Assessment Brief 2024/2025

Assignment Information

1. QUESTION/ DESCRIPTION OF ACTIVITY

This is an individual essay (50%).  

The topic of the assignment is:

“Data snooping becomes a considerable concern when exploiting an extensive dataset whose number of variables is larger than the number of observations. This  issue  results  in  false  discoveries,  especially  when  classical  statistical inference is used in which investors repeatedly test the same single hypothesis without adapting a rejection region.’’

Required: Critically discuss the above statement taking into account Multiple Hypothesis frameworks available in the literature.

Student guidelines:

You will need to plan your answer carefully to provide a focused and succinct essay whatever your approach and reasoning ability within the word-limit.  You should also demonstrate your ability to explain and apply the concepts, theories or models and to justify any conclusion you may reach based on evidence provided by all relevant course materials or any other properly referenced source you may choose to use. The assignment answers should be written in an academic and logical manner, not a journalistic style. The assignment refers to the related lecture in Multiple Hypothesis Testing and focuses on data-snooping and the frameworks available to address the relevant statistical bias. You should research MH testing in finance and critically assess the importance of data-snooping test. Evidence of wider reading and critical thinking are strongly encouraged. The word limit is 2000 words. You should not exceed it.

2. ADDITIONAL INFORMATION FOR GROUP ASSIGNMENTS

Not relevant for this assignment

3. ASSESSMENT RUBRIC/ CRITERIA

Overall, successful individual essays should:

•  Provide a clear, well-structured and well-written essay according to the academic expectations of the School and based on the required task.

• Demonstrate the student’s ability to analyse, elaborate and critically discuss the relevant literature and provide meaningful and coherent arguments regarding the MHT framework and its utility and importance in financial applications

• Demonstrate deep knowledge on how MHT evolves around methods such as FWER and  FDR  and  how  their  application  have transformed financial  research  in machine learning.

A holistic rubric provides a list of assessment criteria together with broad description of the characteristics that would be expected for each level of performance.