The reading behavior tracking and reading outcome measuring
In the e-book reading tasks, the reading behavior tracking technique developed in the IELS was used to simultaneously record the reading behavior on the reading process profile of every child (each record comprises reading duration, texts, and page number).
These profiles were collected by a central server, and then analyzed to identify any differences in the reading behavior between the genders throughout the reading processes. Upon completing each reading task, the children were asked to take a 10-minute retrieval test, as shown in
Figure 4b. Specifically, they were asked to freely recall some of the knowledge they had just learned by reading (Carver, 1990; Fraser, 2007), and the results were regarded as their individual reading outcomes. If a child did not completely read and understand the text, then the new knowledge may be poorly stored in the long-term memory, or even
1- lost from the working memory,
Based on the theory of cognitive memory (e.g. Ecker, Lewandowsky, Oberauer, & Chee, 2010; Nation & Cocksey, 2009; Swanson & Howell, 2001). In every retrieval test the children were asked to write 10 sentences about the scientific information contained in the e-book they had just read.
The reading outcomes were then measured according to the correctness of each sentence based on five separate levels (the rating criteria were as follows: four points for completely correct, three points for almost correct, two points for half correct, one point for less correct, and zero points for no answer or completely wrong).
2- Two raters assessed the sentences
based on these criteria. The Cronbach's alpha coefficient of their preliminary rating results was .68, showing a good fit for the reliability measure. The raters were then asked to review the results together and reach a consensus through discussion if any discrepancies were found.
3- The findings for the reading behaviors and outcomes
Gender differences in reading behaviors To confirm that the students of both genders had similar reading comprehension levels (Joshi & Aaron, 2000), the children took two comprehension tests before the reading experiment.
As shown in Table 5, no significant differences were found between genders, and thus the male and female groups had approximately the same comprehension level, and so this did not interfere with the measurement of the reading rate, which was adopted as an indicator of reading behavior in this work.
4- A total of 1,566 on-reading records were collected
from the participants, and then analyzed to identify the gender differences in their reading behaviors. To understand the fluctuation in reading rate throughout the reading process, this study further examined the reading rate in 5-minute periods, and then converted every reading rate into the corresponding reading status according to the categories listed in Table 2.
The independent sample t-test method was adopted to identify the significant differences in the reading rates between the genders. As shown in Table 6, overall, the boys’ reading rates were lower than those of the girls. A significant difference was found between the genders in the reading rates for the 30-minute reading task. This supports the finding of Burman,
5- Bitan and Booth (2008) that the girls tend to acquire
vocabulary faster than boys, perhaps reflecting different ways of linguistic processing between genders. According to Carver’s (1977) definition, Skimming is used to superficially process a large quantity of text. Duggan and Payne (2009) claimed that while Skimming can improve memorization of the important ideas in a text, but may pass over the details within it.
Our results show that the girls preferred Skimming, which is an important reading strategy essential for acquiring a good comprehension level of the gist of the text, as the reading behavior of spotting keywords shown in Table 2.
Reader and Payne (2007) pointed out that an adaptive reading strategy allows a reader to save more time to allocate to better reading, such as improving memorization of the important ideas by Skimming, but which reading strategy is the best relies on a complex set of issues (Lawless, Mills, & Brown, 2002; Salmerón & García, 2011), such as the texts, reading tasks, and reading skills.
6- Salmerón and García (2011)
claimed that children, starting at around 11 years old, possess the necessary reading skills to implement reading strategies. Nevertheless, how these strategies affect ebook reading comprehension is still an open question.
The boys’ reading rates, all in the Rauding status, were very stable throughout the reading process, as expected with the silent reading behavior. However, the girls’ reading rates were mainly in the Skimming status, which was found for none of the boys. By contrast, the boys read in the Rauding
status, at the pace of their cognitive speed (Carver, 1984; Fraser, 2007) throughout the reading process. The reading strategies that were not used by the boys may have been neglected due to their undeveloped reading skills or strategies, if the performance of the girls is regarded as ‘the norm’. This is a gender difference in reading behavior when the children were reading e-books.
Therefore, instruction in various reading skills and strategies seem to be needed for children, as noted by earlier researchers (e.g., Mayer, 1996; Park & Kim, 2011; Reutzel, Smith, & Fawson, 2005).
In practice, teachers can provide a proper instruction of reading skill and strategy towards the gender differences in reading of e-books for promoting them a better reading experience (Park & Kim, 2011; Reutzel et al., 2005; Sung, Chang, & Huang, 2008), and prevent such differences lead to the barrier for children when e-books are used for learning