Assignment 1
General instructions
Academic integrity
This is an individual assignment. When submitting it on Blackboard (Turnitin), you acknowledge that you have answered the questions individually and in accordance with University guidance on Academic Integrity. You may use notes and online resources when answering these questions, but you must not get help from any person, be it in person or online. For clarification questions, please contact the instructors. Plagiarism will be detected and penalised. The assignment is due via Turnitin in pdf format and will be thoroughly examined for issues of academic integrity. You must use your own words in your descriptions and explanations. Please submit your assignment with your student number, but without your name, to allow for anonymous marking. Turnitin will check for similarity, so please submit a computer-written file in pdf form.
Deadline
You can submit this assignment until Friday 1st of March at 23:59. Late submissions will be penalised following University guidelines.
Task 1
A researcher is interested in whether there are gender differences in giving in a dictator game, that is, whether women are more generous than men. She conducts an experiment with 120 participants, 60 females and 60 males, who are all students at the University where the researcher works. In the experiment participants are given £10 and are asked to determine the division of this amount between himself/herself and an anonymous recipient.
##Data
The data collected by the experimenter is contained in the file “Gender.csv” that you will need to use in this exercise. There are 3 columns in this dataset: the first contains the ID number of the participant, the second their gender, and the third the amount they donated.
Questions
- Question 1.1 Formulate the null and alternative hypotheses that the experimenter wants to test. [5 points]
- Question 1.2 Power analysis: What is the power of a z-test to detect gender differences in this experiment, if the significance level is set at 5% and the standardised minimum detectable effect size (δ over the standard deviation of the outcome variable) is equal to 1/3? What do you conclude? Why is it important to consider the power of the experimental design? [20 points]
- Question 1.3 Using “Gender.csv” from Eckel and Grossman (1998) and the statistical software of your choice or by hand, carry out a t-test of the hypothesis that women’s and men’s donations are equal. What do you conclude? [5 points]
- Question 1.4 Using “Gender.csv” and the statistical software of your choice or by hand, carry out a test of the hypothesis that the probability of a specific donation being made is independent of the dictator’s gender. What do you conclude? [5 points] * Question 1.5 Discuss the internal and external validity of this experiment. [15 points]
Data for Tasks
Task 2
This exercise is based on the on a research paper that is available here: /Assignment Reading 1
Questions
- Question 2.1 This RCT is meant to test if a program “t hat” provides a set of 6 services/activities to the very poor is sufficient to help them to start a productive and durable self-employment activity. How could the authors have tested if each of the program services/activities is individually necessary? [20 points] [max 100 words]
- Question 2.2 How did the authors of this RCT measured the outcomes of interest? [10 points] [max 100 words]
- Question 2.3 In three countries the authors randomized half of the villages to the treatment group and half to the control group. Why did they do this randomization? Explain. [20 points] [max 100 words]