Introduction to Survey Design A
Choosing the Survey Participants and Collecting the Data
Within this workshop we cover: how to minimise the differences between the target population (the members of a population that we ideally want to collect data on) and the sampling frame (those members that we actually can collect data on), survey sampling (the random selection of members from the sampling frame to include within a particular survey), data collection methods (how to choose between methods such as telephone, mail, and face-to-face interviews for a particular survey), and interview techniques (how to structure an interview to minimise potential bias introduced by the interviewer).
This workshop is intended for audience members who are new to the methods of survey research, and who desire a solid foundation in these methods. A simple knowledge of statistics is assumed (where audience members should be comfortable discussing terms such as mean and variance), and a basic understanding of Microsoft Excel would be helpful for one computer exercise.
Each session comprises a lecture and a practical as per the program format below.
Session 1: Who do we ideally want in our survey and who can we have in practice?
In an Australian census questionnaires are sent to every household in Australia, and we know that there are portions of the Australian population not covered in the census (e.g. those people in hospital or who are travelling for business or vacation). In this session we explore the difference between the ideal target population and the population we can survey in practice, and we describe the biases that can occur due to differences between these populations.
Session 2: Choosing a representative subset from a list of possible survey participants
When conducting a survey we generally only survey a small subset of people chosen from the list of all possible survey participants. This is done in order to achieve a balance between the time and cost involved in conducting the survey, and the desired accuracy of survey results. In this session we will describe this survey sampling process and introduce two standard sampling procedures – simple random sampling and cluster sampling.
Session 3: Methods of data collection
There are many options for collecting data for a given survey including face-to-face interviews, telephone surveys, and computer-assisted surveys. These data collection methods differ in a number of ways including the amount of time and money required, the proportion of participants that complete the survey, and the accuracy of completed surveys. In this session we compare and contrast the different methods for data collection.
Session 4: Interview techniques, and accounting for bias
Surveys generally involve some level of input from the interviewer, where there is always the potential that the interviewer might influence the choice of responses provided by the study participant. The reasons for this influence can include a reduced reporting of behaviour and opinions that might not be deemed socially acceptable by the interviewer, and can further be influenced by factors such as the age, gender or race of the interviewer. In this session we introduce many of the reasons for bias during a survey interview, and methods for minimising these types of bias.
This workshop uses a combination of three teaching styles:
Lectures and classroom discussions
Small group discussions
During the lecture sessions the theory of statistics will be presented, and will be discussed in an interactive manner with the class.
Small Group Discussions
During one of the practicals in this workshop, we will read through a number of application papers from a range of fields (including medicine, education, business, and environmental sciences). We will explore what research question is being asked in the paper, the choice of statistical methods used, and an explanation of the results obtained and their interpretation.
Each workshop will involve the use of laptop computers. For these sessions participants will be asked to bring their own laptops and will be able to choose which statistical software they would like to use during the workshop. For this workshop, participants will be able to choose which package (R or SPSS) they would like to use during individual hands-on exercises throughout the workshop.
Introduction to Statistics – can choose between SPSS and R
Introduction to Regression, Longitudinal Data Analysis – can choose between Stata, SPSS and R
Please note that a copy of R will be given to all participants at the start of the workshop, if participants would like to use one of the other software packages then it will be the responsibility of the participant to ensure that they have that software package available on their laptop.
The workshop will adhere to the following format. Please note that both teas and lunch are catered on both days, so please be sure to include dietary requirements on your registration form.
8:30 - 9:00 Registration
9:00 - 10:30 Lecture 1
10:30 - 11:00 Morning Tea
11:00 - 12:30 Practical 1
12:30 - 1:30 Lunch
1:30 - 3:00 Lecture 2
3:00 - 3:30 Afternoon Tea
3:30 - 5:00 Practical 2
9:00 - 10:30 Lecture 3
10:30 - 11:00 Morning Tea
11:00 - 12:30 Practical 3
12:30 - 1:30 Lunch
1:30 - 3:00 Lecture 4
3:00 - 3:30 Afternoon Tea
3:30 - 5:00 Practical 4
Dr Mark Griffin is the Founding Director of Insight Research Services Associated (www.insightrsa.com), and holds academic appointments at the University of Queensland and the University of Sydney. Mark is the Chair of the IIBA Business Analytics Special Interest Group and the IIBA Asia-Pacific Regional Director. Mark also serves on the Executive Committee for the Statistical Society of Australia, and is the Chair of their Section for Business Analytics. Mark has previously taught over 80 two-day workshops and 10 five-day workshops in the fields of Business Analytics and Statistics. Major analytics projects that Mark is or has been involved in include:
Mark leads a research group at the University of Queensland conducting analysis of incident reports collated by the Queensland Ambulance Service. The QAS visits approximately 700,000 incidents per year where QAS staff complete a report detailing each incident. This project uses R for text analytics, market segmentation, and spatial mapping (GIS) (2017 to present).
Mark is leading a research group at the University of Queensland that are creating an online sample size calculator in R. This software will be used by managers of medical trials who wish to know how many patients to enrol in their trials. This work is being conducted in partnership with research collaborators at Harvard University. This project uses R for developing a web interface and for the mathematical equations involved (2017 to present).
Mark has developed software in R for SeqWater (where SeqWater monitors the water quality of all 28 water reservoirs in South-East Queensland). This project uses R for developing a web interface and for statistical analysis using time-series data (2017).
Mark led a project team evaluating the delivery of the Positive Parenting Program for the Queensland Department of Communities, Child Safety and Disability Services. This included the collection and analysis of data from 140,000 parents and 1000 practitioners (psychologists) involved in the program. This project used R for statistical analysis and data visualization (2016-2017).