1 Introduction

The World Wide Web provides people with various sources of information, but one might get lost easily within enormous resources. It is not always easy to find the exact information at once, in a short duration, with one keyword, or by typing an exact URL. Web search engines are the tools helping us to eliminate the unrelated information within the vast majority of candidate resources. That is why; people utilize them as the primary search strategy (Hsieh-Yee 1998). The earliest search engine named Archie was very limited with regards to the resources it offered (Koster 1994). From Archie to Google, many features have changed. The former were listing the resources in an index-like fashion, but the contemporary search engines have much more space to access and list the detailed information of the found resources. Compared to the earlier examples, today, the users have much more freedom to customize the queries, including image search. Nonetheless, general-purpose search engines may not be useful all the time for all types of users. As a result of certain safety concerns, the safer search engines emerged for kids in time. Yahooligans, KidRex, Gogoligans, etc. are some examples to this kind. In addition, such search engines as scholar.google are also available for more specific audiences. Those efforts can be considered as important initiatives to manage the related resources more easily.

Although the range of Internet users is very expansive, the university students constitute the major part of the search engine users since they use those for their projects, especially at the beginning of the process (Biddix et al. 2011; Colon-Aguirre and Fleming-May 2012; Hsieh-Yee 1998). This may be due to the fact that accessing electronic resources is more convenient than accessing printed ones. Utilizing a search engine can bring about timesaving with less effort. The user may hope to find the exact information within one click after typing the keyword, but this is not generally the case. Depending on the searched information, the clicks can be more than one, whereas, some information can be found on search engine results page (SERP) without any clicks, i.e. just scanning or reading the title or explanation listed on SERP. Today, one more challenge has been added to web search process, which is the fact that the browsers allowing working with multi-tabs can make the process more sophisticated. The user needs to manage both SERP and the content opened in different tabs.

The type of a search task can shape the search patterns of users. While a close-ended (or ready-to use) task is considered as easy to search, an open-ended task is the exact opposite (Sendurur and Yildirim 2015; White and Iivonen 2002). According to Hogan et al. (2011), Google as a search engine is not appropriate for complex tasks because these tasks require collection of various resources with heterogeneous interfaces. A close-ended task can be readily found with simple queries, however, open-ended tasks demand customized search, more than one query sessions, or the comprehension of visited sites’ contents, which may in turn result in the keyword changes. Although general-purpose search engines provide advance search features, users generally do not prefer to use them (Moukdad and Large 2001). In addition, including customization within the web search process does not always guarantee reaching at the target content. In short, the user experience with search engines have multiple facets including pre-search decisions, interaction with the interface, examination of SERP, locating the related pages or information, and so on.

Initial interfaces of search engines seem very simple. For example, Google’s search engine has a quite simple design, which is exemplary in terms of white space usage. The user only needs to type a keyword, but the complexity of the search process begins before deciding on the appropriate keywords, therefore, the user experience during a web search can be somewhat sophisticated. For some users, SERP can turn into a maze. In terms of cognition, searching the web is much more beyond just typing a keyword. According to Wu et al. (2008), a search process includes certain cognitive stages: “recalling and inputting a query, navigating the query results, understanding the query results, judging the relevance of the results, learning the new information contained in the results, and refining the query if necessary” (p. 1831). If the task is complicated, then these stages can become harder to accomplish in a smooth fashion.

People use search engines to explore many things. As well as academic purposes, people also frequently use them for everyday issues, such as health problems, news, addresses, etc. One of the significant examples for daily use may be the case of travellers, who generally search the web for various purposes such as planning where to stay (Xiang and Gretzel 2010). There are many popular web sites containing real travelers’ comments, pictures, ratings, and so on. When one includes such keywords as “X city hotels” or “X city attractions”, the top ten on the SERP list will probably include such web sites. For a traveler, planning stage is considerably complex if she/he has never previously been to the city/country. A user might not want to feel overwhelmed while experiencing the web search, content of which is already challenging, so she/he might expect the interface and the overall context to let the search flow. On the other hand, while searching for an academic content, the cognitive load might be higher, but the expectations from the web search can be similar to everyday content search.

Although the process, interaction, expectations, decisions, keywords, scanning, filtering, reading, and selecting actions seem to be similar for both types of web searches (everyday vs. academic), there might be differences in terms of user experience. In this study, the search patterns were compared by using two different contents, therefore, the main research question of this study is:

“How do academic search patterns differ from everyday content search patterns?”

2 Method

The research design is a single-case study (Yin 2009), in which graduate students are the units of analysis. 15 undergraduate and graduate students from various departments voluntarily participated in the study. The researcher contacted participants online, and then scheduled the sessions, each of which took approximately 25 min. Before starting the data collection, each participant was informed about rights and ethical issues. They were allowed to leave whenever they want without any excuse. When they agreed on the both screen and voice recordings, the sessions started. 8 female (2 graduate; 6 undergraduate) and 7 male (2 graduate; 5 undergraduate) students were assigned to two web search tasks within two scenarios:

Scenario 1: The University is funding a 3-day trip to London for high-honor students and you are one of them. All expenses are going to be funded by the university, but you need to decide where to stay with a maximum budget of 500 Sterlin (Note that 500 Sterlin should be all spent for accommodation!).

Task 1: With the assigned budget (500 Sterlin), find three best accommodation options to stay between the given dates (1–4 Feb. 2016).

Scenario 2: For one of the courses you attend, the instructor assigned you to prepare a presentation about “Scientific Ethics and Plagiarism”. You can search the web and use maximum 3 web resources.

Task 2: In order to prepare your class presentation, find three best resources about “Scientific Ethics and Plagiarism” topic.

Participants were encouraged to search the web as they do in daily life. The starting point for both tasks were Google search engine. The order of the tasks was the same for all participants. There were no time restrictions for the tasks. During the whole sessions, all participants were encouraged to think aloud about the search process. Their eye-movements, speeches, gestures, and mimics were all recorded in the Human-Computer Interaction Laboratory. Eye tracking data can provide rich, realistic, and real time data (Webb and Renshaw 2008), for that reason, the main data collection and analysis techniques in this study were all based on the collected eye tracking data. Tobii 1750 Eye Tracker with 50 Hz sampling rate was utilized for data collection. Tobii Studio 3.4.0 software was used to analyze the recorded data. Participants’ think aloud data were analyzed qualitatively in order to shed light on their criteria or any action on the web search experiences.

3 Findings

3.1 Task 1

The first task was a regular task that anyone can perform while planning a trip, thus it is considered as an everyday web search task. For this task, the majority of the participants completed the session with keyword entry at once and there were certain popular keywords. Two common phrases were: ‘hotels in London’; and ‘accommodation London’. Only two female participants tried more than one keyword entries. These included ‘booking.com’, ‘cheap stay London’, ‘London hostels’ etc. They explained that the reason why they used these phrases was so that they would find either the cheapest or the highly rated hotels or hostels.

Except for five participants, the majority opened more than one tab on the browser (N = 10). The maximum number of tabs was 8. The number of visited sites also differed among participants. Nine participants performed only one web site visit and this was ‘booking.com’, however, all of the participants visited this web site as well (N = 15). The maximum number of visited sites was 5 (N = 2). The participants who worked with multiple tabs explored the hotels by opening them on another tab. Before making their final decision, they all navigated across tabs, and compared the ratings, images, prices, and other features and comments. Those who worked within one tab first explored the hotel, then immediately decided to choose or not to choose the hotel as an option.

While experiencing the web sites, the participants searched the best places to stay. Their selection criteria were similar, yet differed in terms of priority. From the think aloud data and eye tracker screen recordings, the overall criteria for this task can be listed as follows: photos (N = 14); budget (N = 13); ratings (N = 10); name (N = 5); and special offers (N = 2). When eye-tracking data were investigated qualitatively together with the think aloud data, the first or most preferred criterion of their hotel selection was found: budget (N = 6); photos (N = 3); travelers’ reviews (N = 2); ratings (N = 2); location (N = 1); and free Wi-Fi (N = 1).

Related to the listed criteria, some areas in booking.com website were pointed as Areas of Interest (AOI), because this site was visited by all participants. Price, photo, hotel name, rating, and stars were the main AOIs (Fig. 1). Time to first fixation, fixation count, visit duration, and percentage-fixated metrics were calculated with the help of Tobii software. The summary of the metrics is displayed in Table 1. As soon as the web site opened, it took shorter for price and rating AOIs to be fixated (M = 1.41 s for both), whereas it was longer for photo (M = 8.79 s) and hotel name (M = 13.61 s) AOIs. The highest fixation count belonged to price AOI (M = 8) with the longest visit duration (M = .75 s). In addition, the highest percentage fixated also belonged to respectively; price AOI (M = 57 %), photo (M = 50 %), and hotel name (M = 36 %) and rating (M = 36 %) AOIs came.

Fig. 1.
figure 1

Gaze plot and heatmap of main AOIs of booking.com: price, photo, hotel name, rating, and stars.

Full size image
Table 1. Metrics summary for task 1
Full size table

3.2 Task 2

The second task was very similar to regular academic tasks, since university students generally start their projects with the exploration of the web with the utilization of search engines (Biddix et al. 2011). For this task, most of the participants entered more than one keyword (N = 10). Only five participants entered keywords once, but the rest of the participants’ entries had a range between 2–8 times. Interestingly, ‘scientific ethics and plagiarism’ was the first keyword phrase of all participants.

Most of the participants preferred to work with multiple tabs on the browser (N = 10), but some of them explored the pages within one tab (N = 5). The maximum number of tabs was 7. The students working with multiple tabs visited more pages, and focused less on the main parts of the documents. On the other hand, those working within one tab visited less web site, but were more tended to read the main body of the pages.

For this task, minimum number of web site visits was 4 and the maximum number was 10. Although the number of visits varied among participants, there were some common web sites visited: ‘yok.gov.tr’, ‘researchgate.net’, ‘uvt.ulakbim.gov.tr’, ‘biotek.ankara.edu.tr’. Moreover, most of the participants reached the same document from different web sites. For example, a scientific article was available on both biotek.ankara.edu.tr and researchgate.net. This situation enabled the examination of the same documents in terms of eye movements’ comparison of different users.

During the examination of the electronic resources, the participants tried to eliminate the unrelated content, thus they set certain criteria to select the best ones. Their selection criteria were similar, however, differed in terms of priority. The overall criteria for the selections can be listed as follows: published article (N = 11); title of the author (N = 7); gov and edu extensions (N = 7); well-organized (N = 4); and the number of citations (N = 2). From either think aloud data or eye-tracking data, it could be inferred that among the listed criteria, the primary concern or criterion of participants can be summarized respectively: title of the author (N = 4); rich content (N = 4); being a published source (N = 4); simple and organized content (N = 2); and edu and gov extensions (N = 1). One of the common websites visited for this task was biotek.ankara.edu.tr. Considering the popular criteria, the AOIs for this web site were title, author, abstract, keywords, and name of the journal (publisher) (Fig. 2). It took shortest to the first fixation of title (M = 1.43 s) AOI and longest to the first fixation of keywords (M = 8.33 s) AOI. The highest fixation count was within abstract AOI (M = 12) with the longest visit duration (M = 1.05 s). In terms of percentage fixated metric, the abstract AOI had the highest percentage (M = 43 %), then title (M = 36 %), author (M = 21 %), keywords (M = 21 %), and journal (M = 14 %) AOIs came respectively (Table 2).

Fig. 2.
figure 2

Gazeplot and heatmap of an article

Full size image
Table 2. Metrics summary for task 2
Full size table

4 Discussion and Conclusion

The findings point that web search for everyday content can be less complicated than that for academic content. Keyword entries and visited websites differed in number for academic search. Although common keywords were observed for either task, participants used more keywords for the academic task. Moreover, the number of web site visits was also higher for the academic task. On the other hand, participants visited generally booking.com website in order to decide on the best three hotels, and they did not feel the necessity to check the hotel (with regards to pricing, photos, features, etc.) somewhere else. This was not the case for academic search, because they find it necessary to check whether the information they obtain is reliable.

While participants trust the reviewers on a web site for travellers, they are more apprehensive to trust any academic content, i.e. they consider who the author is. Reviewers have either nicknames or real names with profiles showing the number of reviews. Their reviews are listed on a well-known web site, which may be the reason why participants rely on the reviews on the site; another reason may be because booking.com is located generally at the top of the SERP. The rank of a page can be sign of how prestigious the site is (Witten 2008), therefore, they might not consider it necessary to look somewhere else. In other words, this site can be perceived as an authority like the author of an academic paper or the publisher of an article. Such a web site can also be considered as a social networking site, so the power of reviews is crucial while planning a trip (Kim et al. 2014; Xiang and Gretzel 2010).

It seems that participants have some common beliefs for the everyday content. They trust a website presenting real travelers’ reviews, and do not feel the need to visit any other website; however, for academic task, they first judged the formality of the website. They did not prefer to visit blogs, wikis, or news web sites. Such small pieces of information as budget (or price) were the main concern or primary selection criteria for the everyday task; whereas it was the origin of the sources for academic task. The users all valued where it was published, who the authors were, and the extension of the web site providing access.

Working with multiple tasks can be a sign of multitasking, which was observed for both tasks, yet, there were a few students searching within one tab. The one-tab users’ overall search process was not very detailed for everyday task, whereas, they read main body of the documents more than others. Comparing this situation for either tasks may demonstrate that academic search may demand more comprehension; because in everyday task, small pieces of information may be enough to evaluate whether it is a good option or not. Since the web search is a way to distinguish what we do not know within a vast majority of information (Witten 2008), not only multitasking but also single tasking can be overwhelming during any search process.

For the everyday task, participants’ most priorities criterion was budget (or price); which did photographs of the hotel, and travelers’ reviews follow respectively. The eye tracking data asserted that it took less time to the first fixation to the price AOI, then rating (or review) AOI. Price AOI had also the highest fixation count and percentage fixated was higher. In short, all data about criteria for the everyday task were consistent. Although the photos of hotels seemed important at first sight, which was inferred from participants’ responses, that AOI did not strike too much attention. It came after price or ratings, which might mean that when the search task has limits, the visual attention can be shaped along with that.

For academic task, participants’ primary criterion was the title of the authors; which was respectively followed by content and its publishing status. However, the eye tracking data did not support it; since the title of the article was the AOI stroke attention first. On the other hand, the abstract AOI had the most frequently fixated location, which may mean that participants wanted to understand the content in a deeper way. The overall meaning should be extracted from the whole text, and it is possible through reading the text.

The findings of the study are limited to 15 participants. Although the direct observation and eye tracking data, which is objective in nature, gave insights of web searchers, more detailed analyses with more participants are required. A true experimental design can be appropriate in order to compare the tasks or to see how different level searchers conduct the web search.