1 Project Assessment and Organization Your understanding of quantitative methods in this module is assessed in three steps: i) in a mid-term exam worth 15% of your mark, ii) in a group-work project worth 25% of your mark and iii) in the final exam worth 60% of your mark. This text sets out the main aspects of the group-work element. This element will cover a question central to the study of contemporary management: • Whether firms should let their employees telecommute, namely do some or all of their work from home. This question has recently been explored by colleagues who have shared their data with us. We want you to analyze their data and report on your conclusions regarding an aspect of the question. We will assign you to work in a group of three students. Assignment will mean you do not influence whom you are working with, just as in a professional work environment in which you act as an employee. To minimize problems, which can arise in such a setting, we request that you fill in diaries when you meet every week. These diaries will record what you have discussed and how your project has progressed. Deliverables At the end of term you are expected to submit two pieces of work: 1. a 5 minute presentation to be held in your seminar class; 2. a report of 1500 words. Both pieces of work will be based on analysis of a data set, which we will make available on QM Plus. You will be expected to report on the following aspects of the data: 1 i) Descriptive statistics for the variables in the data; ii) Graphical analysis of the main correlations in the data; iii) Your main hypotheses and output tests of these hypotheses; iv) Regression analysis testing the same hypotheses; v) Your recommendations and conclusions from this analysis. We will go through each of these steps of analysis with other data in the seminars this term. Your task will be to repeat the work we do in the seminars using the data for your projects. Your first task is to choose which research question you will work on as a group. There is data for six specific questions around the main theme of telecommuting. Please record your choice of question on the courses QM Plus pages. In each seminar group we will allow maximally two teams to choose the same question. Questions will be assigned to you on a first come first serve basis. 2 Telecommuting This study focuses on the effect of allowing workers who previously worked at the office to telecommute for four days a week. The company running this experiment wanted to know whether their employees would work less hard at home or whether the fact that they did not have to commute and could work in a more personal space without noise would improve their concentration and effort levels. The academics running the experiment measured how many breaks employees took, how many calls they took per day, whether employees working from home were more likely to quit, were more or less satisfied and how working from home affected promotion. The company is a large travel agency and employees telecommuting were either taking or making bookings for trips or hotels or they were correcting orders. There are six topics under this theme one of which you can will be assigned work on. For each topic there is a separate data set. The data sets describe different numbers of employees, some focus only on the employees taking calls, some on all types of employees. All data sets contain information on the employees in question, e.g. their sex, their age, whether they have children, the length of their commute. The data also contain a variable “experiment treatement”, which indicates whether a worker was working from home in the week in which the data were recorded. Next we describe the questions you can study with each data set: 2 1. Number of breaks taken – data set “calls”. This data set contains a measure of the number of minutes spent taking calls, “logcalllength”. The question to analyze is whether employees working from home worked longer than employees working in the office. 2. Number of calls – data set “calls”. This data contains a measure of the number of calls each employee took, “phonecall”. The question to analyze is whether the employees working from home took more calls than employees working in the office. 3. Attrition (how many workers leave) – data set “attrition”. This data contains an indicator whether the employee left the company during the experiment, “quitjob”. The question to analyze is whether the employees working from home were more or less likely to quit their jobs than the employees working from the office. 4. Work Satisfaction – data set “satisfaction”. This data contains three measures of satisfaction from surveys of the employees undertaken before and during the experiment. The question to analyze is whether employees working from home were more or less satisfied with their work than enployees working in the office. 5. Exhaustion and Attitude – data set “exhaustion”. This data contains measures of exhaustion and attitudes from weekly surveys of employees undertaken before and during the experiment. The question to analyze is whether employees working from home were more or less exhausted with their work than enployees working in the office. 6. Promotion – data set “promotion”. This data contains an indicator of whether the employee was promoted after the experiment, “promote switch”. The question to analyze is whether employees working from home were more or less likely to be promoted than employees working from the office. 7. Performance – data set “performance”. This data contains a measure of how the employee performed during the experiment, “perform1”. This is a score of performance that is comparable across different groups of employees. The question to analyze is whether employees working from home performed better on this measure than employees working from the office. 2.1 Description of variables – personid Identification number for each individual. – year week Year and week of the measurement. – surveyno Round of the survey in the satisfaction . – expgroup This is a dummy variable that equals 1 if an individual belongs to the treatment group defined by having an even-numbered birthday. 3 – treatment This is a dummy variable that equals 1 for the experimental period December 6 2010 to August 14 2011. – experiment treatment This is a dummy variable that equals 1 if an individual belongs to the treatment group and the measurement was taken in the experimental period. In other words this variable identifies measurements taken at home. – logcalllength Minutes on the phone – a measure of high-frequency labor supply or a measure of number of breaks taken relative to other employees. – logcallpersec Phone calls answered per second on the phone, a measure of productivity. This is negative as the original data will all have been below 1 – no employee can finish one or more calls in a second. – phonecall – perform1 Performance score normalized to mean = 0 and standard deviation = 1 based on pre-experiment performance for each task. As different employees had different tasks normalisation is used to make scores across tasks comparable. – perform10 Score for the period before experiment. – perform10 Score for the period during experiment. – gross wages A measure of performance that captures bonus payments. – promote switch This is a dummy variable that equals 1 if the individual was promoted after the experiment was over. – quitjob This is a dummy variable that equals 1 if the individual quit their job. – commute Length of commute in minutes. – tenure Months worked at the company.