Browsing by Author "Luostarinen, Sini"

Now showing 1 - 2 of 2

Challenges of implementing learning analytics in the context of blended university courses: a reflexive case study of a course completion prediction modelling project
(2021) Luostarinen, Sini
School of Business | Master's thesis
As during the last decade, the use of learning management systems (LMS’s) as course platforms became the norm in courses of higher education, and vast quantities of learning behavior data was thus collected, also the practices of educational data mining and learning analytics began growing significant popularity in the science community. For instance, much research on predicting course outcomes, such as course dropouts and course grades were conducted, in both; the online and blended learning contexts. However, research on the challenges which a learning analyst faces while implementing a learning analytics project in a higher education institution, were nearly non-existent. When working with higher education institutions which haven’t yet developed many internal processes for supporting the implementation of learning analytics, learning analysts may face challenges with data access permission and data quality, for instance. Thereby, their projects may suffer from schedule delays or deteriorated results. With a reflexive case study conducted for Aalto University in Finland, this research identifies the main challenges learning analysts face while implementing each stage of the project life cycle. The case project used in this reflexive study is a course completion prediction modelling project that utilizes learning data of marketing students. The challenges identified during the case project occurred only in the data collection and pre-processing stages. First of all, data privacy issues initiated a legal process which due to a heavy work load, slightly delayed the schedule of acquiring permission to accessing student data. Secondly, hand-picking data manually from two databases and integrating that data from two data systems lead to a missing data problem. The data quality also proved to be rather poor, due to unstandardized and partially nonexistent data collection in the LMS. Finally, most of the modelling data had to be eliminated due to incomparability and inoperability challenges, and the prediction model was conducted successfully only for one marketing course. Based on the experienced challenges, this research introduces a set of propositions for managers of higher education institutions, which help overcome these challenges by improving the institution’s internal processes and developing new ones. The propositions are: 1) Create a standardized process for applying for data access; 2) Adopt a learning analytics information system to ease data integration; 3) Harmonize course and course-site structures to enable automated variable construction; and 4) Organize courses in a way which supports data generation. Considering these improvement propositions, this research can help higher education institution managers both improve the life cycle times of learning analytics projects, and the quality of the project results, while simultaneously enhancing the learning of their students.
Impacts of celebrity endorser scandals to brands
(2018) Luostarinen, Sini
School of Business | Bachelor's thesis
This literature review defines the various influencing factors of celebrity endorser scandal outcomes recognized in research today, and studies scandal impacts on company-, service- and product brands endorsed by celebrities, through these influencing factors. Moreover, connections between the different influencing factors and their impacts are examined and inferred in order to create as comprehensive, full picture of the different influencing factors in the equation as possible. To entirely understand the causations in celebrity endorsement scandal situations, the theories of celebrity endorsement are first introduced. The literature review was conducted by examining, analysing and combining existing research on the areas of celebrity endorsement and celebrity endorsement scandals. Seven established influencing factors of celebrity endorsement are reviewed; the meaning transfer process, attractiveness of the celebrity, credibility of the celebrity, congruence, celebrity’s level of blame, consumers’ previous attitudes and type of scandal. Although all of these factors are found important to consider when estimating scandal consequences, celebrity’s level of blame is found to be the most valuable factor in terms of determining the direction and magnitude of the scandal consequences for the brand, and moreover profoundly connected with the impacts of all other factors of influence.