Influence of gender and academic ability in a computer-based Spanish reading task.

For the last several decades, research dealing with gender differences has consistently found that girls and boys differ to some degree in their academic achievement and their academic interests. The difference is often partially attributed to social stereotypes which may be the factor pushing females to choose more verbally oriented disciplines such as language arts, whereas males are drawn toward more quantitatively oriented disciplines such as sciences and math (Benbow, 1988; Callahan & Reis, 1996; Feingold, 1988; Lubinski, Benbow, & Sanders, 1993). The efforts made by schools to counteract these stereotypes have been quite successful over the years and have managed to narrow the gap between the genders significantly, with one notable exception--the stratum of high ability males and females from high-school age and beyond. It is in this particular group that gender differences remain virtually unchanged. According to Callahan and Reis, "Although the gap in achievement test scores between males and females may be decreasing, the gap that still remains seems largely attributable to differences in scores among the top 10% to 20%" (1996, p. 184). This gap is in favor of boys where quantitatively oriented subject matters are concerned. Girls have been reported to more frequently choose, and to consider less challenging, verbal disciplines such as language arts and foreign languages (Swiatek & Lupkowski-Shoplik, 2000; Olszewski-Kubilius & Turner, 2002).

Within this framework, the aim of the present study is to determine if there are differences in the attitudes and performance of male and female learners of different ability levels when faced with a foreign language reading task. After a previous study (Nikolova & Taylor, 2003) in which we showed that the level of ability had a bearing on student scores in a creative reading task, we took our analysis a step further by examining in detail the part gender combined with ability level played in these differences. It is our contention that the more learner characteristics we can identify as crucial in performance of these reading tasks, the better we can attune our teaching tools and materials.

Review of the Literature

In the Nikolova and Taylor (2003) study, students were identified as high ability or average ability for the purposes of the investigation, and randomly assigned to two versions of a reading task. One task, the experimental method, required the students to read a short passage in Spanish on a computer screen and annotate target words, with the help of a dictionary. The control group read the same story but with the target words already annotated on the screen by the experimenters. Both groups were then tested on immediate and delayed (one month later) vocabulary retention. Results showed that the high ability students recalled significantly more target words in both the immediate and the delayed tests, and remembered significantly more idea units from the story than the average ability students in the experimental group. No differences, however, were found in the control group.

In an earlier study, Nikolova (2002) found a difference between students performing an annotation task compared to a control group during a French reading task. In a separate study (Nikolova, 2000), no differences were found in attitude between students in the annotating group and those in the nonannotating group in a French reading task.

This study involves several areas of research in the domains of foreign languages, educational psychology and computer-assisted language learning (CALL). Therefore, this section will discuss each area separately and attempt to show connections among them as is germane to the study at hand.

Issues in the Study of Gender Differences

Some researchers argue against the study of sex or gender differences. The study of sex differences is controversial for a number of well-founded reasons. Researchers point to the danger of creating a self-fulfilling prophecy by the publication of differences between the sexes, and a continuation of biased educational practices that favor one sex over another (Baumeister, 1988). Furthermore, it is argued that the research on gender differences overemphasizes the differences between sexes and disregards the differences within sexes (Hare-Mustin & Marecek, 1994; Hollway, 1994). However, Halpern and LaMay (2000) maintain that:

   ...despite the dangers inherent in answering questions
   about group differences, censorship, even self-censorship,
   does not promote equality and can be far more
   dangerous and counterproductive than directly addressing
   the question. Stereotypes and prejudice are not
   caused by an open process of scientific inquiry; in fact,
   they seem to flourish in the absence of data. (p. 233)

Nevertheless, it is not our intention to dismiss the cautions about research in sex differences. In an excellent paper, Caplan and Caplan (1997) review the body of literature that has shown sex differences in mathematical, spatial and verbal abilities. In discussing verbal ability, for example, they argue that the staggering number of ways that the construct has been operationally defined by various studies (they list 32), leaves us without a clear understanding of what verbal ability is. They point out, furthermore, that sex differences are not found invariably, they do not exist at all ages, and they account for very little of the variance in individual performance. However, keeping these cautions in mind, we continue to concur with the opinion of Halpern and LaMay (2000) that research in gender differences is important and informative.

Giftedness and Gender

Though the issue of gender differences is not clear, as stated in the introduction, there appears to be a marked differentiation between males and females in the population of high ability adolescent learners. Some reasons for the discrepancy may be differences in socialization or in learning styles and preferences.

Research on gender differences in gifted students has shown that academically talented female students may be socialized to hide their giftedness. Women often believe they must decide between academics and being found "socially valuable" (Noble & Smyth, 1995, p. 49). Reis and Dobyns (1991) report findings that academically talented women underachieved in comparison with academically talented males (see also Lindley & Keithley, 1991). In addition, there is little institutional support at the university level for these women, and when there is support, often women do not seek out such programs (Nolden & Sedlacek, 1997; Wolleat, 1979). According to Noble (1987), at least half of the students at the elementary level identified as gifted, talented, or highly capable are girls, but by junior high school, girls make up less than a quarter of gifted students. She points out that almost all academically talented girls and women have had occasion to hide their abilities to survive socially, having been conditioned by parents and teachers to see themselves as less capable than males, to avoid risk-taking, and to have lowered expectations for success. In addition, both male and female teachers seem to perceive those same qualities that they view positively in academically talented males as negatives in females.

There may also be gender differences in learning style preferences. There is support from Dimitrov (1999), who found that high ability junior high school boys have a predilection for open-ended tasks and score better than high ability girls at the same grade level when confronted with more creative tasks, at least in the fields of math and science. Other factors which have been proposed to account for the discrepancy between male and female academically talented students are issues of assertiveness, fear of success, and external versus internal attribution of ability (Fox & Ferri, 1992; Hollinger & Fleming, 1984; Luscombe & Riley, 2001; Noble, 1987).

Verbal ability is an area in which female students have consistently shown superiority to male students (Caplan, Crawford, Hyde, & Richardson, 1997; Maccoby & Jacklin, 1974; McGlone, 1986). However, in a meta-analysis of 165 studies representing nearly a million and a half subjects, Hyde and Linn (1988) report negligible differences in seven verbal abilities including vocabulary and general verbal ability. Furthermore, they compared studies published before and after 1973 and found a slight decline in the magnitude of sex differences.

Gender and Language Learning

In a similar vein, there has also been much evidence that females are better language learners than males (Burstall, 1975; Carroll, 1975; Lewis & Massad, 1975; Lynn & Wilson, 1993). The Lynn and Wilson study is particularly interesting. They studied students learning Irish in Ireland between the ages of 8 and 14. Because the study of Irish as a second language is required of all students, it gave the researchers an opportunity to study second language learners across the whole ability range, rather than being restricted to the more select number of students who study foreign language in countries such as the United States. Substantial differences were found favoring girls and the differences remained constant throughout secondary school. These findings led the authors to reject the explanation of social expectations theory that society expects girls to be good language learners, and that secondary school girls are more conscious of conforming to these expectations than primary school girls Instead, they favor the theory that sex differences have a biological basis (Benbow, 1988; McGlone, 1986).

Cross (1983), on the other hand, found that boys outperformed girls on a battery of nine tests of French. He attributed this to the fact that the teachers were men, providing a role model for the boys. Role modeling of teachers as a factor in the performance of boys in foreign languages was touched upon by Clark and Trafford (1995). Though, according to them, teachers tended to dismiss the effect of sex of teacher on boys' performance, the authors mention that in a particular school where boys scored above the national average, the two highest achieving classes were both taught by men who shared interests with the boys.

Studies specifically examining sex differences in second language (L2) vocabulary learning have been inconclusive as well. Nyikos (1990) found that females outperformed males in learning German vocabulary, while Grace (2000) found no differences in French vocabulary learning between sexes in a computer-assisted language-learning (CALL) environment.

Gender and Computer-Assisted Learning 

Another area of investigation germane to this study is that of gender and computer use. It has been reported that males exhibit more interest in computers than females, use computers more, and generally have a more positive attitude and less anxiety (Adam & Bruce, 1993; Collis, 1985; Murray, 1993). However, more recent studies have shown that the difference has diminished (Teh & Fraser, 1995; Parish & Necessary, 1996; Taisir, 1999; Mitra, LaFrance & McCullough, 2001), with previous experience being a better indicator than sex for attitude and performance (Hunt & Bohlin, 1993). Grace (2000) found no gender differences in the ability of beginning students of French to retain vocabulary from a dialogue, while Meunier (1995-1996) found that achievement was more strongly related to personality differences and keyboard control than to gender differences.

Gender and Reading Comprehension in the First and Second Language

Gender differences in reading comprehension of L2 text have been shown for the factors of topic familiarity, background knowledge and interest (Bugel & Buunk, 1996; Carrell & Wise, 1998; Brantmeier, 2001). Bugel (1993, as cited in Bugel & Buunk, 1996) found text biases differentially affecting males and females in the Dutch national examinations of French, English, and German reading comprehension. She found that females do better with topics of education, human relations, art, and philosophy, while males do better on economic and technological topics, politics, sports, and violence. To our knowledge, no L2 studies support the contention that women are better L2 readers than men are. In fact, the commonly held belief that females are generally better readers than men does not seem to hold up beyond elementary school (Becker & Forsyth, 1990; Hogrebe et al., 1984; Sheridan & Fizdale, 1981).

Research Questions of the Present Study

The purpose of this study is to examine the performance of males and females of high and average ability in vocabulary recall and reading comprehension during a computer-based open-ended reading task requiring more learner involvement (annotating) versus a more teacher-generated and discrete activity (nonannotating) to discover if there are differences for the factors of Gender and Ability. The research questions then are:

1. Are there differences in the immediate and delayed vocabulary recall of females and males engaged in a computer-based reading task with annotating versus a computer-based reading activity without annotating?

2. Is the vocabulary recall of males and females differentially affected by their ability levels?

3. Are there differences in the reading comprehension of male and female students engaged in a computer-based reading task with annotating versus a computer-based reading activity without annotating?

4. Is the reading comprehension of males and females differentially affected by their ability levels?

5. Are there differences between males and females in attitude toward each of the tasks?

6. Are there differences between ability groups in attitude toward each of the tasks?

7. Is the attitude of males and females differentially affected by their ability levels?

Method

Participants

Participants were first-year second-semester students of Spanish at a large Midwestern university. The students were of traditional college age (18-25). There were 181 students participating, of which 90 were female and 91 male. Ninety-eight of the 181 were identified as high ability students. Of these 98 students, 54 were female and 44 male. Of the 83 students not identified as high ability, 36 were female and 47 were male. There were, therefore, four separate groupings of students: Female/High Ability (FHA), Male/High Ability (MHA), Female/Average Ability (FAA), and Male/Average Ability (MAA).

Ability level was determined by an analysis of the demographic survey questions pertaining to academic history. Students who said that they had been identified as "gifted" in high school, had taken advanced placement courses, and had GPA's above a certain level were considered high ability for the purposes of this experiment.

We want to state emphatically that we make no claims beyond the scope of the present experiment that the students who participated in this study are gifted, or academically talented, or conversely, suffer from learning disabilities. We simply identified, for the purposes of the study, those students who had higher GPA's and class rankings, and had participated in honors and advanced placement courses, and we designated them by the term high ability learners. Those students who did not fit the requirements were not so identified.

Materials

A demographic questionnaire was given to all students. This was used to ascertain general personal information and their academic history. It consisted of questions about their previous language experience and success, computer attitudes and skills, as well as questions about their GPA, whether they had been identified as gifted, or academically talented, and/or had taken honors or advanced placement courses.

An annotation program designed by the lab director at the university, called SmarTText was used for the reading tasks. It consists of a screen onto which text can be input by the author and simple procedures for creating textual, still visual, and video annotations for any word in the text.

A passage called La leyenda del origen de cafe (The Legend of the Origin of Coffee) was used. It contained 126 words, 16 of which one of the experimenters chose for annotation. The experimenter selected the passage because of its appropriateness to the level of the students, which was determined by its length and relative simplicity, as well as the preponderance of words likely to be unknown to first-year students of Spanish. (See the Appendix for the entire text).

A vocabulary test designed by the researchers was used in which students had to supply the English equivalent to the annotated words. In addition, students indicated whether or not they had known the word before that day. Students were also given a sheet of paper with instructions to write a synopsis of the story they had just read as accurately and fully as they could. Finally, students were given an attitude survey to determine their enjoyment of the task and their opinion of its effectiveness for learning Spanish. All of the above instruments have been previously published in Nikolova and Taylor (2003).

Procedures

In the interest of clarity, we propose to call the experimental group the annotating group and the control group the nonannotating group.

On the first day, the researchers went to the 10 classes and briefed the students on the procedures for the study. Students were told that their participation was voluntary and were asked to sign a consent form if they wished to participate. Students then chose random numbers between 1 and 300 from a box. From this point on, they were identified by these numbers for the sake of anonymity. Last, they were given the demographic survey already described.

On day two, we escorted students to the language media center and assigned each a computer. We gave them a 20- to 30-minute orientation on how to create annotations using the SmarTText program. Annotations are words or pictures that are placed in the margins of a text or are accessed through hypertext links. They help the reader to understand certain words and phrases that they may not know. In this case, the annotations were text-based. A short passage was used as a sample. The orientation lasted until every student had demonstrated that they could successfully accomplish the task of annotation. The task was simple, and students readily understood it. The training took place during the regular class period. At this point, all students received the annotation training, because they had not yet been assigned to the annotating or nonannotating group.

The following day, students again came to the language media center and were assigned computers according to their group assignment, annotating or nonannotating. All computers had the same passage displayed, La leyenda del origen de cafe. For the annotating group, a list of the words taken from the story was placed by the computer along with photocopies of pages from a Spanish-English dictionary on which the meanings of the words were found.

The principal task for the annotating group was to create text annotations for these words using the dictionary pages and the SmartText program. We instructed them to read the passage through before deciding on the proper English equivalents of the target words, cautioning them that, because words often have different meanings according to context, they should read the passage carefully for comprehension before beginning the annotation task.

For the nonannotating group, the annotations had already been created for them. Therefore, this group's task was simply to read the passage, La leyenda del origen de cafe, using the annotations as an aid to understanding the text.

Both groups were told to indicate to us when they felt that they could write a synopsis of the story. We then gave them the paper to write the synopsis, followed by the vocabulary recall test, and finally the attitude survey. One month later they were given the same vocabulary recall test.

No pretest for the vocabulary was given. We felt that the words were unlikely to be known by a first-year student, thus obviating the need for the pretest, and its possibly deleterious practice effects.

The vocabulary tests were scored one point for each correct English equivalent consistent with the item's context in the passage. No credit was given for approximations or alternate meanings (possible range 0-15). Story comprehension scores were calculated as the number of propositions the students included in their summary of the reading passage. Students' scores were the raw numbers of propositions they included in their synopsis (possible range 0-16). These 16 propositions were determined by the researchers.

Comparisons were made for the factors of Gender, Ability, and Method on the outcome variables of Immediate Recall, Delayed Recall and Reading Comprehension. Separate factorial analyses of variances were performed for each outcome variable using the General Linear Model procedure (PROC GLM) due to the unequal cell n's. T-tests were then performed to determine any significant differences between the various groupings.

The Freidman test for blocked designs was used for the attitude scores. The scores were converted to rank scores using the SAS Proc Rank procedure and then entered into a Proc GLM procedure to determine any main effect differences and interactions. An overall analysis comparing the two tasks, annotating and nonannotating, for the factors of Gender and Ability was performed. Two additional analyses, one for each task, were then performed for the same factors.

Results

Vocabulary Recall

The ANOVA for Immediate Vocabulary Recall found no main effects for Gender. However, a significant interaction for Gender and Ability was found, F(l,180) = 7.70, p = .006 (see Table 1). The overall difference in recall was largely due to the performance of average ability males (MAA) who scored well below the other three groupings with a mean of 4.94 (SD = 2.96) words recalled, compared to MHA (M = 7.27, SD = 3.71), FHA (M = 7.13, SD = 3.54), and FAA (M = 7.00, SD = 3.17).

Given that the students differed in performance due to the different demands of the two tasks, it seemed appropriate to look at them separately for any gender differences. As can be seen in Table 2, the difference for males appeared to exist largely on the annotating task, which showed MHA recalling 9.53 target words on average as compared to 5.32 for MAA, t(1, 42) = 4.59, p < .0001, effect size (d) = 1.40. Females differed less between ability levels, with high ability females (FHA) outscoring average ability females (FAA) by 8.93 to 7.14 words recalled, t(1,40) = 1.91, p = .06, d = .22. The data is displayed graphically in Figures 1 and 2. Results of the t-tests for these data are presented in Table 3.

[FIGURES 1-2 OMITTED]

For delayed recall, there was also an interaction effect of Gender and Ability, F(1, 180) = 4.62, p = .03, with again the majority of the overall difference in recall between high and average ability students found in the male sample, while females did not differ overall. Once again, as shown in Table 2, the difference was shown to be mostly in the annotating group, with MHA students significantly outscoring MAA, t(1, 42) = 5.24, p < .0001, d = 1.06. Females again did not differ between ability levels, t(1, 40) = 1.54, p = .13, d = .30). The data are shown graphically in Figures 3 and 4. Once again, the results of the t-tests are shown in Table 3.

[FIGURES 3-4 OMITTED]

As shown in Table 2, in immediate recall the annotating group outperformed the nonannotating group for both males and females - MHA: t(1, 41) = -3.98, p = .0003, d= 1.22; FHA: t(1, 52) = -4.53, p <.0001, d = 1.24. This difference maintained in the delayed recall for males, but not for females - Males: t(1, 41) = -4.28, p = .0001, d = 1.30; Females: t(1, 52) = -1.19, p = .24, d = .32. The results of the t-tests on these data are presented in Table 4. Table 5 also shows that, whereas there was no difference in performance between high ability males and females on the annotating task in immediate recall, t(1,45) = -0.69, p = .49, d = .20, in delayed recall MHA significantly outperformed FHA, t(1, 45) = 1.35, p = .02, d = .71.

Finally, Table 6 shows that FAA students significantly outperformed MAA students on the nonannotating task for immediate recall, t(1, 42) = 2.52, p = .02, d = .76, but not in delayed recall, 1.36 versus 1.08; t(1, 42) = 1.29, p = .20, d = .39. Thus, reiterating the data for vocabulary recall:

1. For comparison of males versus females, MHA and FHA performed equally well on the annotating task in immediate recall, but in delayed recall MHA outperformed FHA. In addition, FAA outperformed MAA on both tasks in immediate recall, but no differences were found between them in the delayed recall.

2. For comparison of the sexes between recall tests, both MHA and FHA outperformed those of average ability in immediate recall, but only MHA outscored MAA in delayed recall.

3. For comparison of sexes between ability levels, on the annotating task in both immediate and delayed recall, MHA outscored MAA, but FHA and FAA did not differ.

Reading Comprehension

There were no significant main effects or interactions for Gender in relation to Ability found in reading comprehension. Because of this, no further analysis was warranted.

Attitude

Comparisons over attitude between tasks for the factors of Gender and Ability yielded no significant results. For the annotating group, no significant main effects or interactions were found for attitude, though there was a slight interaction effect, F(1, 89) = 2.81, p = .10, with MHA scoring somewhat higher than MAA, while for females, FAA scored higher than FHA.

Each question was then analyzed separately in order to see if any revealed significant differences. In the annotating group, items 1, 2, 6, and 10 revealed significant interactions, while item 8 approached significance at p = .06. Whereas MHA had higher ratings than MAA, FHA had lower ratings than FAA. This trend was found for all items in the annotating group. See Table 9 for the means, standard deviations and F statistics for the questionnaire items with significant interactions. Figure 5 shows the data graphically.

[FIGURE 5 OMITTED]

In the nonannotating group, although results were not significant for the main effect of Gender (F(1, 97) = 2.83, p =. 10, d = .17), there was a tendency for females to like the task better than males across ability levels. Questions 1, 2 and 9 showed significant differences for the main effect of Gender, with females scoring higher than males across ability levels. The same trend was observed for all questions with the exception of Question 7 in which FHA scored somewhat less than all others. Table 10 contains the means, standard deviations and F statistics for the questions yielding significant main effects for Gender. Figure 5 shows the data graphically.

ANOVA is robust for the assumption of normality even for very skewed distributions (Stevens, 1990), particularly with large cell n's. The assumption of homogeneity of variance is robust for unequal cell n's up to a difference of 1.5 to 1. In this study, no differences in group n's exceeded this ratio. Intra-class correlations (R) were calculated on the data according to a procedure outlined in Stevens (1990) and were found to be in line with the assumption of independence of observations.

Discussion

This is the third in a series of studies to test the effect of a student-centered computer-based task on the performance of beginning second language learners. The first, Nikolova (2002), found that students asked to create annotations during a reading task remembered significantly more vocabulary items afterwards than those who did not create the annotations. Nikolova and Taylor (2003) then found that this difference was concentrated in the performance of higher ability learners. The results of the present study concur with the findings in Nikolova and Taylor (2003) that high ability students fare better when confronted with a more involved task. This is probably due to the higher interest that a task of this type and difficulty arouses in more able students. In addition, high ability students are less likely to suffer from cognitive overload, which may be the reason for a weaker performance on the part of the average ability students.

Now, with the inclusion of the factor of Gender in the analysis, we appear to have clouded the picture somewhat. So, what is the relationship between Gender and Ability, and in what ways did it affect students' performance?

On the whole, it seems to us that ability level had the most impact upon the performance of the male students in the annotating group. Ability level also seemed to be a factor in retention of vocabulary for MHA students, who significantly outscored all others, including FHA students on the delayed test. MAA students on the other hand, performed significantly worse on both tasks, while FAA students scored relatively well.

In terms of the delayed recall, however, it is difficult to credit any practical significance to the findings given the low means and differences. The reporting of confidence intervals and effect sizes can be used to determine practical significance. For example, for the difference between MHA and MAA for the annotating group in delayed recall (Table 3), the large effect size of 1.06, and the p-value of <.0001 gives us some assurance of the existence of a real phenomenon in the population. However, care should be taken in interpreting the results.

Therefore, it appears that while high ability men and women do equally well on immediate vocabulary recall, average ability females perform better than average ability males (and even high ability males in the nonannotating task). We believe that we can attribute the better performance of the FAA to motivation for learning languages. Recall that in the nonannotating task, females always had more positive attitudes than males, while in the annotating task, there was an interaction of Gender and Ability. FAA students also performed respectably in the annotation task, much better than MAA students did, which may be expected given their somewhat more positive attitudes toward the task. Interestingly though, FHA students' performance essentially mirrored that of the MHA students, even though FHA attitudes were somewhat more negative on average.

We will venture here another speculation based on findings in a study by Dimitrov (1999), according to which males and particularly high ability males are more motivated and score better on open-ended, creative tasks than females. Is it not possible then to claim that, for males, the status of high ability student is often achieved thanks to a more creative and more adventurous mindset, while females are more likely to reach this status through hard work and a predilection for more traditional academic tasks?

Certainly, the results for Reading Comprehension do not support the contention that females are better L2 readers than are males, but they do make more clear the findings of Nikolova and Taylor (2003) that average ability students are somehow hindered by the annotating task. We know now that both men and women are about equally affected. We have no certain answer as to why there should be a gender factor for vocabulary recall but not for reading comprehension. Perhaps motivation is less able to affect performance in the case of reading comprehension than vocabulary retention. It must be remembered that the reading comprehension test was announced while the vocabulary test was not, with the intention of eliciting incidental vocabulary recall. Thus, there would not seem to be any incentive for the students to learn the vocabulary items while reading the passage. It could also be argued that the difficulty of the annotating task was detrimental to understanding, arguably a more difficult process than vocabulary recall, yet all students were given as much time as they needed to satisfy themselves that they could summarize the passage, before taking the reading comprehension test, and, of course, were not told of the vocabulary test.

The added exposure to the vocabulary items afforded to the students in the annotating group benefited all students with the exception of MAA's, perhaps due to lack of motivation. MHA and FHA students perhaps performed well in the annotating task due to their academic talents, the FAA students perhaps because of their motivation. Motivation, it seems, was a factor for MHA's and FAA's, but not necessarily for MAA's and FHA's.

In conclusion, we propose that motivation is the overriding factor to consider when interpreting the results of this study. The differences in motivation among men and women of various ability levels, and the differential effect of that motivation on individuals belonging to these demographic groupings are important in deciding the kinds of activities that would most benefit all students. Gender and Ability are learner variables that have not been widely explored in language learning and even less so taken together. In spite of the very real dangers of an over-emphasis on differences of sex and ability, we believe that the data in the study clearly demonstrate the importance of attempting to match activities with individual students to maximize learning. We have also demonstrated the complexity of the factors involved and have not reached any conclusions as to, for example, what effect the computer-based nature of the task had on performance or what role the types of tests used played. These are certainly important topics for further study.

Conclusions, Recommendations, and Limitations

The present study made an attempt to analyze in depth, for the first time, some quantitative parameters of vocabulary learning, reading comprehension, and attitude for academically talented and average-ability male and female students performing verbal annotations to words in a Spanish text. Without replication of the experiment, the exploratory character of this research makes it impossible to generalize the findings of the study beyond the conditions described in this article.

The study showed that students could successfully create verbal annotations for computer-based instructional materials in a foreign language text and thus upheld previous findings (Nikolova, 2002). All subjects accomplished their task in a reasonable time and created high-quality materials.

The main questions of the study were "Students of which gender learn vocabulary better and have a better reading comprehension when they create computer-based verbal annotations for a foreign language text rather than when they use a text with previously created annotations?" and "Were the performance and attitudes of different genders influenced by the ability level of the students?" The complexity of the questions was reflected in the complexity of the findings. The current study brought evidence supporting previous studies (Nikolova, 2002, Nikolova & Taylor, 2003) that high ability students per/brm significantly better than average-ability students when confronted with a more sophisticated task. The poor performance of the average-ability students was not equally distributed among the genders. It was primarily due to the MAA group, whereas the gap was much more narrow between FHA and FAA. We attributed the better performance of females in the average-ability group to a higher level of motivation registered within this population. A question remains as to why FHA students perform well in the annotating task even though their attitude toward this task was somewhat less positive than FAA. In addition, why was there a difference between the attitude of high ability males and high ability females toward the annotation task? A possible explanation offered in the discussion section proposes a strongly expressed orientation toward open-ended learner-involved tasks in the high ability male population.

It is our opinion that the results of the study should be interpreted cautiously because of the small raw score differences in some cases and the fact that the authoring activity was performed only once. A possible novelty effect may have been at play, which might have acted as a confounding variable. A future study may observe students working on longer texts in several sessions over a longer period of time, their motivation and their overall impressions from the treatment.

An important limitation of the present study is the issue of identifying the high ability students. We decided that a more practical, easy-to-use approach should be implemented, which would allow us a fast selection procedure. Thus, the present approach was adopted whereby students were identified based on former high school selection. We were hoping that high schools have used a multitude of selection criteria, which conform with a more multifaceted approach to intelligence. We are aware that different high schools may have had different criteria and sometimes even student/parent desire to belong to the "gifted" population may have been at play. Therefore, a future study may look at an improvement of the selection procedures.

In real-life conditions, the activities described in this paper may be modified in different ways. For example, students may be asked to find their own texts for annotation, including texts from the Internet. They may also be encouraged to annotate texts and to link them to other Internet sites, create their own web pages with annotated texts, and the like.

The design of the current study may be changed in various ways to accommodate real-life requirements of the classroom and to target other variables. Different types of texts may be used and the impact of the nature of the text on the performance of different student populations studied. Using different text types and multiple sessions will give the researchers the opportunity to collect more diverse data and attain more reliable results. A qualitative study in which the students are observed over a longer period of time, working in realistic classroom conditions on multimedia units (including creating picture and sound files) as pan of their curriculum, would most likely bring more insightful information about the real-life application of the idea of student authoring and the predilection of different ability and gender groups for one type of activity or another.

Finally, it is our hope that the present study will help orient researchers' attention toward student authoring activities in foreign languages, the role gender and ability differences play in them and the more thorough investigation of their many and complicated aspects.

Appendix

Reading passage: La leyenda del origen del cafe

Hay varias leyendas sobre el origen del cafe. La leyenda mas conocida y antigua del nacimiento del cafe se remonta a 1440. En Etiopia, un pastor llamado Kaldi noto en una ocasion que sus cabras exhibian cierta euforia despues de comer un fruto pequeno color granate. Intrigado, el pastor probo las hojas del arbusto y sintio los mismos efectos de euforia y <<claridad mental>>. Kaldi tomo frutos y ramas del arbusto y se los mando al abad de un convento cercano. El abad preparo una infusion, y la probo. El abad fue tan ingrata que arrojo los restos de la planta al fuego. Con el calor, los frutos desprendieron un aroma muy agradable. Asi nacio la idea de tostar los granos antes de preparar la infusion.

Note: Target vocabulary words are italicized.

Table 1
ANOVA Tables for the Factors of Gender, Method, and Ability on
Vocabulary Recall.

                              Immediate Recall

Source          df        SS         MS        F        p

Gender (A)#       1#      28.90#    28.90#    3.04#    .08#
Method (B)        1      207.71    207.71    21.87    <.0001
Ability(C)        1       77.11     77.11     8.12     .005
AxB#              1#       1.80#     1.80#    0.19#    .66#
AxC#              1#      73.18#    73.18#    7.70#    .006#
BxC               1      119.75    119.75    12.61     .0005
AxBxC#            1#       0.34#     0.34#    0.04#    .85#
Error           173    12290.89    170.71

                               Delayed Recall

Source          df        SS         MS        F        p

Gender (A)#       1#       0.24#     0.24#    0.10#    .75#
Method (B)        1        2.97      2.97     1.23     .27
Ability(C)        1       18.33     18.33     7.56     .007
AxB#              1#       7.53#     7.53#    3.10#    .08#
AxC#              1#      11.20#    11.20#    4.62#    .03#
BxC               1       41.22     41.22    17.00    <.0001
AxBxC#            1#       2.93#     2.93#    1.21#    .27#
Error             1    73419.50      2.42

Note: Boldface type denotes results of GLM with factor of Gender.

Note: Results of GLM with factor of Gender indicated with #.

Table 2
Vocabulary Recall

                 Immediate Vocabulary Recall

               Annotating         Nonannotating

Source       n    Mean    SD    n    Mean    SD

MHA          19   9.53   2.69   25   5.56   3.48
MAA          25   5.32   3.22   22   4.50   2.63

FHA          28   8.93   3.04   26   5.19   3.01
FAA          14   7.14   2.41   22   6.91   3.62

                  Delayed Vocabulary Recall

               Annotating         Nonannotating

Source       n    Mean    SD    n    Mean    SD

MHA          19   3.58   1.87   25   1.56   1.26
MAA          25   1.08   1.04   22   1.64   1.22

FHA          28   2.21   1.95   26   1.65   1.72
FAA          14   1.36   1.01   22   2.23   1.77

Table 3
T-tests for Equality of Means for Recall in Annotating Group.
Male High Ability (MHA) vs. Male Average Ability (MAA),
Female High Ability (FHA) vs. Female Average Ability (FAA)
(Equal Variances Assumed)

Level                  t     df      p         Mean      Std Error
                                            Difference   Difference
Immediate recall
MHA vs. MAA           4.59   42    <.0001      4.21         0.92
FHA vs. FAA           1.91   40     .06        1.79         0.93

Delayed recall
MHA vs. MAA           5.24   42    <.0001      2.39          .46
FHA vs. FAA           1.54   40     .13        0.86          .56

                           95% Confidence
                        Interval of the Mean

Level                   Lower   Upper    d

Immediate recall
MHA vs. MAA              2.36    6.05   1.40
FHA vs. FAA             -0.10    3.67    .22

Delayed recall
MHA vs. MAA              1.47    3.31   1.06
FHA vs. FAA             -0.27    1.98    .30

Note. d = effect size

Table 4
T-tests for Equality of Means for Recall for High Ability Students
By Group. Male Annotating (MA) vs. Male Nonannotating (MN),
Female Annotating (FA) vs. Female Nonannotating (FN) (Equal Variances
Assumed)

                                             Mean      Std Error
Level                 t     df     p      Difference   Difference

Immediate recall
MA vs. MN           -3.98   41    .0003     -3.90         0.98
FA vs. FN           -4.53   52   <.0001     -3.74         0.82

Delayed recall
MA vs. MN           -4.28   41    .0001     -2.02         0.47
FA vs. FN           -1.19   52    .24       -0.60         0.50

                      95% Confidence
                   Interval of the Mean

Level               Lower   Upper    d

Immediate recall
MA vs. MN           -5.88   -1.92   1.22
FA vs. FN           -5.39   -2.08   1.24

Delayed recall
MA vs. MN           -2.97   -1.07   1.30
FA vs. FN           -1.60    0.41    .32

Note. d = effect size

Table 5
T-tests for Equality of Means for Recall in Annotating Group.
Male High Ability (MHA) vs. Female High Ability (FHA),
Male Average Ability (MAA) vs. Female Average Ability (FAA)
(Equal Variances Assumed)

Level                t     df    p       Mean      Std Error
                                      Difference   Difference

Immediate recall
MHA vs. FHA        -0.69   45   .49     -0.60         0.86
MAA vs. FAA         1.84   37   .07      1.82         0.99

Delayed recall
MHA vs. FHA        -2.40   45   .02     -1.35         0.57
MAA vs. FAA         0.81   37   .42      0.28         0.34

                        95% Confidence
                     Interval of the Mean

Level                Lower   Upper    d

Immediate recall
MHA vs. FHA          -2.34    1.14   .20
MAA vs. FAA          -0.18    3.83   .62

Delayed recall
MHA vs. FHA          -2.51   -0.22   .71
MAA vs. FAA          -0.42    0.97   .22

Note. d = effect size

Table 6
T-tests for Equality of Means for Recall in Nonannotating Group.
Male average ability (MAA) vs. Female Average ability (FAA)
(Equal Variances Assumed)

Level                 t     df    p       Mean      Std Error
                                       Difference   Difference
Immediate recall
MAA vs. FAA          2.52   42   .02      2.41        0.955

Delayed recall
MAA vs. FAA          1.29   42   .20      0.59         0.46

                        95% Confidence
                     Interval of the Mean

Level                Lower   Upper    d

Immediate recall
MAA vs. FAA           0.48    4.34   .76

Delayed recall
MAA vs. FAA          -.033    1.52   .39

Note. d = effect size

Table 8
ANOVA Table for
Annotating Group Attitude

Source             df         SS         MS       F      P

Gender (A)#          1#     10.65#     10.65#   0.01#   .90#
Ability (B)          1     192.19     192.19    0.26    .61
AxB#                 1#   2041.05#   2041.05#   2.81#   .10#
Error              173    2538.94      14.68

Note: Boldface type denotes results of GLM with factor of Gender

Note: Results of GLM with factor of Gender indicated with #.

Table 9
Items with Significant Interactions

Questionnaire Item                 MHA     MHA     MAA     MAA
                                    M      SD       M      SD

Item 1: I find the method         59.48   18.01   39.66   27.22
used in this experiment
an interesting way of
learning Spanish.

Item 2: I find the method         53.93   19.00   37.88   27.59
used in this experiment
a fun way of learning
Spanish.

Item 6: I prefer to have          50.98   21.33   38.38   26.89
homework assignments
based on the method in
this experiment.

Item 10: I think that learning    38.93   25.19   55.52   27.09
how to use the software for
this experiment was a
waste of time.

Questionnaire Item                 FHA     FHA     FAA     FAA
                                    M      SD       M      SD

Item 1: I find the method         43.54   29.54   47.44   23.50
used in this experiment
an interesting way of
learning Spanish.

Item 2: I find the method         44.68   28.78   55.50   24.40
used in this experiment
a fun way of learning
Spanish.

Item 6: I prefer to have          43.73   26.31   59.97   28.13
homework assignments
based on the method in
this experiment.

Item 10: I think that learning    48.66   25.65   38.75   22.61
how to use the software for
this experiment was a
waste of time.

Questionnaire Item

Item 1: I find the method         F(1,88) = 4.64, p = .03, f = .23
used in this experiment
an interesting way of
learning Spanish.

Item 2: I find the method         F(1,88) = 5.88, p = .02, f = .25
used in this experiment
a fun way of learning
Spanish.

Item 6: I prefer to have          F(1,88) = 6.77, p = .01, f = .27
homework assignments
based on the method in
this experiment.

Item 10: I think that learning    F(1,88) = 5.90, p = .02, f = .25
how to use the software for
this experiment was a
waste of time.

Note: MHA=Male High Ability, MAA=Male Average Ability, FHA=Female
High Ability, FAA=Female Average Ability

Table 10
Items with Significant Main Effects for Gender

Questionnaire Item                                      Males    Males
                                                          M       SD
Item 1: I find the method used in this experiment
an interesting way of learning Spanish.                 44.98    29.77

Item 2: I find the method used in this experiment
a fun way of learning Spanish.                          44.46    28.76

Item 9: I would rather study a Spanish text in
paper format than use the method in this experiment.    43.75    27.66

Questionnaire Item                                     Females  Females
                                                          M       SD
Item 1: I find the method used in this experiment
an interesting way of learning Spanish.                 56.67    27.21

Item 2: I find the method used in this experiment
a fun way of learning Spanish.                          57.16    28.11

Item 9: I would rather study a Spanish text in
paper format than use the method in this experiment.    58.69    28.63

Questionnaire Item

Item 1: I find the method used in this experiment      F(1, 97) = 4.19,
an interesting way of learning Spanish.                p = .04, f = .20

Item 2: I find the method used in this experiment      F(1, 97) = 4.91.
a fun way of learning Spanish.                         p = .03, f = .22

Item 9: I would rather study a Spanish text in         F(1, 97) = 7.17,
paper format than use the method in this experiment.   p = .01, f = .27

REFERENCES

Adam, A., & Bruce, M. (1993). The expert system debate: A gender perspective. In J. O. E. Green. & D. Pain (Ed.), Gendered by design? Information technology and office systems. Washington, DC: Taylor and Francis.

Baumeister, R. F. (1988). Should we stop studying sex differences altogether? American Psyhologist, 43(12), 1092-1095.

Becket. D. F.. & Forsyth, R. A. (1990. April). Gender differences in academic achievement in grades 3 through 12: A longitudinal analysis. Paper presented at the annual meeting of the American Educational Research Association, Boston, MA.

Benbow, C. P. (1988). Sex differences in mathematical reasoning in intellectually talented preadolescents: Their nature, effects and possible causes. Behavioural and Brain Sciences, 11, 169-232.

Brantmeier, C. (2001). Second language reading research on passage content and gender: Challenges for the intermediate-level curriculum. Foreign Language Annals, 34(4), 325-333.

Bagel, K., & Buunk. B. P. (1996). Sex differences in foreign language comprehension: The role of interests and prior knowledge. The Modern Language Journal. 80(1), 15-31.

Burstalk C. (1975). French in the primary school: The British experiment. Canadian Modern Language Review. 31(5), 399-402.

Callahan, C. M.. & Reis, S. M. (1996). Gifted girls, remarkable women. In K. D. Arnold, K. D. Noble, & R. F. Subotnik (Eds.), Remarkable women: Perspectives on females' talent development (pp. 171-192). Cresskill. NJ: Hampton Press.

Caplan, P. J., &, Caplan, J. B. (1997). Do sex-related cognitive differences exist, and why do people seek them out? In P. J. Caplan, M. Crawford, & J. Richardson (Eds.), Gender differences in human cognition (pp. 52-80). New York: Oxford University Press.

Caplan, P., Crawford. M., Hyde. J., & Richardson, J. (1997). Gender differences in human cognition. New York: Oxford University Press.

Carrell, P. L., & Wise, T. E. (1998). The relationship between prior knowledge and topic interest in second language reading. Studies in Second Language Acquisition, 20(3), 285-309.

Carroll, J. B. (1975). The teaching of French as a foreign language in eight countries. New York: Wiley.

Clark, A., & Trafford. J. (1995). Boys into modern languages: An investigation of the discrepancy in attitudes and performance. Gender & Education, 7(3), 315-325.

Collis, B. (1985). Sex differences in secondary school students' attitude toward computers. The Computing Teacher, 12, 33-36.

Cross, D. (1983). Sex differences in achievement. System, 11, 159-162.

Dimitrov, D. M. (1999). Gender differences in science achievement: Differential effect of ability, response format, and strands of learning outcomes. School Science and Mathematics, 99(8), 445-448.

Feingold, A. (1988). Cognitive gender differences are disappearing. American Psychologist, 43, 95-103.

Fox, M. F., & Ferri, V. C. (1992). Women, men and their attributions for success in academe. Social Psychology Quarterly, 55(3), 257-271.

Grace, C. A. (2000). Gender differences: Vocabulary retention and access to translations for beginning students of CALL. The Modern Language Journal, 84(2). 214-224.

Halpern, D. F., & LaMay, M. L. (2000). The smarter sex: A critical review of sex differences in intelligence. Educational Psychology Review. 12(2), 229-246.

Hare-Mustin, R. T., & Maracek, J. (1994). Asking the right questions: Feminist psychology and sex differences. Feminism & Psychology, 4(4), 531-537.

Hogrebe, M. C. et al. (1984, April). Are there sex differences in reading achievement? An investigation with the high school and beyond data. Paper presented at the annual meeting of the American Educational Research Association, New Orleans, L.A.

Hollinger. C. L,, & Fleming, E. S. (1984). Internal barriers to the realization of potential: Correlates and interrelationships among gifted and talented female adolescents. Gifted Child Quarterly, 28(3), 135-139.

Hallway, W. (1994). Beyond sex differences: A project for feminist psychology. Feminism & Psychology, 4(4), 538-546.

Hunt, N., & Bohlin. R. (1993). Teacher education students' attitudes toward using computers. Journal of Research on Computing in Education, 25, 487-497.

Hyde, J., & Linn, M. (1988). Gender differences in verbal ability: A meta-analysis. Psychological Bulletin, 104, 53-69.

Lewis. E. G., & Massad, C. E. (1975). The teaching of English as a foreign language in ten countries. New York: Wiley.

Lindley, H. A., & Keithley, M. E. (1991). Gender expectations and student achievement. Roeper Review, 13(4). 213-215.

Lubinski, D., Benbow, C. P.. & Sanders, C. E, (1993). Reconceptualizing gender differences in achievement among the gifted. In K. A. Heller, F. J. Monks, & A. H. Passow (Eds.), International handbook of research and development of giftedness and talent (pp. 693-707). Oxford, England: Pergamon.

Luscombe, A., & Riley, T. L. (2001). An examination of self-concept in academically gifted adolescents: Do gender differences occur? Roeper Review, 24(1), 20-22.

Lynn, R., & Wilson. R. G. (1993). Sex differences in second language ability. School Psychology International, 14, 275-279.

Maccoby. E., & Jacklin, C. (1974). The psychology of sex differences. Stanford. CA: Stanford University Press.

McGlone, J, (1986). The neuropsychology of sex differences in human brain organization. In G. G. R. Tatter (Ed.), Advances in Clinical Neuropsychology (Vol. 3, pp. 1-30). New York: Plenum Press.

Meunier, L. (1995-1996). Human factors in a computer assisted foreign language environment: The effects of gender, personality and keyboard control. Calico Journal, 13(2 & 3). 47-72.

Mitra, A., LaFrance, B., & McCullough, S. (2001). Student differences in attitudes between women and men toward computerization. Journal of Educational Computing Research, 25(3), 227-244.

Murray, F. (1993). A separate reality: Science technology and masculinity. In J. O. E. Green. & D. Pain (Ed.), Gendered by design? Information technology and office systems. Washington, D. C.: Taylor and Francis.

Nikolova, O. R. (2000). Affective aspects of student authoring for foreign language learning. ReCALL, 12(2), 129-142.

Nikolova, O. R. (2002). Effects of Students' participation in authoring of multimedia materials on student acquisition of vocabulary. Language Learning and Technology, 6(1), 100-122.

Nikolova, O. R., & Taylor, G. (2003). The impact of a language learning task on instructional outcomes in two student populations: High ability and average ability students. Journal of Secondary Gifted Education, 14(4), 205-217.

Noble, K. D. (1987). The dilemma of the gifted woman. Psychology of Women Quarterly, 11(3), 367-378.

Noble, K. D., & Smyth. R. K. (1995), Keeping their talents alive: Young women's assessment of radical, post-secondary acceleration. Roeper Review, 18(1), 49-55.

Nolden, D., & Sedlacek. W. E. (1997), Gender differences in attitudes, skills, and behaviors among academically talented university freshmen. Roeper Review, 21(2). 106-109.

Nyikos, M. (1990). Sex related differences in adult language learning: Socialization and memory factors. The Modern Language Journal, 74(3), 273-287.

Olszewski-Kubilius. P. & Turner, C. (2002). Gender differences among elementary school aged gifted students in achievement, perceptions of ability, and subject preference. Journal for the Education of the Gifted, 25(3), 233-258.

Parish, T. S., & Necessary, J. R. (1996). An examination of cognitive dissonance and computer attitudes. Education, 116, 565-566.

Reis. S. M., & Dobyns, S. M. (1991). An annotated bibliography of non-fictional books and curricular materials to encourage gifted females. Roeper Review, 13(3), 129-134.

Sheridan, E. M., & Fizdale, B. (1981). Cultural factors and sex differences in reading and Mathematics. (ERIC Document Reproduction Service No, ED205936)

Stevens. J. (1990). Intermediate statistics: A modern approach. Hillsdale. NJ: Lawrence Erlbaum.

Swiatek, M. A, & Lupkowski-Shoplik, A. (2000). Gender differences in academic attitudes among gifted elementary school students. Journal for the Education of the Gifted, 23, 360-377.

Taisir, S. (1999). Attitudes toward computers of gifted students and their teachers. High Ability Studies, 10(1), 69-84.

Teh, G. P. L., & Fraser. B. J. (1995). Gender differences in achievement and attitudes among students using computer-assisted instruction. International Journal of Instructional Media, 22(2), 111-120.

Wolleat. P. L. 1979). School age girls. Counseling Psychologist, 8(1), 22-23.

Gregory Taylor is assistant professor of Spanish in the Department of Foreign Languages and Literatures at Southern Illinois University Carbondale. His research interests include second language acquisition, computer-assisted language learning, vocabulary acquisition, listening comprehension, and pragmatics. E-mail: [e-mail]

Ofelia Nikolova is currently assistant professor of French in the Department of Foreign Languages and Literatures at Southern Illinois University Carbondale. Her research interests are in second language acquisition, computer-assisted language learning, and foreign language instruction for high ability students. E-mail: [e-mail]