected, their choices were between not sure and agree. However, they agreed or strongly agreed to oral and written correction. Regarding the items whether pronunciation or grammatical errors in class should not be corrected if they do not interfere with comprehensibility, their choices were close to disagree. Concerning who is to correct, their choices were also similar to each other. If we arrange them from strongly agree to strongly disagree, they are self-correction under guidance (but learners hesitated when teachers' correction and selfcorrection were mixed), teachers' correction, and peers' correction. Likewise, they were consistent with each other in when to correct. Specifically, both male and female learners agreed or strongly agreed to immediate correction of errors in pronunciation, vocabulary, grammar and language use. As to immediate correction of errors in answering questions, reading the text aloud, males (4.06; 4.04) showed a little bit stronger desire than females (3.89; 3.91). But for the general statement of immediate correction of errors(items 10 and 13), the means were lower than those for the concrete ones (items 7, 8, 16, 17, 18 and 19). Equally, they displayed little discrepancy in how to correct. They agreed or strongly agreed to the statement of telling students what is wrong before telling them the correct form. The means of the females in telling directly the correct form (4.11) and explaining the relevant grammar rules (4.05) were a bit higher than those of the males (3.91; 3.96). They showed no certainty in direct correction "No, you should not say in this way, you should say...". Male and female learners didn't show any significant differences in how often to correct, either. Their means were approximately between sometimes and often whether it was in pronunciation, lexical, grammar or language use errors, with female learners' means (4.05; 3.91; 4.20; 4.11) respectively a bit higher than male learners '(3.91; 3.83; 3.98; 3.92). Moreover, they demonstrated little differentia in which type of CF is most effective. Detailly, it was not obvious as whether direct or indirect correction is better, for all their means were below 4. They also showed little difference in the effects of CF except in except in items 26 and 27. Specifically, significant gender difference existed in "I feel lucky when my teachers correct my errors in class" for its p was less than 0.05(F=5.295，p=0.024). The mean of the male in this item was 3.87, obviously higher than that of the female (3.27). Great significant gender difference also appeared in "I feel grateful to my teacher for correcting my errors" for its p was less than 0.01(F=7.331，p=0.008).
And the mean of the male in this item was 4.17, obviously higher than that of the female (3.52). Overall, they agreed to or strongly agreed to the positive effect of CF provided by teachers. But the effect of peers' correction was not as apparent as that of teachers'. The positive psychological effects were also revealed from the means. Teachers' CF did not make learners unwilling to speak, afraid to speak, feel humiliated, nervous and anxious, or reluctant to answer the questions for a period. But it was not certain whether learners would feel cheated without teachers' CF.
Then, the form of the questionnaire was transformed into 5 questions to know more about how and when to correct in oral and written classes. Here are the results of one-way anova about male and female learners' beliefs in CF in LCFL in Section 3 (see Tables 2). As the table reveals that male and female learners did not display any significant differences in when and how to correct in oral and written courses, either (p>0.05). Their choices in items 52-55 mainly focused on correction after finishing speaking in oral practice class. As to the written errors, their choices were between underlining the wrong places for students to correct them by themselves and teacher's giving the correct forms directly.

Discussion
Both male and female learners demonstrated little significant differences in perception of committing verbal errors in Chinese classes, and in whether, what, who, when, how, how often, which and what effects in CF in LCFL. These results accept H 1 and H 2 except in items 26 and 27. Male learners showed more gratefulness and sensitivity to teachers' CF than female ones, which is quite beyond the traditional images of males who are masculine and insensitive to emotions. A further research has found that "The human male is, on most measures, more vulnerable than the female. Part of the explanation is the biological fragility of the male fetus, which is little understood and not widely known." (Kraemer, 2000) It is the biological factor that partially accounts for the emotional gap between male and female learners. Therefore, masculine stereotype on males is harmful to learners and teachers. This finding enables the author to stand firm on the belief that males, like females, are individuals with similarities and differences.
However, the results do not accept H 3, i.e. culture did not categorize these international learners into male or female group with different beliefs in CF in LCFL although they were from 33 cultural backgrounds with diverse educational customs. In other word, male and female learners' beliefs are not influenced by their cultures.

Limitations and Suggestions for Future Research
This paper compared male and female learners' beliefs in CF in LCFL by means of one-way anova. The results found in the research may be conducive to LCFL or TCFL. Yet, limitations in the research constrained the researcher from going into the deeper side of the topic.
Firstly, all the results in this survey were based on a network questionnaire survey without other any parameters like interviews, learners' achievements, etc. Next, the 5-point Likert scale remained unaltered without flexible options that learners may choose from. Finally, the reasons for the choices were not supplied in the questionnaire survey. As a result, it is hard for the author to analyze the reasons when differences emerge.
Multiple research methods might be adopted for future research in the similar topics, case studies, ethnographic studies, narrative inquiry, etc. They may involve more detailed designs in each specific category to explore gender and learners' beliefs in CF in LCFL.

Conclusion
Results in one-way anova reveal that broadly there were little significant differences between male and female learners in CF in LCFL. These results echo the findings in Halpern (2012). The author compared male and female beliefs in CF in LCFL, she did not mean to distinguish males from females in learning. But rather she treated them equally or as a single group with different emotions and motivations in LCFL.
For a long time there has been existing a serious and common misunderstanding about sex differences: women are like this, men are like that, then treat each sex differently according to these differences. As a result, many similarities between males and females are obscured, and people as individuals are neglected (Halpern, 2012, 151). However, in modern scientific research every learner is an individual with individualized character, verbal ability, aptitude.