Autumn 2022/ 2023
Example:
Focus on comparing analytical techniques for management questions
• Key focus (research aim): understanding and comparing two (or more) prediction models and apply to employee turnover
• Literature review (4000-5000 words):
• theoretical background of employee turnover, main causes – which variables are required in the model (~1000 words)
• Explain the prediction models, assumptions and interpretation
• Review empirical studies regarding prediction in theory and prediction models
• Prediction methods
• Start with the linear model (or logistics model) as a simple version
• Use at least one advanced model, for example a statistical learning model (a decision tree is not advanced)
• Discuss similarities and differences between both prediction models, in terms of key explanatory variables (à theory) and prediction performance
Focus on understanding a management question using analytics tools
• Key focus (research aim): understanding the triggers (causes) why employees leave a firm
• Literature review (4000-5000 words):
• Explain at least two theoretical approaches in detail and derive reasons why employee leave the firm
• Summarise empirical studies that analyse drivers of employee turnover
• Methodology:
• Explain the prediction model (in this case probably a logistic model or LPM)
• Consider moderation or mediation models to dig into causes
• Discussion:
• Explain why your findings (causes of employee turnover) are in line or different from the literature
• Reflect if your data and model included all relevant drivers of employee turnover
Classical quantitative management dissertation with evidence of coding
• Key focus (research aim): understanding the triggers (causes) why employees leave a firm
• Literature review (4000-5000 words):
• Explain at least two theoretical approaches in detail and derive reasons why employee leave the firm
• Summarise empirical studies that analyse drivers of employee turnover
• Methodology:
• Identify how you can measure the key drivers of employee turnover in your questionnaire (use the measurements in the literature)
• Collect your data
• Explain a simple linear model to predict employee turnover (linear model)
• Discussion:
• Explain why your findings (causes of employee turnover) are in line or different from the literature
• Reflect if your data and model included all relevant drivers of employee turnover
• Research question (aim):
• A precise research question is welcomed but a broader aim is also possible
• Data:
• Secondary data analysis
• Primary data analysis (only classical dissertation)
• Variable mapping – how key variables are measured in the data
• Methodology:
• Advanced (necessary if you want to focus on the method)
• Simple
The data should include the key variables to measure the research
question
• The topic choice is restricted by the available datasets
• Secondary data analysis relies on publicly available datasets
• Firms do not publish their internal data
• For the proposal, you should check the questionnaires (or data descriptions) if your key variables (explanatory and outcome variable) are in the data
• Topics that require `within firm data’ are not always feasible
• Topics can be twisted to fit to data
• using turnover intention instead of actual turnover
• using motivation instead of individual productivity
• Using data from Kaggle is not allowed for dissertations
• The individual assignment in Business Intelligence uses a secondary dataset
• Identifying a dataset is the crucial part
• Example for variable mapping:
• Research question: Does job autonomy increases job satisfaction?
• Dataset EWCS 2015
• Job autonomy: Q61N, four point Likert scale with the question “You can influence decisions that are important for your work”
• Job satisfaction: Q88, four point Likert scale with the question “On the whole, are you very satisfied, satisfied, not very satisfied or not at all satisfied with working conditions in your main paid job?”
• See materials for the Research method unit for details
• Time to explore possible routes for the dissertation:
• Look at different data and ideas
• Motivation and Research aim (1 page)
• Link to literature (1 page)
• Key theory (short description)
• Empirical paper that analysed the question with the same data
• Dataset and key variables (outcome variable, key explanatory variables)
• Identify the dataset
• How you want to measure the key variables in the research aim (variable mapping)
• You need to send the data to your supervisor (except you know that (s)he has it)
• We expect that you comply with the regulations
• structure and marking criteria (expectations) for a dissertation
• reference style
• Learning new analytical techniques (statistics, programming) and theoretical concepts
• A three months dissertation requires
• project management skills
• time management skills
• You are not expected to invent research
• new topics are not expected, you are assessed on your ability to
analyse a dataset and interpret the data
• applying a well-researched question to another dataset or to
subsample of the data or with a conceptual twist is sufficient
• Use empirical papers as a blueprint for your analysis
• Definition of variables
• Theoretical framework
• The second assessment in Business Intelligence can be your starting
point
• Use the material from spring term (R scripts for programming, slides
and additional materials for methods, books, blogs and datacamp) to develop
your skills
You are entitled for supervision, ask your supervisor for feedback
• Four weeks for the literature review (Introduction and Literature
Review)
• Six weeks for the data analysis (Methodology and Analysis)
• Supervisor cut-off (probably) in End of April
• you should have finished your data analysis before the cut-off day
• Supervisor feedback is
most valuable for the proposal and the data analysis
• Write afterwards: discussion, conclusions, revising for consistency,
proof reading
• During a long project, something will go wrong or you underestimated
the required time, hence plan with some buffers