Design and evaluation of surveys and A/B tests

Vaso Dzhinchvelashvili

Master's thesis

NÁVRH A HODNOCENÍ PRŮZKUMŮ A A/B TESTŮ

Abstract:

This thesis reports on comparison of methodologies of designing and evaluation of tests, which include surveys, A/B tests, natural experiments, with a main focus on A/B tests. The aim is to compare different approaches, highlight advantages and limitations based on review of literature, pick the best suitable one and simultaneously test on real enterprise data. The outcome of the research could be used to contribute to a decision making in enterprise. The approaches to randomisation, feature selection are described as well.

Abstract:

This thesis reports on comparison of methodologies of designing and evaluation of tests, which include surveys, A/B tests, natural experiments, with a main focus on A/B tests. The aim is to compare different approaches, highlight advantages and limitations based on review of literature, pick the best suitable one and simultaneously test on real enterprise data. The outcome of the research could be used to contribute to a decision making in enterprise. The approaches to randomisation, feature selection are described as well.

Keywords

A/B test, experiment, propensity score matching

Keywords

A/B test, experiment, propensity score matching

Language used: English

Date on which the thesis was submitted / produced: 29. 6. 2022

Thesis defence

• Date of defence: 25. 8. 2022
• Supervisor: Adam Čabla
• Reader: Martin Potančok

Vysoká škola ekonomická v Praze

Master programme:
Information Systems Management