Recently, many studies investigates the role of individual (e.g. gender) and cognitive differences (e.g. cognitive style) during Web navigation and Web searching. Despite this settled interest, limited works have considered the role of the individual differences in real-environment navigation during Web navigation. For this reason, the main aim of this work is to investigates the effect of different spatial cognitive styles, Landmark style (LS), Route style (RS) and Survey style (SS), on the Web searching behaviour. More specifically, we hypothesize that individual that generally use landmark strategies during real environmental navigation have more difficulties during the Web exploration. To establish the hypothesis, we asked to 30 University students (10 LS, 10 RS, 10 SS) to solve three Web information tasks. The spatial cognitive style was assessed with the Spatial Cognitive Style Test, then participant had to fill in a questionnaire about internet and computer use. We also developed an ad-hoc key-logger program for browser, in order to collect Web behaviour measures. Particularly, measures considered are: search engines instruments used (e.g. back button), pages visited and revisited, time spent in information searching and mouse cursor movements. Results showed significant differences between the spatial cognitive styles in the number of pages revisits, where individuals with LS seem to use a trial and error strategy, in order to get the relevant information. Furthermore, differences emerge also in the distribution of mouse cursor movements during Web navigation. Implications for future works are discussed.


Most people in Western world have a computer with an Internet access at home; many of them use the computer and search engines daily both at home and at work. These people have different age and habits with computers: teenagers and young adults of the 21st century were born and grown up with computers and Internet, for this reason they are defined by Prensky as “Native Digital”; conversely, older people, whom began to use computer for work and other necessities, are called “digital immigrant” (Prensky, 2001). Furthermore, the human interaction with computer technology has changed over the past 50 years, so we can consider three waves of computing (Weiser, 1993; Shiode, 2004): a) first wave goes from 1960 to 1980 and it is called “Mainframe Era”, in which there was only one computer shared by many people; b) second wave is called “Personal Computing Era”, it goes from 1980 to 2000, in which every person has a personal computer; c) the third wave is called “Ubiquitous Computing Era”, that goes from 2000 to now, in which a person has many computer and other devices.

Due to this variability, it seems necessary for the field of computer science, to understand which factors affect computer users’ behaviour during computer interaction. In particular, an interesting part of research focused on Web searching experience. Literature tried to identify the internal factors, such as computer expertise or individual characteristics, and external factors, like Web browser features, may influence Web searching behaviour (for a review see Spink & Jansen, 2006). Evaluating the role of these elements means have a deeply understanding of which criteria are fundamental in Websites design, to better provide contents and improve search engine efficacy.

An external factor considered is, for example, the impact of display size during Web searching, in terms of usability, in particular smartphone or tablets screen compared to bigger computer displays, and the two different Web interfaces used in these devices, during Web navigation researches. At the dawn of the spread of devices with Web browsing capabilities, the existing Website interfaces were not appropriate to this new tools, and this led to general difficulties during Web searching (e.g. Jones, Marsden, Mohd-Nasir, Boone & Buchanan, 1999); for this reason, designers and researchers are more and more pushed to improve Web navigation on smaller devices (e.g. Xie, Miao, Song & Ma, 2005).

Within internal factors, a huge number of studies focused on individual differences; one of the most important, is gender. Many authors explored how gender affects computer skills and Web searching. For example, several studies showed that men are more interested in using computers and Web browsing compared to women (Light, Littleton, Bale, Joiner & Messer, 2000; Schumacher & Morahan-Martin, 2001). Furthermore, men and women purposes during Web browsing seem different: women tend to search the Web for relational concerns, like sharing ideas on communities or join Webchats; conversely, men are more interested in personal activities and facing a minor cognitive burden to find information (Smith & Whitlark, 2001; Jackson, Ervin, Gardner & Schmitt, 2001).

Numerous contributions have also discovered links between personality and internet usage. People high in Neuroticism (a personality trait of the Big Five, e.g. Goldberg, 1990), often use Internet and especially social networks, to avoid loneliness and socialize with other people (e.g. Amichai-Hamburger & Vinitzky, 2010) while those high in Extraversion tend to make friendship outside the virtual world and use internet as an instrument to keep in touch (Ross et al, 2009).

Cognitive style (e.g. Riding & Cheema, 1991) is another individual difference considered, which emerge as an important factor during Web searching behaviour and information processing. Between the 1940s and 1980s, many researchers have developed their own theory and instruments to evaluate cognitive style. Riding and Cheema (1991) grouped the cognitive style into two dimensions: wholistic-analytic (or field dependent-independent) and verbal-imagery (or verbalizer-visualizer). The wholistic or field-dependent cognitive users tend to see a situation as a whole picture (Riding, 1997). They are able to structure and analysing problem solving and learning. Analytic or field-independent cognitive style users see a situation as a collection of parts and focus on one or two aspects of the situation at time. They are good to see similarities, detecting differences, and providing their own structuring in intellectual activity. The verbal-imagery cognitive style describes the person tendency to elaborate the information using a verbal code or a picture one (Riding & Cheema, 1991). Specifically, verbal (verbalizer) cognitive style users think in terms of words and consider the information they read, see, or listen to, in words or verbal associations. Instead imagers (visualizer) cognitive style users think in terms of mental pictures: when they read, see, or listen to, they consider it in pictures. Only limited studies have been conducted to explore cognitive styles among different information and web user (Kinley et al., 2013). For example, Palmquist and Kim (2000) also found that field-dependent cognitive style and lower Web searching expertise are related: users take more time and actions than required, to find information. More recently, Kinley and Tjondronegoro (2010) found that verbal users tend to navigate in Internet in a nonstructural mode – for example they often reformulate queries and scan several pages quickly; conversely, imagery users (who represents knowledge in mental pictures) seem to have a more linear and structural navigational behaviour – for example, they read all pages and spend more time for searching information. Kinley et al. (2014) showed that wholistic and verbalizers followed a top-down search approach while searching information on the web: that is, they search for general information and then gradually searched for specific information. On the other hand, analytic and imagers preferred a bottom-up approach while performing web searches, searching for specific information by using a lot of search terms in their succeeding query. Moreover, participants’ search queries were categorized into New, Add, Remove, Replace, and Repeat. A significant difference was found among wholistic and analytic: wholistic were found to use more new and repeat queries than analytic. Moreover, verbalizers executed a higher number of Add, Remove and Replace query reformulation then imagery cognitive style users. Cognitive style is also a key factor in the development of hypermedia learning systems, because of the individual differences in information processing (Liu & Reed, 1994; Papanikolau, Grigoriadou, Magoulas & Kornilakis, 2002; Lee, Cheng, Rai & Depickere, 2005), even though some studies do not attribute the same importance to this factor (Calcaterra, Antonietti & Underwood, 2005).

Obviously, Internet expertise arise as significant element that affects Web searching behaviour. In Lazonder, Biemans and Woperis (2000), Internet expertise is associated with better performance during Web searching, in facts expert users find a larger number of corrected information in a shorter time compared to novice users. Moreover, users with high level of expertise utilize complex queries and advanced search operators –e.g. Boolean operators- usually not used by average and novice users (e.g. Hӧlscher & Strube, 2000). An interesting effect related to Web searching experience, is a spatial disorientation during navigation, which means that users are unable to locate their position in hypermedia during Web interaction (Thüring, Hanneman & Haake, 1995). Herder and Juvina (2004) discovered two navigational styles associated to the perception of disorientation: 1) “Flimsy navigation style” reflects a low Internet expertise and it is related to users’ perceived disorientation, with high rate of home page and less number of pages visited; 2) “Laborious navigational style”, in which users employ a trial and error strategy, open many link for see if they are useful, otherwise return to previous pages, often using back button, and follow other links. This navigation style is also related to low score in mental rotation tasks (Juvina & Oostendorp, 2004).

Despite the importance of individual and cognitive differences, few studies considered the potential contribute of spatial and environmental navigation skills on Web searching behaviour. Nevertheless, spatial abilities, in particular spatial orientation, seem to affect performance during information retrieval tasks in Internet (e.g. Pak, Rogers & Fisk, 2006). It would be interesting to deepen the knowledge about the relations between sense of direction, which is the ability to find a destination from a starting place (e.g. Golledge, 1999; Lawton & Kallai, 2002) and Web navigation.

People differ significantly in navigational skills, in fact there are those who easily find how to reach a place and those who get lost frequently. It is well known in literature that men have a better sense of direction compared to women (e.g. Lawton, 1994; MacFadden, Elias & Saucier, 2003; Nori & Piccardi, 2011; Piccardi, Risetti & Nori, 2011; Nori & Piccardi, 2015). Another important factor that help us to explain differences in sense of direction is the way in which people acquire and represent the spatial information. Indeed, Siegel and White (1975) assume that there are three mental representations: landmark, route and survey. According to Pazzaglia, Cornoldi and De Beni (2000) these spatial representations correspond to three spatial cognitive styles, which reflect the strategy used for self-orienting. Specifically, in landmark cognitive style, people utilize only salient elements of the environment, like buildings, churches, train station, etc., to orient themselves, thus there is no spatial information and relations between these elements and in order to reach their goal, they use a trial and error strategy. In route cognitive style, people utilize pathways that generally connect landmarks to each other. It is a sensorimotor strategy, organized on egocentric coordinates (i.e. right-left and front-behind). Every step of the path is necessary to find the final destination: if one element of the sequence is missing, the person cannot reach the goal. Latest, in the survey cognitive style, the person has a complete spatial configuration of the environment. The spatial knowledge is based on allocentric coordinates, like cardinal points, so it does not depend on person’s position.

The aim of this study is to investigate whether spatial cognitive style may influence Web searching behaviour. Findings can provide a better comprehension of computer users’ behaviour during Web browsing to designers and researchers. Specifically, the assumptions underlying this research are:

i) Spatial cognitive style affects the modality of using the search tools provided by search engines. Specifically, it is hypothesized that landmark style users tend to open more links, revisit more pages and use more the back button comparing to route style users and survey style users. This behaviour could be due to the characteristics of landmark cognitive style, in which, during environmental navigation, a person with this style utilizes a trial and error strategy, that is expected to be used also during Web navigation, like in the laborious navigational style found by Juvina and Oostendorp (2004) and the sporadic navigation style found by Kinley and Tjoindronegoro (2010); ii) Spatial cognitive style affects the success in finding Information during Web searching. In particular, landmark users are expected to spend more time and find less information with respect to route users and especially with respect to survey users, because they use less useful and linear strategies considering time and actions and they find less information as a result; iii) Spatial cognitive style affects the screen and page exploration area that reflects the modality of exploration in real environment. Particularly, landmark users tend to explore less the hypermedia space with the mouse compared to route users and especially to survey users, reflecting the real world navigation, in which they tend to navigate only in a well known environment and avoid largest space, because of fear to get lost.



Thirty undergraduates from several academic disciplines took part in the study (11 men), the age range is from 18 to 30 years (M = 24.03, S.D. = 3.12). The spatial cognitive style was assessed with the Spatial Cognitive Style Test (Nori & Giusberti, 2006), described below, and the thirty participants were subdivided in: 10 landmarks style (LS: 1 men and 9 women), 10 routes style (RS: 4 men and 6 women) and 10 surveys style (SS: 6 men and 4 women). In order to exclude participants with self-reported topographical orientation disorders, the general sense of direction was evaluated in all participants by the Familiarity and Spatial Cognitive Style scale (FSCS: Piccardi et al., 2011). None of the participants showed the presence of navigational deficits or developmental topographical disorientation (see Bianchini et al., 2010; Iaria et al., 2005, 2009). The study was approved by the local Ethics Committee of the Psychology Department of Bologna University and all participants gave their written informed consent, in accordance with the Declaration of Helsinki.

Materials and Procedure

The study was composed of a part in which the participant had to search information on the Web and a part in which he or she had to complete test and some questionnaires, including a part in which demographic information are required.

Technical Instruments

We used a computer with Windows XP operative system, Google Chrome browser, two screens (one for the participant’s Web session and one for checking the software) and a camera for video records participant’s behaviour and the information they found in the search task.

To perform the study, a specific software was developed, using application programming interface (API) exposed by Google Chrome Web browser. Application was able to: a) record mouse movements and scrolling –e.g. x and y coordinates on the display or document; b) record keyboard input –e.g. key pressed, key released; c) query browser state –e.g. on loading, windows geometry, etc.; d) query single browser tab state –e.g. active, URL, etc.; e) collect time spent on each page and total navigation time; f) perform screen recording. We decided to consider these measures because they were taking into account also in previous works on Web searching tasks (e.g. Herder & Juvina, 2004; Pak, Rogers & Fisk, 2006).

The software is composed of three main components:

The software is free available at: https://github.com/MatteoRagni/LauraKeylogger.

Search Tasks

We decided to use the Web search tasks developed by Kinley and Tjondronegoro (2010), based on the concept of “simulated work task situation” (Borlund & Ingwersen, 1997), which are designed for being as close as possible to real situation of Web information researches. This conception allows participants to reinterpret subjectively the problem, and it was used in other studies on Web interaction behaviour (Kim, 2009). We decided to use this kind of tasks because the participants could search the required information as they usually do. These tasks have different level of difficulty (from 1= easy to 3=hard). Taking into account the subjects’ natural context and avoiding biases in test’s results, the first task –that in the American original version required knowledge of the judicial American field and a high level of English language–was modified and the participants received the following instructions:

  1. “You have become a parent recently and you would like to know the Italian laws regarding child safety while traveling in vehicles. Identify three such rules.”

    The other two tasks were the original two used in Kinley and Tjondronegoro (2010), translated in Italian language:

  2. “You, with two friends, are planning a trek for one week in Solukhumbu in Nepal. The trekking will occur next month. You are told that tourists trekking in the place may get high-altitude illness. You decide that you should know more about the place, and symptoms, seriousness and prevention of high-altitude sickness.”

  3. (3) “There has been talk of the Bermuda Triangle mystery for the last three decades or so. You are curious about the mystery and want to know more about it. So, you want to search any incidents, people’s views and any other relevant information (literature, images and videos) about it.”


For every task, we decided to attribute a score based on information found, to be able to establish the second hypothesize: 1 point for correct information, 0.5 points for not complete information and 0 points for uncorrected information.

Spatial Cognitive Style Test

The Spatial Cognitive Style Test (SCST, Nori & Giusberti, 2003; 2006) was used to assess the participants’ spatial cognitive style. It consists of nine spatial tasks, each one composed of 7 items. The score goes from 0 to 7.

The first two tasks (photo task and figure task) are representative of landmark cognitive style, because in order to solve them, participants do not need to refer to any spatial information. The subsequent four tasks (sequence task, map description task, right-left discrimination task and 2D rotation task) reflects a route spatial strategy because in order to solve them, participants have to rely on egocentric point of view and a right-left discrimination. The last three tasks (path task, 3D rotation task and sum and straighten task) represent survey cognitive style, because to solve them correctly, participants have to rely on an abstract representation and an allocentric point of view.

The order of these tasks is randomized. To establish the cognitive style, the score in every task is summed with other in the same category (landmark, route and survey tasks); afterwards they are divided for the number of the tasks in the correspondent category. If the number obtained in landmark tasks is equal or major to the 80% of the correspondent items and minor of 50% of the route and survey items, the participant utilize landmark strategy or cognitive style. Whether the score is at least 80% of correct answer for landmark and route tasks but inferior to 50% of the survey tasks, the participant utilize route strategy or cognitive style. Lastly, if the score is at least 80% of correct answers in every category of tasks, the participant utilize survey strategy or cognitive style.

Computer Use Questionnaire

The CUQ (Antonietti, 1998) it is a self-report instrument that investigates the capabilities and the frequency in the use of computer. Specifically, focuses on the following categories: a) videogames, b) word processor, c) calculation and statistical programs, d) e-mail, e) programming, f) educational programs, g) Web browsing, h) chats. Due to the lack of references to mobile Technologies and social networking, some categories are added to the previous. Specifically, the new items were about: a) smartphone/tablet usage, b) browser used, d) time spent for social network browsing vs. search engine browsing –this because from some studies emerged a different behaviour during social networks browsing, like Twitter, and general Web browsing in search engines (Teevan, Ramage & Morris, 2011). Two initial questions ask the average time spent on computer during work and free time. Then, for every activity, participant had to respond on a 5 scale item for the frequency (from 1= never’ to 5 = ever) and for the ability (from 1 = very poor to 5 = very good).


After having signed the written informed consent and fill in the questionnaire about demographics information, participants were subdivided, in a balanced order, in those who did the Web session before and those who, vice versa, first filled in the questionnaires and did the SCST. In the Web session, the experimenter administered in a randomized order the three Web search tasks. Then, he explained to the participant that s/he had to Web browsing to research the required information, communicate the information found and explain what s/he did during the session. Furthermore, it was explained that there is not a limit of time to complete the task and the participant could use every instrument provided by the search engine. When the participant read the first task, the experimenter started the software and the video recording, and the session finished only when the participant decided to stop it, thus when s/he thought that have found every information required. All the participant’s actions were automatically logged. The same procedure was repeated for the other tasks. When a task was completed, the experimenter stopped the software and saved the data.

The whole procedure took about 50 minutes.


We performed a series of regression analysis considering gender and age as independent variables, on the following Web navigation measures information found in every task, total number of page visited, number of return to homepage, back button usage, number of time participant returned to previous pages (obtained by the total number of pages visited minus the different URL visited), total time spent on a task, mean time per page in order to investigate if these variables could predict the participants’ performance. The analysis revealed no significant differences for what concerns information found, number of page visited, amount of home page return, total time spent and mean time spent in every page for completing a task, number of home page visit in task 1 and 2 and use of back button in task 1 and 2 (F2,29 = from 0.03 to 3.00, r2 = from -.07 to .12, ps = from .07 to .96); otherwise, the analysis revealed significant differences of gender in the number of home page return in task 3 (F2,29 = 7.97, r2 = .32, p < 0.01), in which men tend to return to the homepage more frequently than women, and in the back button usage in task 3 (F2,29 = 4.14, r2 .17, p < 0.03), in which, conversely, women tend to use more frequently the back button.

Furthermore, other regression analysis were performed considering the frequency of use computer and the other aspects of computer activities, assessed with the CUQ, as independent variables, on the Web navigation measures again and no significant differences were found (F5,29 = from 0.2 to 1.74, r2 = from - .13 to .05, ps = from .16 to .89).

To test the first hypothesize, that is spatial cognitive style affects the modality of use the search tools provided by search engines, we performed ANOVAs with the three levels of task as within independent variable (task 1-task 2-task 3) and the three levels of spatial cognitive style (LS-RS-SS) as between independent variable on the Web navigation measures considered; we considered also gender as covariate variable on back button usage measure and home page revisit measure in the appropriate dependent variable.

First we considered the page visited in the three tasks as dependent variable. Results showed that there is not a main effect of the spatial cognitive style (F2,27 = 1.15, p = .33, η2p = .07); conversely, there is a main effect of the task (F2,26 = 17.16, p < 0.01, η2p = .56). Pairwise comparison with Bonferroni adjustment showed that task 1 (M = 4.23, SD = .43) differed significantly (ps < 0.05) from task 2 (M = 8.76, SD = 1.18) and task 3 (M = 7.33, SD = .78). There is not a significant interaction “Task x Spatial Cognitive Style” (F4,54 = .77, p = .54, η2p = .05).

Then, we considered the back button usage in the three tasks as dependent variable and we insert gender as covariate. Results showed that there is not a main effect of the spatial cognitive style (F2,26 = .66, p = .52, η2p = .04); conversely, there is a main effect of the task (F2,25 = 9.66, p < 0.01, η2p = .43). Pairwise comparison with Bonferroni adjustment showed that task 1 (M = .53, SD = .16) differ significantly (ps < 0.01) from task 2 (M = 1.76, SD = .39) and task 3 (M = 2.03, SD = .30. There is not a significant interaction “Task x Spatial Cognitive Style” (F4,52 = .64, p = .63, η2p = .04).

We then considered the homepage return in the three tasks as dependent variable and we insert gender as covariate. Results showed that there is not a main effect of the spatial cognitive style (F2,26 = .92, p = .41, η2p = .06) and neither a main effect of the task (F2,25 = 2.96, p = .07, η2p = .19). There is also not a significant interaction “Task x Spatial Cognitive Style” (F4,52 = .54, p = .70, η2p = .04).

Lastly, we considered the number of revisit previous pages, obtained by taking into account the total number of URL visited minus the number of different URL visited. Results showed a main effect of the spatial cognitive style (F2,27 = 3.87, p < 0.05, η2p = .22). Pairwise comparison with Bonferroni adjustment showed that those with a landmark cognitive style (M = 0.9, SD = .21) revisit more previous pages compared to those with a survey cognitive style (M = .06, SD = .21). There is not a main effect of task (F2,26 = 1.71, p = .20, η2p = .11) and also not a significant interaction “Task x Spatial Cognitive Style” (F4,54 = .94, p = .42, η2p = .06).

To establish the second hypothesize, that is spatial cognitive style affects the success in finding Information during Web searching, we performed ANOVAs with the three levels of task (task 1-task 2- task 3) as within independent variable and the three levels of spatial cognitive style (LS-RS-SS) as between independent variable on the information found, the time spent on every page visited and the total time spent in every task.

We considered the information found in the three tasks as dependent variable. Results showed that there is not a main effect of the spatial cognitive style (F2,27 = 0.62, p = .90, η2p = .00) and neither of the task (F2,26 = 0.09, p = .54, η2p = .04) and there is also not a significant interaction “Task x Spatial Cognitive Style” (F4,54 = 1.18, p = .32, η2p = .08).

Moreover, we considered the mean time spent on a page as a dependent variable. Results showed that there is not a main effect of the spatial cognitive style (F2,27 = .11, p = .88, η2p = .00), and a main effect of the task (F2,26 = .45, p = .64, η2p = .03). There is also not a significant interaction “Task x Spatial Cognitive Style” (F4,54 = .38, p = .82, η2p = .02).

Lastly, we considered the total time spent on a task as dependent variable. Results showed that there is not a main effect of the spatial cognitive style (F2,27 = .23, p = .79, η2p = .01). Conversely, there is a main effect of the task (F2,26 = 21.41, p =< .01, η2p = .62). Pairwise comparison with Bonferroni adjustment showed that task 1 (M = 246.37s, SD = 26.56s) differ significantly (ps < 0.01) from task 2 (M = 496.31s, SD = 55.33s) and task 3 (M = 437.80, SD = 44.50). There is also not a significant interaction “Task x Spatial Cognitive Style” (F4,54 = .29, p = .87, η2p = .02).

In order to establish the third hypothesize, that is spatial cognitive style affects the screen and page exploration area using the mouse, we drawn three graphs, obtained by considering the mouse movements during the three tasks for every spatial cognitive style. Specifically, graphs present multi-modal distribution of mouse coordinates. The multi-modal distribution is the result of sum of normal distributions, centred in mouse coordinates in pixels \(\left( x,y \right)\), with variances equal to mouse cursor dimensions (here as \(\rho\)). Given for a task of one of the participant:

\[\mathbf{p}\left( t \right) = \left\{ x\left( t \right),y\left( t \right) \right\}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ t \in \lbrack t_{i},t_{f}\rbrack\]

the distribution for his/her task is:

\[h(\mathbf{p}) = \int_{t_{i}}^{t_{t}}{\mathcal{N(}\mathbf{p},\rho\mathbb{I})dt}\]

and cumulative distribution for all \(N\) participants use the same spatial cognitive style —landmark, route or survey— is:

\[H\left( \mathbf{p} \right) = \frac{\sum_{i = 1}^{N}{h_{i}\left( \mathbf{p} \right)}}{\max\sum_{i = 1}^{N}{h_{i}\left( \mathbf{p} \right)}}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ H(\mathbf{p}) \in \lbrack 0,1\rbrack\]

that is presented in Figure 1, through contour plots with maximum (\(\approx 1\) ) in black and background (\(\approx 0\)) in white. The symbol \(\times\) shows the mean coordinates of the mouse during a single task, of the participants that use the same navigational strategy or spatial cognitive style.

Figure 1. Multimodal distributions for mouse coordinates through sessions per each spatial cognitive style, with respect to tasks. The red cross is the mean of the distribution.


The work investigates the role of different navigational strategies for environmental exploration, here as spatial cognitive styles, in relation to Web searching behaviour. Particularly, it is hypothesized that people with different spatial cognitive style search on the Web by using different strategies, in terms of Web searching instruments used (e.g. back button), visits and pages’ revisits, number of correct information found, time spent on searching information and mouse movements. We expected that a landmark cognitive style user navigates on the Web redundantly, using often the back button, revisiting pages and returning to homepage more frequently, thus, taking more time to find information because of this unorganized behaviour, compared to user with other cognitive spatial styles. This leads to a lower quality information scraping and a limited exploration of the hypermedia space, i.e. in the use of the mouse cursor. Conversely as expected, most of the metrics considered showed almost no differences between the three spatial cognitive styles. The only difference emerged is in the amount of pages’ revisiting. In particular, participants with a landmark cognitive style revisit more often previous pages, with respect to the participants with survey cognitive style. This behaviour reflects a trial and error strategy evident also during real environment navigation of landmark style people: if they are not sure about the correctness of ”where they are”, they return back to previous known places and try again to reach the goal. During Web navigation, if landmark users do not find the information required in a Web page, they revise the previous pages and then try with a new Web page. Revisiting is not only achieved through back button, but also through a redundant search, as employing the same query used before. This behaviour is in a certain sense, similar to the laborious navigational style found in Herder and Juvina (2004) and also at the sporadic navigational style described in Kinley and Tjondronegoro (2010), in which users tend to use an unstructured navigation during Web searching. These people visited numerous links, pages, switching between browser tabs and windows, return back to previous pages and their Web search is characterized by a short duration. They also tended to visit the homepage more frequently and use the back button more often than others, which is an indication that they felt uncertain about their searching. Users of this kind, as suggested by Kinley and Tjondronegoro (2010) were found to be unorganized. Landmark cognitive style users as sporadic navigational style users tend to revisit previous page and then try with other links. This behaviour could be an indication of “getting lost” (Palmquist & Kim, 2000) in Web searching, Landmarks users stop whatever they have been doing and starting over again this is similar to the strategy used in the environment, that is when they getting lost they tend to search a familiar place or landmark to start again their goal search (e.g., Nori & Piccardi, 2011). These people tend to navigate both in real environment and in web in a non-linear mode. In a very recent study, Piccardi and co-workers (submitted) analyzing eye movements during the learning of a schematic environmental map found that the three cognitive spatial styles are characterized by a different pattern of eye movement exploration. In line with the present results, they showed that landmark style users made the greater number of fixations and runs, and had the shortest fixation durations and every spatial cognitive style differed from one to another.

Here, differently from landmark style users, route and survey style users follow a linear navigation: they revisit less previous links and they appear more confident in Web searching. Similar to environmental navigation, these people keep in mind the relevant information which is useful to orient themselves in a sort of online updating of (spatial) information. Indeed, in real environment, during wayfinding activities, travelers are used to make decisions about spatial sequential process in which the purpose is to match internal with external cues (Stern & Portugali, 1999). Afterwards, during wayfinding activity, the environment is to be considered a dynamic source of information used by travelers in their decision-making processes: when a person arrives at a relevant landmark for his/her trajectory, s/he thinks about where the successive landmark is in relation to his/her position. Whereas route and survey cognitive style users are able to do this, landmark cognitive style users do not and for such a reason have to come back at the previous landmark (i.e., previous link in the web navigation). However, route and survey cognitive style users do not spend less time to keep the right information compared to landmark style users. Therefore, higher performance outcomes are not associated with time spent during navigation. This result is in line with Yin’s (2001) work which suggested that the amount of time spent using a computer-based presentation program in a multimedia learning environment does not predict performance.

These results are also in line with Kinley et al.’s (2014) model which pointed out that users with a certain cognitive style, that is Analytic/Wholistic and Verbalizer/Imagers, followed a specific web search behaviour, while other types of cognitive styles showed different search behaviours. Specifically, verbalizers tend to follow a sporadic navigational style whereas imagers show structured navigational behaviours whereas both analytic and wholistic use structured navigation. It appears as people who have good spatial ability and use visual imagery to acquire and represent information are able to adopt a more functional and optimal behavior when they search for web information. Unlikely, people who have scarce spatial ability and tend to base on verbal strategies to acquire and represent information adopt a confused behaviour which requires to come back in order to reorganize the information and to proceed in web search information.

Another interesting result is that individuals with different cognitive styles show differences in the way in which they explore hypermedia space. We expected that individuals with landmark style tend to explore an hypermedia area smaller compared to whom have a survey style. As expected, we found differences in explored areas’ size, even if not as the way we supposed. Landmarks appear less goal-directed, and they use the mouse continuously during information findings. They seem to have more spread movements, in different position on the screen, in respect of survey users, that have more focalized movements, in few areas of the pages. This could be related to a trial and error strategy as suggested before, in which landmark spatial cognitive style users tend to open links and then to return back to the previous page and open a new link. Whereas, survey spatial cognitive style users tend to focalise on a specific hypermedia space where the useful information are got together.

This finding points to the fact that differences in spatial cognitive styles and as a consequence differences in navigational ability may reflect also in the capability to explore the hypermedia space also when the task does not require to virtual navigate through an environment. This evidence suggests that the spatial mental representation, as well as navigational strategies that people adopt in the everyday life may be significant of other forms of visuo-spatial exploration. Technological innovations provide a great opportunity to study spatial cognition and, in particular, to investigate whether cognitive processes used during paper and pencil tasks may be the same adopted in approaching to the personal computer. In human navigation, it is still matter of debate whether moving in a virtual reality setting may be different or not from moving through a real environment (e.g, Chrastil and Warren, 2012; Morganti, Carassa, & Geminiani, 2007; Waller, 2000, 2005; Hegarty et al., 2006; Nori et al., 2015b; 2015c). To our knowledge the present study is the first attempt to extend the concept of visuo-spatial exploration strategies from real to the hypermedia space trying to combine an innovative software developed ad hoc for quantitative detecting searching strategies on Web navigation and to compare them with individuals’ performance on spatial cognitive style scales that allows to classify individuals in accordance with their navigational proficiency in the real environment.

Limitations and Future works

In future works, it could be interesting to trace back the target of the mouse cursor, for example, a link or a piece of information. Unfortunately, in the present work was not possible because to solve the tasks, participants navigated in an unconstrained hyperspace; there is no other Website opened by every participant besides the Google search engine homepage, and this did not allow us to match the mouse targets of different participants. For this reason, it would be interesting to investigate how people with different cognitive styles search information during a specific task, in a supervised hyperspace, thus allowing a direct comparison of mouse movements. Moreover, it could be interesting to analyse if different spatial cognitive style users adopt different information search strategies, query reformulation and information processing approaches as in Kinley et al ’s model (2014) for analytic/wholistic and verbalizer/imager cognitive style. Finally, more attention should be devoted to eyes movements during Web searching, to understand if people with different spatial cognitive styles have a different way to explore hypermedia space visually.

Taken together our findings confirm the importance of individuals’ cognitive style in Web searching and navigation that should be considered when information are organized into web site to optimize and to make people efficient in personal computer’s use.

Amichai-Hamburger, Y., & Vinitzky, G. (2010). Social network use and personality. Computers in Human Behavior, 26(6), 1289–1295. doi:10.1016/j.chb.2010.03.018.

Antonietti, A. (1998). Computer use questionnaire.

Borlund, P., & Ingwersen, P. (1997). The development of a method for the evaluation of interactive information retrieval systems. Journal of Documentation, 53(3), 225–250. http://dx.doi.org/10.1108/EUM0000000007198.

Calcaterra, A., Antonietti, A., & Underwood, J. (2005). Cognitive style, hypermedia navigation and learning. Computers & Education, 44(4), 441–457. doi:10.1016/j.compedu.2004.04.007.

Goldberg, L. R. (1990). An alternative“ description of personality”: the big-five factor structure. Journal of Personality and Social Psychology, 59(6), 1216 --1229. http://dx.doi.org/10.1037/0022-3514.59.6.1216.

Golledge, R. G. (1999). Human wayfinding and cognitive maps. Wayfinding Behavior: Cognitive Mapping and Other Spatial Processes, 5–45.

Herder, E., & Juvina, I. (2004). Discovery of Individual User Navigation Styles. In Workshop on Individual Differences in Adative Hypermedis at AH 2004 (pp. 40–49). Birkbeck College, University of London.

Hölscher, C., & Strube, G. (2000). Web search behavior of Internet experts and newbies. Computer Networks, 33(1), 337–346. doi:10.1016/S1389-1286(00)00031-1.

Jackson, L. A., Ervin, K. S., Gardner, P. D., & Schmitt, N. (2001). Gender and the Internet: Women communicating and men searching. Sex Roles, 44(5-6), 363–379. doi: 10.1023/A:1010937901821.

Jones, M., Marsden, G., Mohd-Nasir, N., Boone, K., & Buchanan, G. (1999). Improving Web interaction on small displays. Computer Networks, 31(11), 1129–1137. doi:10.1016/S1389-1286(99)00013-4.

Juvina, I., & van Oostendorp, H. (2004). Individual differences and behavioral aspects involved in modeling web navigation. In User-centered interaction paradigms for universal access in the information society (pp. 77–95). Springer. doi: 10.1007/978-3-540-30111-0_7.

Kim, J. (2009). Describing and predicting information-seeking behavior on the Web. Journal of the American Society for Information Science and Technology, 60(4), 679–693. DOI: 10.1002/asi.21035

Kinley, K., & Tjondronegoro, D. W. (2010). The impact of users' cognitive style on their navigational behaviors in web searching. In Proceedings of 15th Australasian Document Computing Symposium (ADCS) (pp. 68–75).

Kinley, K., Tjondronegoro, D., Partridge, H., & Edwards, S. (2014). Modeling users’ web search behavior and their cognitive styles. Journal of the Association for Information Science and Technology, 65(6), 1107–1123. DOI: 10.1002/asi.23053.

Lawton, C. A. (1994). Gender differences in way-finding strategies: Relationship to spatial ability and spatial anxiety. Sex Roles, 30(11-12), 765–779. doi: 10.1007/BF01544230.

Lawton, C. A., & Kallai, J. (2002). Gender differences in wayfinding strategies and anxiety about wayfinding: A cross-cultural comparison. Sex Roles, 47(9-10), 389–401. doi: 10.1023/A:1021668724970.

Lazonder, A. W., Biemans, H. J. A., & Wopereis, I. G. J. H. (2000). Differences between novice and experienced users in searching information on the World Wide Web. Journal of the American Society for Information Science, 51(6), 576–581. Doi: 10.1002/(SICI)1097-4571(2000)51:6<576::AID-ASI9>3.0.CO;2-7.

Lee, C. H. M., Cheng, Y. W., Rai, S., & Depickere, A. (2005). What affect student cognitive style in the development of hypermedia learning system? Computers & Education, 45(1), 1–19. doi:10.1016/j.compedu.2004.04.006.

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(6), 483–496. doi:10.1016/S0959-4752(00)00010-4.

Liu, M., & Reed, W. M. (1995). The relationship between the learning strategies and learning styles in a hypermedia environment. Computers in Human Behavior, 10(4), 419–434. doi:10.1016/0747-5632(94)90038-8.

MacFadden, A., Elias, L., & Saucier, D. (2003). Males and females scan maps similarly, but give directions differently. Brain and Cognition, 53(2), 297–300. doi:10.1016/S0278-2626(03)00130-1.

Nori, R., & Giusberti, F. (2006). Predicting cognitive styles from spatial abilities. The American Journal of Psychology, 119(1), 67–86. doi: 10.2307/20445319.

Nori, R., & Piccardi, L. (2011). Familiartity and spatial cognitive style: How important are they for spatial representation? in J. B. Thomas (Ed.), Spatial memory: Visuospatial processes, cognitive performance and developmental effects (pp. 123--144). New York: NovaPublisher.

Pak, R., Rogers, W. A., & Fisk, A. D. (2006). Spatial ability subfactors and their influences on a computer-based information search task. Human Factors: The Journal of the Human Factors and Ergonomics Society, 48(1), 154–165. doi: 10.1518/001872006776412180.

Palmquist, R. A., & Kim, K.-S. (2000). Cognitive style and on-line database search experience as predictors of Web search performance. Journal of the American Society for Information Science, 51(6), 558–566. DOI: 10.1002/(SICI)1097-4571(2000)51:6<558::AID-ASI7>3.0.CO;2-9.

Papanikolaou, K. A., Grigoriadou, M., Magoulas, G. D., & Kornilakis, H. (2002). Towards new forms of knowledge communication: the adaptive dimension of a web-based learning environment. Computers & Education, 39(4), 333–360. doi:10.1016/S0360-1315(02)00067-2.

Pazzaglia, F., Cornoldi, C., & De Beni, R. (2000). Differenze individuali nella rappresentazione dello spazio e nell’abilità di orientamento: Presentazione di un questionario autovalutativo. Giornale Italiano Di Psicologia, 27(3), 620–627. DOI: 10.1421/310.

Prensky, M. (2001). Digital natives, digital immigrants part 1. On the Horizon, 9(5), 1–6.

Riding, R., & Cheema, I. (1991). Cognitive styles -- an overview and integration. Educational Psychology, 11(3-4), 193–215. http://dx.doi.org/10.1108/10748120110424816.

Riding, R. J. (1997). On the nature of cognitive style. Educational Psychology, 17(1-2), 29–49. DOI:


Ross, C., Orr, E. S., Sisic, M., Arseneault, J. M., Simmering, M. G., & Orr, R. R. (2009). Personality and motivations associated with Facebook use. Computers in Human Behavior, 25(2), 578–586. doi:10.1016/j.chb.2008.12.024.

Schumacher, P., & Morahan-Martin, J. (2001). Gender, Internet and computer attitudes and experiences. Computers in Human Behavior, 17(1), 95–110. doi:10.1016/S0747-5632(00)00032-7.

Shiode, N. (2004). Digitally united?: a case study on the penetration of wireless and ubiquitous information technologies in japan. In Proceedings of the winter international synposium on Information and communication technologies (pp. 1–6).

Siegel, A. W., & White, S. H. (1975). The development of spatial representations of large-scale environments. Advances in Child Development and Behavior, 10, 9–55.

Smith, S. M., & Whitlark, D. B. (2001). Men and women online: What makes them click? Marketing Research, 13(2), 20.

Spink, A., & Jansen, B. J. (2006). Web search: Public searching of the Web (Vol. 6). Springer Science & Business Media.

Stern, E., & Portugali, J. (1999). Environmental cognition and decision making in urban navigation. Wayfinding Behavior: Cognitive Mapping and Other Spatial Processes, 99–118.

Teevan, J., Ramage, D., & Morris, M. R. (2011). # TwitterSearch: a comparison of microblog search and web search. In Proceedings of the fourth ACM international conference on Web search and data mining (pp. 35–44). doi:10.1145/1935826.1935842.

Thüring, M., Hannemann, J., & Haake, J. M. (1995). Hypermedia and cognition: Designing for comprehension. Communications of the ACM, 38(8), 57–66. doi:10.1145/208344.208348.

Weiser, M. (1993). Ubiquitous computing. Computer, (10), 71–72. http://doi.ieeecomputersociety.org/10.1109/2.237456.

Xie, X., Miao, G., Song, R., Wen, J.-R., & Ma, W.-Y. (2005). Efficient browsing of web search results on mobile devices based on block importance model. In Pervasive Computing and Communications, 2005. PerCom 2005. Third IEEE International Conference on (pp. 17–26). doi: 10.1109/PERCOM.2005.16.

Yin, L. R. (2001). Dynamic Learning Patterns: Temporal Characteristics Demonstrated by the Learner. Journal of Educational Multimedia and Hypermedia, 10(3), 273.