£575.00 Excluding VAT
This Course has passed
If someone remarked to you that “the weather is very nice today” or “I didn’t like that person”, it is unlikely that they would have made such statements based on a single variable. It is more likely that a combination of variables were evaluated to arrive at these statements. When we analysis datasets with multiple variables, we are undertaking Multivariate Statistical Analysis.
In this highly practical and interactive course based on seat level data for UK general elections between 2010 & 2019, run by the most accurate forecaster of the 2019 General Election, you will learn how to use and interpret the two types of Multivariate Analysis:-
- Analysis of Correlations between Multiple Variables – Informally known as reducing the dimensionality of your data.
- Analysis of Distance between Many Objects – Informally known as mapping, clustering or segmentation of your data.
Who should attend this course?
This course will be of particular interest to political researchers, analysts & forecasters who want to understand how the British electorate voting behaviour changed between 2010 & 2019 based on constituency level data. Some of the questions that will be considered in the course are –
- Can the Conservatives consolidate the gains they made from Labour’s Red Wall?
- Where should Labour focus their efforts to recover their losses and win back power?
- Which seats should be the long term targets for the Liberal Democrats?
- Will the EU referendum results of 2016 continue to be a key dynamic in the future?
- What differences can we see between the various regions and are these likely to continue in the future?
The techniques taught in this course can be applied to many other datasets outside of politics so do come along even if political data is not your main focus.
Whilst COVID19 is still affecting daily life, this course will be run as an online course using Zoom over 2 days, Thursday 24th June and Friday 25th June. Each day will start at 0930 and finish at 1700.
The course fee is £575 +VAT per person.
What datasets will be used in the course?
The course will use methods of multivariate analysis to understand the General Election results of 2010, 2015, 2017 & 2019 in the UK. You will receive a spreadsheet listing all 632 parliamentary constituencies in Great Britain along with the vote shares for each party, the EU referendum vote of 2016 and demographic and other related data for each seat, mostly taken from the 2011 census. Since the boundaries of the seats did not change during this decade, this means we have a very clean dataset to understand in some depth the extraordinary realignment of British politics between the 2010 & 2019 elections.
Please note that the data for Northern Ireland I have is incomplete and hence this data will not be included in the course.
What you will learn?
Day 1 – You will learn how to prepare your data for any multivariate analysis and apply the main tools of Q-Analysis based on distances between objects.
- Why multivariate analysis is different from univariate analysis
- What are best ways to prepare a dataset for multivariate analysis.
- The importance of distance metrics and how to choose the right one for Q-Analysis.
- How to use and interpret MDS Multi Dimensional Scaling.
- How to use and interpret AHC Agglomerative Hierarchal Clustering.
- How to use other methods of clustering such as K-Means Clustering and Manual methods.
Day 2 – You will learn how to apply the main tools of R-Analysis which are based on correlation between variables.
- The difference between correlation & covariance matrices and which one to use.
- How to use and interpret PCA Principal Component Analysis.
- The similarities and differences between PCA & Factor Analysis.
- How to use and interpret MCA Multiple Correspondence Analysis.
- A brief overview of other methods of Multivariate Analysis such as DA, GPA, CCA, etc.
For an example of the kind of analysis you will be able to perform after this course, take a look at this blog post “Keir Starmer’s train to Downing Street“.
Throughout the course, there will be ample opportunities for you to practice these techniques on the datasets described earlier, ask questions and to discuss the implications of your results. After the course, you will receive a spreadsheet explaining all the answers and demonstrating the relevant calculations.
In order to get the most out of this course, participants will require:
- A laptop with Microsoft Excel installed. A spreadsheet with the data to be analysed during the course will be emailed to you before the course.
- Be comfortable with the main operations of Microsoft Excel such as entering data, producing line, bar & XY scatter plots and be able to use the data manipulation functions of Excel such as IF, VLOOKUP, MATCH, OFFSET etc.
- A firm understanding of basic statistical concepts such as expectation, variance, distribution, correlation, probability & risk.
- A statistical software package capable to undertaking Multivariate Analysis (see next section)
Multivariate Analysis requires the use of specialist statistical software. You are free to use your own statistical software package but please make sure that it can perform the following operations.
- PCA Principal Components Analysis. You may find this under the Factor Analysis option in some packages.
- MCA Multiple Correspondence Analysis.
- Cluster Analysis, specifically AHC (Agglomerative Hierarchal Clustering) and K-Means.
- MDS Multi-Dimensional Scaling
Your course instructor will be using XLSTAT. This is an add-in to Microsoft Excel and has an excellent Multivariate Analysis capability since it was originally developed for the Sensory world (of food & drink tasting) which is inherently multivariate. You can download a free 14-day trial copy of this add-in from the publisher AddinSoft.
Your course instructor
Nigel Marriott is an Independent Statistician and fellow of the Royal Statistical Society. He has been running basic statistics courses aimed at non-statisticians for 20 years. Feedback has consistently noted that Nigel has a rare ability to explain what looks like complicated statistics in a simple manner that makes it easy for you to understand and apply what you are learning.
Find out what previous course attendees think of Nigel.
He was denoted the most accurate forecaster for the 2019 General Election by the Institute of Government Studies. Find out more here!
Nigel also blogs extensively on topical statistical issues including elections and multivariate analysis and you can find all his blogs here.
Other Training Courses in Statistics
We offer a wide range of statistical training courses covering numerous topics. We provide in-house training as well as public courses like this one and you can find out more here.
If you would like to receive notifications of future training courses, you can subscribe to our Newsletter by following this link and ticking the Statistics Training category and any other categories of interest. You will be able to unsubscribe at any time.