Search on video streaming services is significantly different from web search engines and traditional information retrieval systems. Although an internal site search function’s primary purpose is similar to a web search engine’s, the user’s needs differ, which calls for a different design approach regarding the search result. Many previous researchers have investigated the algorithms behind search functions. However, there is a lack of studies related to the user experience of a search's user interface within the video streaming industry. This study investigates how different presentations of search results affect the users to complete a search on web-based video streaming services. The study was done in collaboration with SVT, examining their video streaming service SVT Play. The research aimed to answer three research questions based on the official usability standard ISO 9241–11. A prototype was designed from a state-of-the-art analysis and the findings from three focus group interviews. An A/B test was conducted on the developed prototype, comparing it against the original search function live on SVT Play, including 46 860 test users. A complementary user test with 10 participants was done in addition to the A/B test. The findings from the A/B test did not show any significant difference between the prototypes. It highlighted that users’ search goals are strong enough not to let the layout and design affect their choice, hence the overall click-through rates. However, the user test showed that the new prototype was, in general, easier to use. It also showed that the different prototypes suited different search goals. The participants expressed a wish to have a solution that combines the design elements from the two prototypes. / Sökfunktionen på videostreamingtjänster skiljer sig från webbsökmotorer och traditionella informationssökningssystem. Även om det primära syftet av en sökfunktion på en intern webbplats liknar webbsökmotorns, så skiljer sig användarens behov vilket därav kräver en annan designmetod. Många tidigare forskare har undersökt algoritmerna bakom sökfunktioner. Dock saknas studier relaterade till användarupplevelsen av sökets gränssnitt inom videostreaming industrin. Denna studie undersöker hur olika presentationer av sökresultatet påverkar användarna för att slutföra en sökning på en webbaserad videostreamingtjänst. Studien gjordes i samarbete med SVT och därav utgick undersökningen från deras videostreamingtjänst SVT Play. Uppsatsen syftade till att svara på tre forskningsfrågor utformade efter den officiella användbarhetsstandarden ISO 9241–11. En prototyp togs fram baserad på en toppmodern analys och resultat från fokusgruppsintervjuer. Ett A/B test med 46 860 testanvändare utfördes med syfte att jämföra den framtagna prototypen med den ursprungliga sökfunktionen på SVT Play. Ett kompletterande användartest med tio deltagare gjordes utöver A/B testet. Resultaten från A/B testet visade ingen signifikant skillnad mellan prototyperna. Det framhöll att användarnas sökmål är tillräckligt starka för att layout och design ska påverka deras val och därav den totala klickfrekvensen. Användartestet visade att den nya prototypen generellt var lättare att använda. Studien visade också att de olika prototyperna passade olika sökmål. Deltagarna uttryckte en önskan om en lösning som kombinerar designelement från de två olika prototyperna.

This research is important because it has identified a gap in the existing knowledge base. A term is therefore coined to label a computer-based test mode effect, the so-called Item Randomisation Effect, discussed in detail in this thesis. Item Randomisation Effect is a test mode effect occurring in computer-based testing contexts, especially noticeable in test-takers that may be susceptible to test anxiety. The practise of randomising multiple choice items in computer-based test venues is commonplace, mainly as a deterrent for cheating. Previous research attempted to determine the degree of equivalence across testing modalities of any test. The need was to ensure test-takers in paper-based tests would not have an advantage/disadvantage over test-takers given the same test in a computer-based mode. Such studies have a nomothetic perspective. This research contrasts with those earlier studies in that it has an ideographic perspective because it is concerned with the performance of individuals taking any test in the computer-based modality. This subtle difference in perspective may account for the apparent gap in the existing educational research literature. Evidence of Item Randomisation Effect was found in this study but further research into this test mode effect is necessary. / Dissertation (MEd (Computer-Integrated Education))--University of Pretoria, 2008. / Curriculum Studies / unrestricted