It has been suggested that readers spend a great deal of time skim reading on the Web and that this type of reading can affect comprehension of text. Across 2 experiments, we utilised eye tracking methodology to explore how hyperlinks and navigating web pages affect reading behaviour. In Experiment 1, participants read static web pages either for comprehension or whilst skim reading, while in Experiment 2, participants additionally read through a navigable Web environment. Embedded target words were either hyperlinks or not and were either high-frequency or low-frequency words. Results from Experiment 1 showed readers only fully lexically process linked words when skim reading, as was evidenced by a frequency effect that was absent for the unlinked words. They did fully lexically process both linked and unlinked words when reading for comprehension. In Experiment 2, which allowed for navigating, readers only fully lexically processed linked words compared to unlinked words, regardless of whether they were skim reading or reading for comprehension. We suggest that readers engage in an efficient reading strategy where they attempt to minimise comprehension loss while maintaining a high reading speed. Readers use hyperlinks as markers to suggest important information and use them to navigate through the text in an efficient and effective way. The task of reading on the Web causes readers to engage with the text in a markedly different way from typical reading experiments. The centrality of the Web for scientific research and economic activity has not been matched by our understanding of its complex relationship with the embedding society. In part this is because of its Protean nature and ubiquity. It exists at a variety of scales, from engineering protocols to websites, small communities to giant e-government and e-commerce systems. It is engineered technology, and a network of overlapping social networks.Hence the Web's study is legitimate from many disciplinary perspectives. To engage with it as a first-order object requires an interdisciplinary overview, grounded by an understanding of its engineering principles, that currently few researchers can achieve. The aim of this Doctoral Training Centre (DTC) proposal is to create a cohort of researchers that (a) have appropriate research skills, (b) have a breadth of understanding of multidisciplinary approaches to the Web, and (c) are a coherent community.Such a cohort will enhance our understanding of the Web in a number of ways. They will carry out deep and committed research, which they will be able to situate in a wider context. The number of researchers will create a critical mass, able to promote Web Science in the academic community and help make it a hub for research investment over the medium term.The DTC aims to produce PhDs to fill the need, but who are also aware of each other's work and the work of relevant fields. This community awareness will be fostered in a number of ways. Students will: (i) begin with an MSc to equip them for the technical methods required for Web Science; (ii) have a base in the School of Electronics and Computer Science and the Web Science Research Initiative (WSRI) wherever their main location; (iii) share a core introduction to relevant disciplines, and to the protocols that make up the Web; (iv) be partly assessed on interdisciplinary group projects; and (v) attend joint research seminars. In this way the DTC will add value compared to a stream of isolated DTAs.The DTC will avoid introspection, and will be outward-looking. The Web is an international phenomenon, and WSRI is affiliated with other Web Science labs in a global network. All DTC funded students will be offered the opportunity to spend up to 3 months studying at one of these labs, or in an industrial placement. Furthermore, WSRI will host international scholars of all levels from other labs, and students will be able to attend their seminars and discuss their work.The creation of a cohort of researchers will be beneficial for the Web and our interactions with it. The continuing health of the Web is clearly a matter of interest for everyone, and this DTC will help develop a set of researchers able to understand the Web at varying scales, from varying points of view, and with a clear perception of the international dimension.A PhD in Web Science will be a valuable asset in many industries operating in the digital economy, including e-health, the media, finance or e-defence, where quality of service provided depends both on technological developments and the integration of technology into a social context. Issues such as security, privacy and collective intelligence matter just as much as methods of inference or the structure of the Web. The breadth of understanding that Web Science provides will also be vital in the software and hardware development companies which not only add value to the British economy but also support the Web itself.Finally, the Web is an important tool for government, in terms of communication and coordination within itself, and of delivery of services and supply of information to citizens. Once more, Web Science would enable an understanding of the technology and the social context for the next generation of government officials, allowing them (a) to apply current Web technology to existing problems, and (b) to engage with the future development of the Web in socially beneficial ways.

Across 2 experiments, we utilised eye tracking methodology to explore how hyperlinks and navigating webpages affect reading behaviour. Experiment One consisted of forty edited Wikipedia articles taken from Fitzsimmons et al. (2019, Experiment Three). One-hundred and sixty target words were embedded in sentences (one target word per sentence) and four sentences were inserted into each Wikipedia article. The rest of the text was edited from Wikipedia articles and all words that were links in the original articles were retained for the experimental text. In total there were 8 conditions in a 2 (Task Type: Comprehension, Skimming) x 2 (Word Type: Linked, Unlinked) x 2 (Word Frequency: High, Low) within participants design. Experiment Two was similar in design to Experiment One. The forty wiki pages used in Experiment One were insufficient to allow for a realistic Web environment with full clicking and navigable functions. As such, the stimuli for Experiment Two consisted of 843 edited new Wikipedia articles with experimental sentences inserted into the existing text. The Wikipedia articles were two to twelve lines long (see Figure 5). Participants could follow any hyperlinks they wished to click on and because of this environment, the number of target words observed by each participant varied dependent on the pages they choose to view. Eye movements were measured with an SR-Research Eyelink 1000 eye tracker operating at 1000 Hz. For experiment 2, For more information, see the Methods file.