Gender difference in web search perceptions and behavior: Does it vary by task performance?

Computers & Education 78 (2014) 174e184

Contents lists available at ScienceDirect

Computers & Education journal homepage: www.elsevier.com/locate/compedu

Gender difference in web search perceptions and behavior: Does it vary by task performance? Mingming Zhou* Faculty of Education, University of Macau, Av. Padre Tomas Pereira, Taipa, Macau

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 October 2013 Received in revised form 3 June 2014 Accepted 5 June 2014 Available online 14 June 2014

This study investigates Chinese students' gender differences in their actual use of the web for online information seeking. One hundred and seven Chinese university students responded to questionnaires regarding their perceptions about the use of the web for learning purposes. Afterwards, all the participants were asked to search online to answer two questions about bees' decision for hive location. As they searched, the online system logged participants' search activities during the search, including the type of activities during search, the frequency of each activity and the time spent on each activity. Participants were compared by gender in terms of their web search efficacy, web search anxiety, frequency counts of different web search activities, time spent on each search activity and search task performance. Web search efficacy levels varied by gender but not by performance levels. Anxiety did not vary by gender or performance levels. The interaction effect between gender and performance level was found in several search process variables: significant gender differences were only found in medium-performing students wherein males were engaged in more search activities than females, as seen in the larger number of searches, search queries, and times male students updated the search queries. One factor that could explain the significant gender differences in the medium-level group was their web search efficacy. The more confident medium-performing male students were in web search, the less need they perceived to access information to solve the task. This pattern was reversed for medium-performing females. The high- and low-performing males did not differ much from females in their search activities. It appeared that students' perceptions of their web search ability did not contribute much to their search activities in these two groups. Implications of the findings were also discussed. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Web search Gender Adult learners

1. Introduction In the 21st century, the proliferation of digital information technologies for knowledge construction requires individuals to be able to manage the overload of information adequately. This online information search process encompasses a set of cognitive skills including identifying the purpose of search, locating appropriate information sources, selecting and organizing relevant information, and synthesizing information from multiple sources into cogent, productive uses (Moore, 1995). As the Internet has become a popular platform for various purposes, such as online learning, electronic commerce and information search, individuals with different backgrounds use different approaches in their interactions with the web. Recent research finds that age, education, and gender are among the most important predictors of online information search behavior (Maghferat & Stock, 2010; Singer, Norbisrath, & Lewandowski, 2012; Steinerova & Susol, 2007; Weber & Jaimes, 2011). Therefore, these human factors must be taken into account in our explanation of different levels of performance by using the web to solve problems. A growing body of studies has been conducted to examine gender differences in information seeking on the web. However, previous research has shown that students vary widely in their ability to find and retrieve information in loosely structured information environments (e.g., Brand-Gruwel, Wopereis, & Vermetten, 2005; Tabatabai & Shore, 2005), such as the Internet. A recent review by Chen and Macredie (2010) suggests that major differences between males' and females' online information seeking lie in their navigation patterns,

* Tel.: þ853 83978704; fax: þ853 28831695. E-mail address: [email protected]. http://dx.doi.org/10.1016/j.compedu.2014.06.005 0360-1315/© 2014 Elsevier Ltd. All rights reserved.

M. Zhou / Computers & Education 78 (2014) 174e184

175

attitudes and perceptions. In general, literature has suggested that males report lower levels of computer anxiety than females (Karavidas, Lim, & Katsikas, 2004); females lag behind males in the degree to which they are experienced with and motivated by technology (Leong & Hawamdeh, 1999; Light, Littleton, Bale, Joiner, & Messer, 2000; Schumacher & Morahan-Martin, 2001). In addition, it also seems that males were more successful in finding required information (Kim, Lehto, & Morrison, 2007). However, Agosto (2004) argued that gender as a sole determining factor was too simplistic a way to look at information-seeking behavior. Studies consistently reported that males were more interested and involved with technology than females, yet this is often no longer the case (North & Noyes, 2002). Other factors could also co-influence individuals' search behavior. As several researchers have noted that the effect of gender on learners' self-perceptions and learning processes could vary by different performance levels (e.g., Khalid & Hasan, 2011; Sheu, Wang, & Hsu, 2013), gender differences in web search efficacy and behavior by search task performance was examined in the current study. In the following sections of this paper, a brief review of empirical studies on gender differences in individuals' navigation patterns when searching for information online as well as their perceptions about web search is first presented, followed by the purpose of the current study.

2. Gender difference in perceptions about the use of web Web efficacy has been the subject of many research studies in seeking online information. It has been defined by Savolainen (2002) as a person's perception of “his or her ability to organize and execute action, such as finding information on the Web” (p. 211). This perception is supposed to be clearly reflected in one's web search behavior. Higher web search efficacy may help individuals try better web search strategies and facilitate higher-order metacognitive skills, such as information selection and evaluation. This, in turn, facilitates individual performance in a web-based environment (Tsai & Tsai, 2003). Research has been done on the influence of gender differences on individuals' attitudes/perceptions toward the use of web to seek information. Results are mixed. On one hand, the majority of literature suggests that females reported more computer anxiety, lower levels of competence and higher levels of discomfort than their male counterparts when using the web to search for information (e.g., Hu, Zhang, Dai, & Zhang, 2012; Jackson, Ervin, Gardner, & Schmitt, 2001; Koohang, 2004; Li & Kirkup, 2007; Liaw, 2002; Peng, Tsai, & Wu, 2006; Schumacher & Morahan-Martin, 2001). On the other hand, some studies indicate that females showed stronger positive attitudes than males towards using the web as major resources to gather information (e.g., Kim et al., 2007). Also, several researchers found no significant gender differences in their perceptions toward using the web for trip planning information search (e.g., Koohang & Durante, 2003). Tsai and Tsai (2010) went one step further by examining high school students' perceptions about different uses of the Internet. They found that boys and girls perceived themselves about the same confident with regards to exploring and navigating the web, but girls held significantly more confidence than boys regarding using the web as an online communication tool.

3. Gender difference in navigation patterns It has been repeatedly evidenced that problem-solving strategies directly affect problem-solving success (e.g., Tsai, Hou, Lai, Liu, & Yang, 2012; Tu, Shih, & Tsai, 2008). When searching information online for a given search task, search patterns determined search success to a large extent. Tabatabai and Shore (2005) noted that experts' performance was differentiated by the cognitive and metacognitive strategies they employed. By evaluating and monitoring the search process, expert searchers had higher chances of success. In contrast, poor searchers relied more on trial-and-error with less patience. Impatience led them to navigate more, to click more, and to execute before spending sufficient time planning or evaluating. However, there was less evidence available regarding whether males searched for information online in a different way from females, in comparison to the effort to study gender differences in individuals' attitudes and perceptions about the web use. In Tsai and Tsai's (2010) study, boys were found to be more exploration-oriented who navigated or searched information on the web mostly, whereas the girls were more communication-oriented who mainly communicated via the Internet. The authors attributed the
ne
cal (2011) found that observed different patterns of web use to users' different purposes (or goals) of web use. Arcand, Nantel, and Se women spent more time per page but viewed fewer pages, whereas men spent less time per page but accessed more pages throughout the task. This could be explained by the selectivity model (Meyers-Levy, 1989) whereby females are comprehensive processors who tended to assimilate all available information and elaborated more on it, whereas males were selective information processors who did not generally engage in extensive processing of all available information. Instead, they employed various heuristic devices that served as surrogates for more detailed processing. Large, Beheshti, and Rahman (2002) examined sixth graders' web searching process to complete a class assignment. They found that boys conducted more search activities including formulating more queries, clicking on more links per minute, and following up on more hits. On the other hand, girls spent more time reading documents. The authors speculated this was because girls were generally less skilled or practised in navigating the web who had to spend more time in processing information. Roy and Chi (2003) had similar findings to Large et al. (2002) that middle school boys tended to employ a different search pattern from girls when searching for answers to an academic question and that this difference was related to the search performance. Specifically, boys navigated the Internet in a non-linear way and girls tended to browse entire linked documents and to follow a linear navigation approach. Liu and Huang (2008) also found in university students that male readers preferred non-linear reading than female readers, who tended to spend more time on browsing, scanning and non-linear reading (such as jumps) in their report. The authors again borrowed the selectivity model to explain the findings that men naturally tended to be more selective in web searching. Women, however, tended to make a greater effort and employ a more conscientious approach (Hupfer & Detlor, 2006). With eye tracking data, Lorigo et al.'s (2006) observed that girls more often returned to previously visited abstracts when they searched information while boy's navigation paths were more likely to be strictly linear. These inconsistent findings about whether males search for information in a linear way as compared to females invite further investigation. Moreover, the aforementioned studies did not relate the different web search behavior to any measure of search success. Hence, it is unknown if such differences are relevant to search performance. This will be examined in the current study.

176

M. Zhou / Computers & Education 78 (2014) 174e184

Fig. 1. User interface of the web search platform.

4. Present study There are several reasons for the inconsistent results as to how gender differences influence students' perceptions towards web search and their navigation patterns when searching online. One reason is the specificity of the purpose of using the Internet. In some studies (e.g., Tsai & Tsai, 2010), the purpose was quite general, while in others a specific search task was used (e.g., Roy & Chi, 2003). Thus, the information collected about the users as well as the web use was contextualized differently. Another reason is the use of self-report to obtain such information as to how the individuals navigate the website to find the information they need. Research has repeatedly shown that one's perception about how he or she completes a task could differ from the way the task is actually carried out (e.g., Winne & Jamieson-Noel, 2002). When students are asked to provide cumulative and retrospective accounts about how they perform a task, their accounts may be based partially on biased information arising from incomplete and reconstructed memories plus subjective and implicit theories of the mental processes involved (Ericsson & Simon, 1984; White, 1989). Hence complete reliance on participants' self-reports might obscure or even bias the results. Last, gender might have a differential effect on web perceptions and use patterns for individuals at different performance levels. Although no studies have been conducted to test this assumption directly, several studies have pointed to this conclusion. For instance, Weinburgh's (1995) meta-analysis of the literature in student attitudes toward science showed that the high-performance girls reported a slightly more positive attitude toward science than high-performance boys, yet for the low-performance and general performance students, boys showed more positive attitudes than girls. In a recent study, Preckel, Goetz, Pekrun, and Kleine (2008) found that gender differences in sixth-graders' mathematics-related self-concept, interest, and motivation were larger for gifted than for average-level students, in favor of boys. Khalid and Hasan (2011) observed female high achieving undergraduates experienced more test anxiety as compared to male high achievers, whereas male low achievers experienced more test anxiety than female low achievers. These results seemed to support the assumption that gender differences in students' self-concept, attitude or anxiety varied by the performance level of students. However, Dai (2001, study 2) found high school boys tended to report higher scores than girls on the measure of math self-concept, regardless of their performance level. With regards to strategy use, Sheu et al. (2013) found that female EFL learners in Taiwan generally adopted more use of memory and affective strategies than males for language learning, but the gender difference disappeared in the high performing group. It is noteworthy that ability was used in the above studies, measured by individual performance on achievement tests in a particular domain. Such a way of dividing students into high or low ability groups created concerns when we attempted to translate it into web search contexts (e.g., general search ability). First, web search patterns are subject to the demand of a specific task, and web search task performance does not only depend on student search behavior but also on the task itself. The experts on one search task could easily be novices in another. Hence, task performance was adopted in this study to group participants, which was more accurate than students' general search ability level. As one's ability in a given area is also related to one's performance of tasks in that area (Werner & DeSimone, 2012), it is reasonable to assume an interaction effect between gender and task performance on a specific web search task. Unfortunately, this issue has yet been systematically examined.

M. Zhou / Computers & Education 78 (2014) 174e184

177

As most of the literature on the gender effect in information search behavior was contextualized in non-academic tasks such as online shopping (Laroche, Saad, Cleveland, & Browne, 2000), searching advertising (Jansen & Solomon, 2010) and tourism (Kim et al., 2007; Xie, Bao, & Morais, 2006), in the current study, I examined Chinese students' efficacy about using the web to seek information as well as their search processes and outcomes, to explore what, if any, gender differences would be found. Furthermore, I aimed to explore the extent to which differences in these variables are specific to students with a certain performance level. Therefore, gender differences in online information search were compared among groups with different performance levels. This study was designed to avoid the two drawbacks of previous research. One problem of prior studies on gender differences in using the web was that samples were not differentiated based on their search performance levels. Therefore, it is difficult to generalize and consolidate findings across studies. Second, most of the studies cited above only used self-reported data as the only measure of the actual use of the Internet. To address this issue, self-reports of the searching process was not used as they only reflected students' perceptions about the process. Instead, computer logs were employed in the current study to assess whether and how males solved the task differently from females. Specifically, the following research questions were examined in the current research: 1. Was there any gender difference in students' perceptions towards web search? 2. Was there any gender difference in web search behavior? 3. Did the gender difference in the above variables vary by students' online search task performance?

5. Method 5.1. Participants One hundred and seven Chinese undergraduate students from the same university in south China volunteered to participate in this study (47.7% male; mean age ¼ 21.2 years old). 77.6% of them were studying in their second year in computer or science-related majors, who had extensive experience with web usage. On average, they spent 33.7 h per week on the Internet, mainly for emailing, online chatting and academic studying. 5.2. User interface of the web search platform A web-browser add-on application was developed for Firefox web-browser that detects and records significant events during a session of web search (Zhou, Xu, Su, & Liu, 2011). When a user signs in with an assigned ID, the application is launched and runs without obtrusive interference with user web search. The user will use the search engine as usual after logging in. The application also allows users to highlight information they deem relevant to solve the search problem (see Fig. 1). At the backend, the application logs participants' search activities while they search online, including the type of activities during search, the frequency of each activity and the time spent on each activity. 5.3. Instruments 5.3.1. Web perceptions Six items that measured students' general perceptions about their Internet use for information search were taken from The Inventory of Perceptions of Web-Based Information Seeking (Ford & Miller, 1996). It has been used in specific contexts such as search tasks on a given topic (e.g., Ford & Miller, 1996; Ford, Miller, & Moss, 2005). A sample item was “I usually manage to keep ‘on target’ and avoid too much irrelevant material when using the Internet”. An acceptable model fit was observed based on the standards by Browne and Cudeck (1993): c2 ¼ 13.46, c2/df ¼ 1.50, RMSEA ¼ .070, SRMR ¼ .054, CFI ¼ .95, TLI ¼ .91, IFI ¼ .95. Internal consistency alpha value was .69. Another five items on anxiety were taken from Joiner et al.'s (2007) online search anxiety scale. A sample item was “When I use the Internet, I feel as though I'm not as ‘in control’ as I would like”. A good model fit was observed after removing one item due to its poor factor loading: c2 ¼ .05, c2/df ¼ .02, RMSEA ¼ .000, SRMR ¼ .004, CFI ¼ 1.00, TLI ¼ 1.00, IFI ¼ 1.00. Internal consistency alpha value was .75. Participants responded by agreeing or disagreeing on a 5-point Likert scale. 5.3.2. Search logs The activities identified in this study were based on the log files as well as theoretical accounts of the IPS skills (Brand-Gruwel, Wopereis, & Walraven, 2009). They were searching, reading webpages, adapting search queries, highlighting, and answering questions. As an example, when an individual submitted a search query and clicked on the “search” button in the webpage, this behavior was recognized as a “searching” action. To ensure the validity of the coding procedure (Zhou, 2013b), these activities captured the core skills needed for online information seeking in Brand-Gruwel et al.'s (2009) model: “search information”, “scan information”, “process information”, and “organize and present information”. Inter-rater reliability was determined by randomly selecting ten student logs that were coded independently by the author and the research assistant. The agreement rate was 96.5%, indicating that the coding was highly reliable. 5.3.3. Search task performance Two search tasks were designed for participants to search online in the current study: 1) How do bees choose where to build their new homes? and 2) What do you think are the implications for human life?. Participants' performance on these tasks was not related to their academic results in university, but was used as a context to conduct web search to seek information on a topic they have limited knowledge about. Task performance was measured by Search Performance Index (SPI). It is a measure of the performance in the search tasks, which considers both the accuracy of the answers (effectiveness) and the total time of task completion (efficiency) (see Zhou, 2013a). The SPI measure was calculated with the following formula: SPI ¼ Accuracy/Time  60. The accuracy of each answer was scored between 0 and 5, with a sum of 10 points for the whole task. Each participant's answer was evaluated by two independent researchers. A scoring rubric was

178

M. Zhou / Computers & Education 78 (2014) 174e184

Fig. 2. Procedure of the study.

developed which reflected the degree to which participants presented the answers accurately and logically (see Zhou, 2013b for details). For the first close-ended question, an answer awarded 4e5 points should present correct, elaborative, and logic information, with zero or minimal portions of irrelevant information. An answer was awarded 2e3 points if it included accurate information with small portions of irrelevant information, or information that was insufficiently detailed. An answer was awarded 0e1 point if the answer was off-topic (e.g., the best conditions to build a hive, what makes a good beehive, or categories of beehives) or only a minimal portion of the information was pertinent. The answer to the second question was rated on its soundness, richness and organization. An answer awarded 4e5 points should be logically organized and the claims were supported with rich examples and details. An answer was awarded 2e3 points if it was generally logically organized and the claims were supported with relevant facts. An answer was awarded 0e1 point if it was poorly organized and the claims were supported with few or no facts. Each participant's answer was evaluated by two independent researchers. The inter-rater reliability (Cohen's kappa) was .90. The final score was the average of the two. Any difference more than three points were resolved through discussions. As an example, an answer of “Bees use something called swarm intelligence to decide their new home location” to the first question was agreed between the two raters to receive a score of 1 for the lack of facts to support the point, where an answer of “An individual scout first promotes a location she has found through the waggle dance to indicate direction and distance to others in the cluster. The more excited she is about her finding, the more excitedly she dances. If she can convince other scouts to check out the location she found, they may take off, check out the proposed site and promote the site further upon their return. Usually several different sites are promoted by different scouts. After several hours for examination of all the proposed sites, a group decision will be made.” to the same question received a score of 5 from rater A for its logical and elaborative explanation but a score of 4 from rater B, as she believed the explanation was incomplete by missing the exact information on how the cluster came to a final decision (agreed by 100% or above 80% of the cluster?). In this case, an average score of was this participant's final accuracy score for this answer. The total time of task completion was based on the computer logs, starting from the time participants logged in until the time they logged off.

M. Zhou / Computers & Education 78 (2014) 174e184

179

5.4. Procedure The study was conducted in a computer laboratory at the university. After arriving at the laboratory, the researcher informed the participants of the nature and the procedure of the study. Each participant signed a consent form to agree that the data collected in any form would be used for research purposes only. The participants were asked “What do you know about how bees choose where to live?”. Their responses were assessed by whether or not they could answer this question correctly. The accuracy of their responses served as an indicator of their prior knowledge level about this topic. Results showed that 92.5% of the participants had zero knowledge about the topic. They then responded to the questionnaire on hardcopy. After that, the participants received the search tasks and were told that the purpose was “to study their normal information search strategies”, without knowing that their search behavior was tracked at the same time they searched (also see Zhou & Winne, 2012; Zhou, 2013b). The default search engine was set to be Google, although the participants were allowed to use other search engines. As the participants searched information online, their search process was logged in real time. The time limit for the whole session was limited to one hour. The whole procedure is illustrated in Fig. 2. 6. Results In this section, the results of several statistical tests were presented. The analyses focused upon the calculation of descriptive statistics and analyses of variances statistics to investigate students' experiences in their efficacy of using the Internet and the actual search activities based on their task performance levels. Initially, a series of descriptive statistics were conducted to ensure no violation of the assumptions of analysis of variance relating to the normality of the dependent measures and the absence of extreme outliers. The examination of the skewness and kurtosis of all the variables showed that two variables (i.e., the number of highlights during searching, and the time spent on reading webpages) were not normally distributed. The boxplot identified four subjects as outliers for these variables. The removal of these four cases led to acceptable values of skewness and kurtosis for these variables (ranging from 0.83 to 1.33) as well as a sample size of 103. On average, the participants scored 11.68 (SD ¼ 16.54) in the two tasks. They were divided into three groups (high, medium, and low) according to their SPI scores. Accordingly, high level students were those in the upper one-third of SPI scores, and low level students were those in the lower one-third, while the rest of students were considered average level students. Two subjects were identified as outliers due to their extremely high SPI scores (above 60.0, with the rest below 37.0). This led to the removal of these two cases. Combined with the four outliers aforementioned, the final sample size in this study was 101. The mean and standard deviation of each variable are presented in Table 1. The number of students, percentages, and SPI score range for each performance level group are shown in Table 2. 6.1. Gender differences in web perceptions by search task performance A 2 (male or female)  3 (low, medium or high performance level) multivariate analysis of variance (MANOVA) was conducted with web search efficacy and anxiety as dependent measures. There was a significant main effect of gender for web search efficacy, F(1,99) ¼ 6.87, p ¼ .01, h2 ¼ .067 (see Table 3). Post hoc tests revealed that male students in general were more efficacious towards using the Internet to search for information than females. No significant main effect of performance levels was detected for web search efficacy. No significant main effects or interaction effect of gender and performance level were found in anxiety, although female students reported to be more anxious when seeking information online than their male counterparts, and low-performing students reported to be more anxious than medium or high-performing students. However, these differences did not achieve significance during post hoc testing. 6.2. Gender differences in search process variables by search task performance With regards to the search behavior, 2 (male or female)  3 (low, medium or high performance level) MANOVA results showed a significant main effect of gender for total time spent on the task, time spent on constructing the answers, the number of searches, the number of search queries, and the number of times participants updated the search queries, with males scoring higher in all the above variables except for the time spent on answering the questions (see Table 3). There was also a significant main effect of performance level for total time spent on the task, time spent on searching, time spent on reading webpages, the number of searches, the number of search queries, the number of times participants updated the search queries, and the number of webpages viewed, with the low-performing group scoring significantly lower than the other two groups. A significant interaction effect between gender and performance level was observed for the number of search queries and the number of times participants updated the search queries. A marginally significant interaction effect was also found in the number of searches, Graphic displays of the interactions of gender and performance level for these three variables are shown in Figs. 3e5. Post hoc tests (see Table 4) consistently revealed that significant gender difference stemmed from the medium-performing group: medium-performing males were

Table 1 Means and Standard Deviations (SD) for web perceptions and search process variables for the whole sample (N ¼ 101).

Mean SD

Mean SD

Web search efficacy

Anxiety

Total time (min)

Time on searching (min)

Time on reading webpages (min)

Time on answering (min)

3.15 0.58

2.25 0.66

31.32 10.17

5.46 3.03

9.58 5.22

3.08 3.77

Number of searches

Number of search queries

Number of times of adapting search queries

Number of webpages viewed

Number of highlights

9.40 6.56

8.09 5.45

7.78 6.08

13.74 9.63

6.95 4.57

180

M. Zhou / Computers & Education 78 (2014) 174e184

Table 2 Number of students, percentages and cutoff scores of students in their search performance index (SPI). Quartile

Mean

Percentage

SPI score range

Gender ratio (Male:Female)

Low Medium High

33 34 34

32.6% 33.7% 33.7%

0e3.5 3.6e14.0 14.1e37.0

19:14 15:19 16:18

engaged in a higher frequency of search activities as shown by higher numbers of searches, search queries as well as times search queries were adapted. The high- and low-performing males did not differ much from females in their search activities. 6.3. Gender differences in search process variables by web search efficacy To further investigate the effect of gender and web search efficacy on search processes, participants in each group were labeled by their web search efficacy level. A score of 3.00 was found to be the median efficacy score across three groups. Hence, students with efficacy scores equal to or above 3.00 were classified as high web search efficacy searchers in that group; those with a score below 3.00 were low efficacy searchers. A 2 (high vs low search efficacy)  2 (gender) MANOVA was then conducted within each search performance group with search behavior variables as dependent measures. In the average-performing group, there was a significant main effect of gender in the number of webpages viewed: F(1,30) ¼ 9.40, p < .01, h2 ¼ .238; the number of searches: F(1,30) ¼ 8.89, p < 01, h2 ¼ .229; the number of search queries: F(1,30) ¼ 11.15, p < .01, h2 ¼ .271; and the number of times of updating search queries: F(1,30) ¼ 9.25, p < .01, h2 ¼ .236, all favoring males. Despite no significant main effect of web search efficacy in any search process variables, a significant interaction effect between gender and web search efficacy was observed in the number of webpages viewed: F(1,30) ¼ 4.24, p < .05, h2 ¼ .124. Average-performing males with higher web search efficacy tended to view much fewer webpages during web search (M ¼ 16.00, SD ¼ 10.65) than those with lower efficacy did (M ¼ 28.75, SD ¼ 10.05), whereas average-performing females with higher efficacy tended to view slightly more webpages (M ¼ 12.71, SD ¼ 7.78) than those with lower efficacy did (M ¼ 12.00, SD ¼ 6.47). For the high-performing group, the only significant effect was found on gender's effect on time in answering questions: F(1,30) ¼ 11.72, p < 01, h2 ¼ .281. For the low-performing group, no main or interaction effect was found. 7. Discussion The purpose of this study was to investigate gender differences in university students' web search perceptions and behavior and to examine such differences by varying search task performance levels. First, males in the medium-performing group held a stronger and more positive belief about their search ability than females. This result was consistent with prior literature that women usually had less confidence in finding information on the Internet than men (e.g., Comber, Colley, Hargreaves, & Dorn, 1997; Kirkpatrick & Cuban, 1998; McDonald & Spencer, 2000; Torkzadeh & Koufteros, 1994). Yet the non-significant gender differences in web efficacy in the other two groups seemed to support Sam, Othman, and Nordin's (2005) observation of no significant gender differences in computer self-efficacy. The observed gender difference in web search efficacy in the average-performing group may be due to the sociocultural stereotypes. Gender socialization theory posits that gender differences in academic environments stem from social stereotypes (Konrad, Yang, Goldberg, & Sullivan, 2005). Such stereotypes have a big impact on males' and females' perceptions of themselves and on what they are able to achieve (Martin, 2010), and can bias individuals' expectations for their own competence in the situation independently of their underlying abilities (Ridgeway & Correll, 2004). It is possible that the increased awareness of gender stereotypes in the average-performing group results in the fact that genderrole stereotypes and gender-linked development of self-perceptions become more pronounced in this group of students. Also, similar to Rosen, Sears, and Weil's (1987) finding, females and males in this study did not differ much in their anxiety levels. Neither were there any significant differences across performance levels in participants' anxiety scores. It is probably because that the current research was conducted in a laboratory setting (vs a traditional classroom setting) and the performance scores were not related to their actual academic results in university. This could have reduced the anxiety levels to some extent (as shown in their low ratings in the anxiety scale).

Table 3 Results of the effects of gender and performance level on web perceptions and search process variables. Web search efficacy

Gender Performance level Gender  Performance level

Anxiety

Time on searching (min)

Time on reading webpages (min)

Time on answering (min)

F

p

h2

F

p

h2

F

p

h2

F

p

h2

F

p

h2

F

p

h2

6.87 .45 .41

.010 ns ns

.067 .009 .008

2.37 .17 .02

ns ns ns

.024 .003 .000

4.16 11.40 .17

.044 .000 ns

.042 .194 .004

.21 6.97 .92

ns .001 ns

.002 .128 .019

.04 13.27 .84

ns .000 ns

.000 .218 .017

9.37 .34 .99

.003 ns ns

.090 .007 .020

Number of searches

Gender Performance level Gender  Performance level

Total time (min)

Number of search queries

Number of times of adapting search queries

Number of webpages viewed

Number of highlights

F

p

h2

F

p

h2

F

p

h2

F

p

h2

F

p

h2

4.83 7.52 2.93

.030 .001 .058

.048 .137 .058

5.89 9.52 4.24

.017 .000 .017

.058 .167 .082

5.33 8.55 3.31

.023 .000 .041

.053 .153 .065

2.20 8.35 1.61

ns .000 ns

.023 .150 .033

.51 1.16 2.24

ns ns ns

.005 .024 .045

M. Zhou / Computers & Education 78 (2014) 174e184

181

Fig. 3. Gender differences in the number of searches by different performance groups.

There was a more robust difference by gender in students' search behavior. In line with prior literature (e.g., Large et al., 2002; Liu & Huang, 2008), males were engaged in more search activities than their female counterparts by conducting more searches, experimenting with different search queries, and visiting more webpages. But this was only observed in the average-performing group. For students who were very capable or incapable of searching online, the gender difference in search behavior was rather obscure. One factor that could explain the significant gender differences in the medium-level group was their web search efficacy. The significant interaction of gender and web search efficacy in predicting the number of webpage visits in this group demonstrated that web search efficacy contributed to certain search behavior and this contribution varied by gender. For males in this group, it appeared that the more confidence they had in web search, the less need they perceived to access information to solve the task, because of their strong and positive belief that they could perform well by any means. This pattern was reversed for females, however. Highly efficacious female searchers tended to click more results for information, possibly because they considered it to be an indicator of expert search behavior with a large coverage of information on the given topic. This gendered pattern deserves further investigation. A closer examination of the average-performing group's search performance revealed that males' significantly higher frequency of web search activities and higher level of web search efficacy did not lead to significantly better search performance than females' in this group. Further, females' lower efficacy did not result in significantly poorer performance; rather, their lower frequency of search activities (i.e., a smaller number of searches, search queries and search query reformulations) could to some extent compensate for the poor performance they could possibly have had. According to Tu et al.'s (2008) finding that fewer search keywords led to fewer visited web pages but better search task performance, average-performing females' fewer searches or queries could possibly show better cognitive or metacognitive strategies in performing web search. If the females in this group held the same level of web search efficacy as males did, along with their better search strategies, the females could have achieved better search performance than males'. On the other hand, males in the averageperforming group could have encountered disorientation, which drove them to try more searches or queries (Bilal, 2001). Similar reasons could also be applied to explain why high-performing group's (especially males') number of searches and search queries €lscher and Strube (2000) argued that web novices more often reiterated their was far lower than low- and average-performing groups'. Ho search queries, which led to a higher number of searches. But this behavior did not result in much search success. On the other hand, web experts were significantly more likely to choose a relevant webpage for closer inspection than web novices, which led to a smaller number of

Fig. 4. Gender differences in the number of search queries by different performance groups.

182

M. Zhou / Computers & Education 78 (2014) 174e184

Fig. 5. Gender differences in the number of search query updates by different performance groups.

webpages to view. Although novice searchers visited more webpages, most of them appeared to be irrelevant. Their higher proportion of search query reformulations also showed the unsuccessfulness of previous searches. Hence, the smaller number of searches, search queries and webpage visits in the high-performing group could reflect better cognitive strategies than the other two groups. As to the high- and low-performing groups, students' perceptions of their web search ability did not contribute much to their search activities. As no significant gender difference was found in students' web search efficacy, no significant difference was found in their search behavior and task performance. One striking finding was in the significantly longer time spent on solving the task by high-performing females relative to their male counterparts. Tsai, Liang, Hou, and Tsai (2012) found that university female students might be more readily influenced by contexts than male students during online information seeking. I speculated that this could be particularly true for high-performing students. If a high-performing female student perceived the task to be highly complex, the task could increase the cognitive load which resulted in a higher level of effort than usual. This might not apply to average- or low-performing students, as their search capability might not allow such a level of effort expenditure. Hence, the nature of search tasks could be another factor to consider in web search. 8. Conclusion The Internet has been historically characterized as masculine (Li & Kirkup, 2007; Silva, 2000). The current findings reflect that gender stereotyping is diminishing (Schumacher & Morahan-Martin, 2001) in the new generation in China, especially in students who are either very skilled or unskilled in web search. Rather, the gender gap was most salient in average-performing students. This has particular implications for designing training programs for undergraduate students' Internet skills. On one hand, the programs could consider males and females in similar ways when designing online search practice, as females' perceptions of web use and actual navigation of the web have become similar to males'. On the other hand, educators still need to provide and design specific and effective Internet skill instruction for each gender, according to their learning characteristics and needs (Cheng, Liu, Chen, Shih, & Chang, 2012). This was especially necessary for average level students. Female students at this level of performance needs more encouragement and a stronger sense of success such that they would be more positive towards online search and more successful in online information seeking. Table 4 Means, Standard Deviations for the comparisons of web perceptions and search process by gender in different performance level groups. Whole sample

Web search efficacy Anxiety Total time (min) Time on searching (min) Time on reading webpages (min) Time on answering (min) Number of searches Number of search queries Number of times of adapting search queries Number of webpages viewed Number of highlights Search task performance a b c

Male (N ¼ 50)

Female (N ¼ 51)

Mean

SD

Low-performing group

Medium-performing group

High-performing group

Male (N ¼ 19)

Female (N ¼ 14)

Male (N ¼ 15)

Female (N ¼ 19)

Male (N ¼ 16)

Mean

SD

Mean

Mean

SD

Mean

Mean

SD

Mean

SD

Mean

SD

3.30a 2.15 29.56 5.44 9.83 7.99b 10.82a 9.36a 9.12a

.59 .59 8.22 2.92 4.28 5.19 7.70 6.48 7.26

3.00a 2.35 32.00 5.48 9.34 11.63b 8.06a 6.89a 6.51a

.54 .71 12.56 3.16 6.01 7.60 4.97 3.97 4.43

3.25 2.16 32.86 5.79 9.71 8.61 10.74 9.16 8.95

.52 .49 6.63 3.26 4.38 4.32 6.31 4.98 5.84

3.02 2.39 35.62 7.14 11.06 10.09 10.71 9.21 8.86

.47 .69 13.78 4.24 6.22 10.59 5.88 4.56 5.01

3.32a 2.17 31.88 6.20 12.52 8.37 15.07a 13.47b 13.40b

.63 .70 8.76 3.36 3.62 7.06 9.74 8.05 9.09

2.88a 2.38 37.13 6.16 12.21 12.91 8.32a 7.11b 6.84b

.51 .72 11.86 2.35 6.17 6.52 4.70 3.70 4.32

3.34 2.11 23.46 4.31 7.44 6.89c 6.94 5.75 5.31

.67 .63 6.12 1.58 3.35 4.14 4.81 3.99 4.48

3.11 2.28 26.73 3.80 5.84 12.74c 5.72 4.64 4.28

.62 .76 5.96 2.03 2.98 5.10 3.68 2.96 3.25

15.46 7.32 9.49

10.82 4.50 9.05

12.11 6.60 10.19

8.13 4.65 7.91

17.37 8.11 .92

10.97 5.13 1.15

17.50 5.79 1.54

9.61 3.68 1.21

19.40a 8.67 8.65

11.69 4.06 2.91

12.26a 7.00 8.21

6.77 5.42 2.99

9.50 5.13 20.46

7.26 3.36 5.96

8.56 7.17 19.01

6.27 4.76 4.98

Gender difference was significant at the p level of .05. Gender difference was significant at the p level of .01. Gender difference was significant at the p level of .001.

SD

SD

Female (N ¼ 18)

M. Zhou / Computers & Education 78 (2014) 174e184

183

Further, the current findings help us gain insight into how to support the development of learners' perceptions and skills in online learning so as to make intelligent decisions about online course design, delivery and pedagogy. Previous studies have found that gender could affect the outcomes of student online learning (e.g., Lim & Kim, 2003; Muilenburg & Berge, 2005). This study adds specificity to previous work by highlighting the need for differentiating online instructions and designs by student gender as well as performance levels. Stemming from the current results, online instructional strategies can be recommended to fully accommodate those differences in learner characteristics and augment learning opportunities. For example, one strategy is to make online instructions in such ways that provide ample opportunities to meet student needs and match student profiles (e.g., gender and performance level). This could include adopting different teaching styles, implementing online system adaptations, and customizing online learning activities and assignments to expand the learning opportunities for those students (especially average-performing students) and allow students with different backgrounds find effective support for their learning. For average-performing female students, the activities/assignments could be more encouraging to cultivate a stronger sense of achievement such that they could be more active and achieve better in online learning. The current findings need to be interpreted with cautions. One major limitation was the employment of biased samples e a majority of the participants were science/technology majors. Further research should work with a sample with a more representative cross-section of the population. Further studies are also needed to investigate whether the current findings can be replicated in other cultures and in students of different ages with regards to their information seeking skills. In addition, Navarro-Prieto et al. (1999) concluded that the type of search task had a stronger influence on the search strategy choice for experienced users than novice users. Thus, task type, along with task performance levels, could be examined to better understand gender differences in search behavior in future studies. In a nutshell, the formation of gender stereotypes in web search activities is a complex issue, which is influenced by many factors and largely documented in education in the West (Li & Kirkup, 2007). Wilson (1997) noted that gender is one of the intervening variables in information seeking behavior. The current findings suggested that these gender differences in online information search did not hold across performance levels. The roots of gender differences in web searching among adolescents can only be better understood when the issue is explored with individuals at different performance levels. The present study contributes new evidence and perspective to the current research on students' web search perceptions and skills, particularly concerning gender differences in different search performance groups. Given the lack of attention paid to Eastern students, more research should be conducted with eastern samples, and future research should point to the differences in how males and females orchestrated individual search behaviors to result in different overall search patterns. Moreover, cross-cultural comparisons of gender differences will be interesting since gender role and stereotypes vary across cultures (Xie et al., 2006).

References Agosto, D. E. (2004). Gender, technologies, and the school library. School Libraries Worldwide, 10, 39e51.
ne
cal, S. (2011). The effects of sex and Internet usage on search efficiency and effectiveness. International Journal of Computer Information Systems Arcand, M., Nantel, J., & Se and Industrial Management Applications, 3, 663e670. Bilal, D. (2001). Children's use of the Yahooligans! Web search engine: II. Cognitive and physical behaviors on research tasks. Journal of the American Society for Information Science and Technology, 52(2), 118e136. Brand-Gruwel, S., Wopereis, I., & Vermetten, Y. (2005). Information problem solving by experts and novices: analysis of a complex cognitive skill. Computers in Human Behavior, 21, 487e508. Brand-Gruwel, S., Wopereis, I., & Walraven, A. (2009). A descriptive model of information problem solving while using Internet. Computers & Education, 53, 1207e1217. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen, & J. S. Long (Eds.), Testing structural equation models (pp. 136e162). Beverley Hills, CA: Sage. Chen, S. Y., & Macredie, R. (2010). Web-based interaction: a review of three important human factors. International Journal of Information Management, 30(5), 379e387. Cheng, S.-S., Liu, Z.-F., Chen, N.-S., Shih, R.-C., & Chang, C.-S. (2012). Gender differences in college students' behaviors in an online question-answer discussion activity. The Asia-Pacific Education Researcher, 21, 244e256. Comber, C., Colley, A., Hargreaves, D. J., & Dorn, L. (1997). The effects of age, gender and computer experience upon computer attitudes. Educational Research, 39(2), 123e134. Dai, D. Y. (2001). A comparison of gender differences in academic self-concept and motivation between high-ability and average Chinese adolescents. Journal of Secondary Gifted Education, 13(1), 22e32. Ericsson, K. A., & Simon, H. A. (1984). Protocol analysis: Verbal reports as data. Cambridge, MA: MIT Press. Ford, N., & Miller, D. (1996). Gender differences in Internet perception and use. In Proceedings of the Third Electronic Library and Visual Information Research (ELVIRA) Conference, London, UK. Ford, N., Miller, D., & Moss, N. (2005). Web search strategies and human individual differences: cognitive and demographic factors, Internet attitudes, and approaches. Journal of the American Society for Information Science and Technology, 56(7), 741e756. €lscher, C., & Strube, G. (2000). Web search behavior of Internet experts and newbies. Computer Networks, 33(1), 337e346. Ho Hu, T., Zhang, X., Dai, H., & Zhang, P. (2012). An examination of gender differences among college students in their usage perceptions of the Internet. Education and Information Technologies, 17(3), 315e330. Hupfer, M. E., & Detlor, B. (2006). Gender and web information seeking: a self-concept orientation model. Journal of the American Society for Information Science and Technology, 57, 1105e1115. Jackson, L. A., Ervin, K. S., Gardner, P. D., & Schmitt, N. (2001). Gender and the internet: women communicating and men searching. Sex Roles, 44(5e6), 363e379. Jansen, B. J., & Solomon, L. (2010). Gender demographic targeting in sponsored search. In Proceedings of the 28th International Conference on Human Factors in Computing Systems (CHI) (pp. 831e840). New York: ACM. Joiner, R., Brosnan, M., Duffield, J., Gavin, J., & Maras, P. (2007). The relationship between Internet identification, Internet anxiety and Internet use. Computers in Human Behavior, 23, 1408e1420. Karavidas, M., Lim, N. K., & Katsikas, S. L. (2004, July). The effects of computers on older adult users. Poster presented at the annual meeting of the American Psychological Association, Honolulu, HI. Khalid, R., & Hasan, S. S. (2011). Test anxiety in high and low achievers. Pakistan Journal of Psychological Research, 24(2). Kim, D. Y., Lehto, X. Y., & Morrison, A. M. (2007). Gender differences in online travel information search: implications for marketing communications on the internet. Tourism Management, 28(2), 423e433. Kirkpatrick, H., & Cuban, L. (1998). Should we be worried? What the research says about gender differences in access, use, attitudes, and achievement with computers. Educational Technology, 38, 56e61. Konrad, M. A., Yang, Y., Goldberg, C., & Sullivan, E. S. (2005). Preferences for job attributes associated with work and family: a longitudinal study of career outcomes. Sex Roles, 53, 303e315. Koohang, A. (2004). Students' perceptions toward the use of the digital library in weekly web-based distance learning assignments portion of a hybrid programme. British Journal of Educational Technology, 35(5), 617e626. Koohang, A., & Durante, A. (2003). Learners' perceptions toward the web-based distance learning activities/assignments portion of an undergraduate hybrid instruction model. Journal of Information Technology Education, 2, 105e113.

184

M. Zhou / Computers & Education 78 (2014) 174e184

Large, A., Beheshti, J., & Rahman, T. (2002). Gender differences in collaborative web searching behavior: an elementary school study. Information Processing and Management, 38, 427e443. Laroche, M., Saad, G., Cleveland, M., & Browne, E. (2000). Gender differences in information search strategies for a Christmas gift. Journal of Consumer Marketing, 17(6), 500e524. Leong, S. C., & Hawamdeh, S. (1999). Gender and learning attitudes in using web-based science lessons. Information Research, 5(1). Li, N., & Kirkup, G. (2007). Gender and cultural differences in internet use: a study of China and the UK. Computers & Education, 48(2), 301e317. Liaw, S. S. (2002). An Internet survey for perceptions of computers and the world wide web. Computers in Human Behavior, 18, 17e35. Light, P., Littleton, K., Bale, S., Joiner, R., & Messer, D. (2000). Gender and social comparison effects in computer-based problem solving. Learning and Instruction, 10, 483e496. Lim, D. H., & Kim, H. (2003). Motivation and learner characteristics affecting online learning and learning application. Journal of Educational Technology Systems, 31(4), 423e439. Liu, Z., & Huang, X. (2008). Gender differences in the online reading environment. Journal of Documentation, 64(4), 616e626. Lorigo, L., Pan, B., Hembrooke, H., Joachims, T., Granka, L., & Gay, G. (2006). The influence of task and gender on search and evaluation behavior using Google. Information Processing & Management, 42(4), 123e131. Maghferat, P., & Stock, W. G. (2010). Gender-specific information search behavior. Webology, 7(3). Martin, J. L. (2010). Gender difference: the arguments regarding abilities. In M. A. Paludi (Ed.), Feminism and women's rights worldwide: Heritage, roles and issues (Vol. 1); (pp. 27e42). Santa Barbara, CA: ABC-CILO. McDonald, S., & Spencer, L. (2000). Gender differences in web navigation. In E. Balka, & R. Smith (Eds.), Women, work and computerization: Charting a course to the future (pp. 174e181). US: Springer. Meyers-Levy, J. (1989). Gender differences in information processing: a selectivity interpretation. In P. Cafferata, & A. Tybout (Eds.), Cognitive and affective responses to advertising (pp. 219e260). Ann Arbor, MA: Lexington. Moore, P. (1995). Information problem solving: a wider view of library skills. Contemporary Educational Psychology, 20, 1e31. Muilenburg, L. Y., & Berge, Z. L. (2005). Student barriers to online learning: a factor analytic study. Distance Education, 26(1), 29e48. Navarro-Prieto, R., Scaife, M., & Rogers, Y. (1999, July). Cognitive strategies in web searching. In Proceedings of the 5th Conference on Human Factors & the Web (pp. 43e56). Gaithersburg, Maryland. North, A. S., & Noyes, J. M. (2002). Gender influences on children's computer attitudes and cognitions. Computers in Human Behavior, 18, 135e150. Peng, H., Tsai, C.-C., & Wu, Y.-T. (2006). University students' self-efficacy and their attitudes toward the Internet: the role of students' perceptions of the Internet. Educational Studies, 32, 73e86. Preckel, F., Goetz, T., Pekrun, R., & Kleine, M. (2008). Gender differences in gifted and average-ability students. Gifted Child Quarterly, 52, 146e159. Ridgeway, C. L., & Correll, S. J. (2004). Motherhood as a status characteristic. Journal of Social Issues, 60, 683e790. Rosen, L. D., Sears, D. C., & Weil, M. M. (1987). Computerphobia. Behavior Research Methods, Instruments, & Computers, 19, 167e179. Roy, M., & Chi, M. T. H. (2003). Gender differences in patterns of searching the Web. Journal of Educational Computing Research, 29(3), 335e348. Sam, H. K., Othman, A. E. A., & Nordin, Z. S. (2005). Computer self-efficacy, computer anxiety, and attitudes toward the Internet: a study among undergraduates in Unimas. Journal of Educational Technology and Society, 8(4), 205e219. Savolainen, R. (2002). Network competence and information seeking on the Internet: from definitions towards a social cognitive model. Journal of Documentation, 58(2), 211e226. Schumacher, P., & Morahan-Martin, J. (2001). Gender, Internet and computer attitudes and experiences. Computers in Human Behavior, 17(1), 95e110. Sheu, C. M., Wang, P. L., & Hsu, L. (2013). Investigating EFL learning strategy use, GEPT performance, and gender difference among non-English major sophomores at a technological university. Asian EFL Journal, 15(1), 128e164. Silva, E. B. (2000). The cook, the cooker and the gendering of the kitchen. The Sociological Review, 48(4), 612e629. Singer, G., Norbisrath, U., & Lewandowski, D. (2012). Impact of gender and age on performing search tasks online. arXiv preprint arXiv:1206.1494. Steinerova, J., & Susol, J. (2007). Users' information behaviour e a gender perspective. Information Research, 12(3). paper 320. Retrieved October 12, 2009, from http:// informationr.net/ir/12-3/paper320.html. Tabatabai, D., & Shore, B. M. (2005). How experts and novices search the Web. Library & Information Science Research, 27(2), 222e248. Torkzadeh, G., & Koufteros, X. (1994). Factorial validity of a computer self-efficacy scale and the impact of computer training. Educational and Psychological Measurement, 54(3), 813e921. Tsai, M. J., Hou, H. T., Lai, M. L., Liu, W. Y., & Yang, F. Y. (2012a). Visual attention for solving multiple-choice science problem: an eye-tracking analysis. Computers & Education, 58(1), 375e385. Tsai, M., Liang, J., Hou, H., & Tsai, C. (2012b). University students' online information searching strategies in different search contexts. Australasian Journal of Educational Technology, 28(5), 881e895. Tsai, M. J., & Tsai, C. C. (2003). Information searching strategies in web-based science learning: the role of Internet self-efficacy. Innovations in Education and Teaching International, 40(1), 43e50. Tsai, M. J., & Tsai, C. C. (2010). Junior high school students' Internet usage and self-efficacy: a re-examination of the gender gap. Computers & Education, 54, 1182e1192. Tu, Y. W., Shih, M., & Tsai, C. C. (2008). Eighth graders' web searching strategies and outcomes: the role of task types, web experiences and epistemological beliefs. Computers & Education, 51(3), 1142e1153. Werner, J. M., & DeSimone, R. L. (2012). Human resource development. Mason, OH: South-Western College. Weber, I., & Jaimes, A. (2011). Who uses web search for what: and how. In Proceedings of the fourth ACM international conference on Web search and data mining (pp. 15e24). ACM. Weinburgh, M. (1995). Gender differences in student attitudes toward science: a meta-analysis of the literature from 1970 to 1991. Journal of Research in Science Teaching, 32(4), 387e398. White, P. A. (1989). Evidence for the use of information about internal events to improve the accuracy of causal reports. British Journal of Psychology, 80, 375e382. Wilson, T. D. (1997). Information behavior: an interdisciplinary perspective. Information Processing and Management, 33(4), 551e572. Winne, P. H., & Jamieson-Noel, D. L. (2002). Exploring students' calibration of self-reports about study tactics and achievement. Contemporary Educational Psychology, 27, 551e572. Xie, H., Bao, J., & Morais, D. (2006). Exploring gender differences in information search among domestic visitors to Yellow Mountain and Guilin. In Proceedings of the 2006 Northeastern Recreation Research Symposium, Bolton Landing, New York. Zhou, M. (2013a). A systematic understanding of successful web searches in information-based tasks. Journal of Educational Technology and Society, 16(1), 321e331. Zhou, M. (2013b). “I am really good at it” or “I am just feeling lucky”: the effects of emotions on information problem-solving. Educational Technology Research and Development, 61(3), 505e520. Zhou, M., Xu, Y., Su, Y., & Liu, L. (2011). A research tool for exploring student online search strategies and performance. In Proceedings of International Conference on Education and Educational Psychology (p. 159). Istanbul, Turkey. Zhou, M., & Winne, P. H. (2012). Modeling academic achievement by self-reported versus traced goal orientation. Learning and Instruction, 22(6), 413e419.