Deliveroo (www.deliveroo.com.au), the online food delivery company, wants to understand the causal impact of its website design on the probability of a consumer completing a transaction on the website. Specifically, they want to know whether the presence of an AI chatbot improves customers' probability of completing the transaction (remember, not all customers visiting the site end up ordering food). The chatbot service facilitates user experience on the website. The AI chatbot appears at the bottom of the screen when the customer appears on the website, greets them and helps them navigate through the website. It answers frequently asked questions (answers can be pre-programmed) and recommends related products. It aims to provide a personalized experience for every customer and relieves the burden on the company's call centre making it a cost-effective investment. However, some customers may find it intrusive or annoying. Therefore, the company wants to conduct an experiment with 1000 individuals to test whether it improves sales. You are a member of a researcher team investigates whether the probability of purchase for a user (whether the user completes the food order by paying at the checkout page) is influenced by the AI chatbot (whether the AI chat was enabled when the enter the website). Your task is to design this experiment.

1. Who are the subjects in this study? Who will be your respondents?

2. What is the dependent variable in this study? How are you going to measure it?

3. What is the treatment condition(s) - i.e. what is your independent variable? How are you planning to operationalise it?

4. What experimental research design will be used? Describe the design and represent in mathematical symbols.

5. What confounding/extraneous variables should be controlled?

6. How will the subjects be assigned to conditions/groups? How many subjects will be in each group?

7. Briefly comment on the internal and external validity of the design used.