Analysis
Please note that I am currently offering any of my Face to Face Courses as an online course, delivered through Microsoft Teams (or similar tool of your choice). The content will remain the same, although the way the material is presented, and how activities are completed, may change.
Principles of Analysis: Descriptive and inferential Statistics
Aims:
To introduce descriptive statistics including levels of measurement, central tendency and dispersion
To introduce concepts of rates (Prevalence and Incidence)
To discuss data presentation methods
To give a basic introduction to inferential statistics - statistical tests for 2 groups, P values and confidence limits
To discuss the importance of data quality
To discuss the principles of statistical process control
Content:
This session gives a high level overview of analytical methods. Unlike other sessions there is no group activity for participants. Instead the session is delivered in a discursive manner, and aims to give a taster of some of the more detailed sessions on offer.
Duration:
2 hours
Pre-requisites
Basic Numeracy
Introduction to Descriptive Statistics
Aims:
To present the key concepts of data measurement
To introduce summary statistics and methods of calculation
To introduce the concepts of prevalence and incidence
To outline measures of central tendency and dispersion
To introduce the 'bell shaped curve' (the normal distribution)
To investigate methods of data presentation
Content:
Starting from first principles, we will discuss basic methods for summarising data. The session aims to introduce the concepts of central tendency and dispersion in a participative manner, enabling students to calculate such summary statistics themselves. The normal distribution will be introduced, alongside concepts such as skewness and kurtosis. Methods of presenting data will also be introduced, with students gaining insight into the appropriateness(or not) of each technique.
Duration:
3 hours
Pre-requisites
Basic Numeracy
Introduction to Inferential Statistics
Aims:
To introduce the concept of sampling and standard error
To relate the concept of standard error to that of confidence limits
To outline the principles behind formulating and testing statistical hypotheses.
To explore correlations and the principles of association and causation
Content:
Building on the previous introduction to descriptive statistics sessions, participants will be introduced to more detailed statistical methods, and the concepts behind them. Sampling issues will be addressed, in particular how they relate to the underlying population. Students will be given the opportunity to explore these relationships and develop a more detailed understanding of statistical confidence. Statistical hypotheses will be discussed, including here relationship to one-tailed and two-tailed statistical tests. Finally, correlations will be discussed alongside the principles of association and causation. Real data will be used to demonstrate the role of confounding within such associations.
Duration:
3 hours
Prerequisites:
Basic numeracy
Completion of “Introduction to Descriptive Statistics” session or similar understanding
(Understanding) Statistics and Data Insight
Aims:
To assess a research paper for statistical validity
To understand how we currently use data, how it is used by the media and how it can be misrepresented and misinterpreted
Looking at outcomes and how they are measured including Nominal, ordinal and interval measures
Measures of central tendency (mean, median, mode)
Measures of dispersion (Range, IQR, Standard Deviation)
Percentages, rates and ratios (Prevalence and Incidence)
Data distributions, and the importance of normality
Presentation rules including Chart types and their relationship to data types
The two major study designs for two groups, Independent measures and Repeated measures
Identifying whether the correct statistical test has been used based on the stated outcome
Interpreting the test results (the role of chance) including P values and Confidence Intervals
Understanding why the study result isn’t the most important statistic, including te role of power, and the importance of the outcome
Understanding the advantages of stratification of data
Understanding standardisation of data - when is it used, and the difference between direct and indirect standardisation
Understanding the advantages of looking at time series data, including run charts, moving averages and other smoothing methods
Understanding correlations and cause and effect, using scatterplots and the line of best fit (and its relation to simple forecasting– simple forecasting
Content:
This session differs from other statistics courses on offer as it has few exercise, and is instead content rich. Consequently a lot of content is covered in a relatively short time. A research paper will be sent out before the sessions that participant will be asked to read and attempt to interpret the study results. The morning and afternoon sessions will not only introduce participants to data and statistical concepts but will enable them to interpret the results of the paper and draw meaningful conclusions. The morning session will introduce participants to data and how it is commonly used / misused. We will look at how data is measured, and how it can be summarised. We will look at how data is distributed and why this is important to understanding statistical testing. Data presentation methods will also be looked at, describing the correct way of presenting different types of data. Finally the two key study designs will be introduced. The afternoon session will introduce the most common statistical tests related to the study designs and data types we have discussed. We will discuss how the role of chance affects results and how to interpret them. We will then look at other ways to present and interpret data, including stratification and standardisation of data, time series data and correlation of variables. Lastly we will revisit the results tables from our research paper and use what we have learned to interpret the results.
Duration:
1 Day
Pre-requisites
Basic Numeracy
Analysing Data: Statistical tests for 2 groups
Aims:
To understand the importance of organising and presenting data
To understand how the achieved sample size may affect results
To apply statistical principals to investigation of data
To formulate a statistical hypotheses based on the research aim
To apply relevant statistical tests to the data
Content:
This session builds on the theory introduced in the previous statistics sessions and attempts to set this in a practical context. In particular, through group work and participative learning, we aim to demonstrate how the choice of analysis links both to the underlying question, via statistical hypotheses, and choice of presentation. The session will highlight some of the common pitfalls and errors made in data analysis, and how data can be manipulated to give misleading conclusions.
Duration:
3 hours
Pre-requisites
Basic Numeracy
Completion of "Introduction to Descriptive Statistics" session, or similar understanding
Completion of "Introduction to Inferential Statistics" session, or similar understanding
Key Statistical Concepts 1 (1 Day)
Aims:
To present the key concepts of data measurement
To introduce summary statistics and methods of calculation
To introduce the concepts of prevalence and incidence
To outline measures of central tendency and dispersion
To introduce the concept of 'degrees of freedom'
To understand how data is distributed and introduce the 'bell shaped curve' (the normal distribution)
To investigate methods of data presentation
To understand the principles behind the Central Limit Theorem
To introduce the concept of sampling and standard error
To relate the concept of standard error to that of confidence limits
To understand statistical confidence (p values and confidence limits), and how to interpret
To understand and construct one sided and two sided statistical hypotheses
To outline the principles behind formulating and testing statistical hypotheses.
To explore correlations and the principles of association and causation including Pearson, Kendall’s tau and Spearman
Content:
This session will discuss some of the key statistical concepts associated with descriptive and inferential statistics. It will start from first principles and explore the different types of data that we collect, and how we can describe it usefully for ourselves and others. It will explore how we can use our data to infer things about larger populations and explore the variability in our own data. Concepts of statistical confidence and certainty will be described.
Duration:
1 Day
Prerequisites:
Basic Numeracy
PC Skills – Keyboard and mouse skills
Key Statistical Concepts 2 (1 Day)
Aims:
Formulating analytical questions and corresponding hypotheses
data considerations, including data quality
analytical design (independent vs related samples),
statistical power.
tests for two independent groups (chi-squared, independent samples t-test, Wilcoxon signed ranks),
tests for two related groups (McNemar, Paired samples t-test, Mann Whitney U),
homogeneity of variance,
tests for multiple groups (e.g one way and two way Anova, Kruskal-Wallis etc)
Content:
This session will build on the principles discussed in Key Statistical Concepts 1. Statistical testing will be discussed, and the concept of independent samples and repeated measures will be discussed. The importance of the statistical hypothesis will be emphasised, and how it relates to one tailed and two tailed significance.
Duration:
1 Day
Prerequisites:
Basic Numeracy
PC Skills – Keyboard and mouse skills
Quantitative analysis using SPSS
Aims:
To enable participants to utilise SPSS in simple database setup and manipulation
To understand how SPSS can be utilised in analysing quantitative research data.
Content:
Workshop containing all basic elements of using the SPSS package effectively.
Duration:
3 hours
Prerequisites:
Completion of Introduction to Descriptive Statistics session, or similar understanding
Completion of Introduction to Inferential Statistics session, or similar understanding
Completion of Analysing Your Data session, or similar understanding
PC Skills – Keyboard and mouse skills
Quantitative Analysis using SPSS - part 2
Aims:
To enable participants to utilise SPSS for more complex data manipulation
To understand how to write replicable code
To understand how to perform more complex analysis
Content:
Workshop containing more advanced features of SPSS, including executing syntax.
Duration:
3 hours
Pre-requisites:
Completion of "Introduction to Descriptive Statistics" session, or similar understanding
Completion of "Introduction to Inferential Statistics" session, or similar understanding
Completion of "Analysing Data" session, or similar understanding
Completion of "Quantitative Analysis using SPSS - part 1" or similar understanding
PC Skills – Keyboard and mouse skills
Introduction to Epidemiological Methods
Aims:
To understand different study methods, in particular cohort, cross-sectional, and case-control
To understand event rates in different groups
To understand prevalence and incidence rates and
To understand relative risk and when it can be used
To understand the concept of odds ratios
To understand how to calculate relative risk reduction and absolute risk reduction
To understand the concept of numbers needed to treat
Content:
In this session participants will be introduced to epidemiological methods and their application. In addition to learning some simple epidemiological techniques, participants will be able to discuss the relative strengths and weaknesses of different data collection mechanisms and their associated analysis.
Duration
3 hours
Prerequisites:
Basic Numeracy
Epidemiological / Population Methods
Aims:
To understand descriptive epidemiological design strategies including Correlational, Case Study, Case Series and Cross-Sectional
To understand analytical epidemiological design strategies including Case-Control and Cohort
To understand interventional epidemiological design strategies, specifically randomised controlled trials
To understand the role of chance (p-values and confidence intervals), bias, confounding and effect modification on study results
To be able to stratify data to help remove the effect of confounding
To understand cause-effect relationships and the contribution of strength of association, biological credibility, time sequence and dose-response
To understand and calculate prevalence and incidence, crude and category specific rates
To understand and be able to perform standardisation of data – direct and indirect
To understand Relative Risk and Odds Ratios and their relation to different study designs
To understand other ratios including Standardise Mortality Ratios, Proportional Mortality Ratios
To understand and calculate a number of measures of association, including relative risk reduction, attributable risk, absolute risk reduction, numbers needed to treat / harm
To understand the Chi-Square test and associated p-values
To be able to calculate confidence intervals for Relative Risk
To understand the concept of sample size and calculate for a case control study
To understand statistical power and calculate for a case control study
To understand the need for screening tests including the nature of disease
To understand and calculate the Sensitivity, Specificity, Predicted positive and negative values of a screening test
To understand how optimal sensitivity and specificity can be determined using ROC curves
Content:
In this session participants will be introduced to epidemiological and population methods and their application. The session not only covers the most common epidemiological analysis, but details the different types of epidemiological studies alongside their strengths and weaknesses. Common issues with epidemiological studies will be discussed as well as techniques for overcoming these including stratification and standardisation methods. In addition to learning epidemiological techniques to understand how exposure and disease are associated, participants will be able to perform sample size and power calculations. Finally the issue of pro-active screening for diseases will be discussed and concepts such as sensitivity and specificity will be introduced as well as methods to calculate and maximise them, including the use of ROC curves
1 Day
Prerequisites:
Basic Numeracy, although knowledge of basic statistics would be an advantage
Statistical Principles in Data Analysis (2 Days)
Aims:
To introduce summary statistics and methods of calculation
To introduce distributions, including the normal distribution
To introduce the concept of sampling and standard error
To outline the principles behind formulating and testing statistical hypotheses
To introduce confidence limits and p values and their interpretation
To introduce a framework for choosing the appropriate statistical test
To understand how SPSS (or similar packages) can be utilised in analysing quantitative data
Content:
This two day course is designed to give participants an overview of basic statistics and analysis of data that is most commonly used within research practices and other areas of health care; focusing primarily on key elements of quantitative data analysis. This workshop is aimed at those currently doing research, intending to do research or those with research interests and therefore assumes participants will be familiar with some research and statistical terminology.
Duration
2 days (Four 3 hour sessions)
Pre-requisites:
Basic Numeracy
PC Skills – Keyboard and mouse skills
Statistical Process Control
Aims:
To emphasise the importance of understanding your data
To understand time series data
To highlight problems with existing visualisation methods
To understand natural and special cause variation
To understand different types of process control charts
To understand rules for 'out of control' processes
To discuss process capability
Content:
This session will introduce participants to the concepts of statistical process control, what it is and when it can be used effectively. In particular, participants will have the opportunity to bring their own data and produce simple process control charts from first principles.
Duration
3 hours
Prerequisites:
Basic numeracy
Regression Methods and Models (1 Day)
Aims:
To understand the difference between independent variables, dependent variables, confounding variables and other influencing variables
To understand simple correlation methods, including correlation coefficients and R-squared
To understand 'line of best fit' based on least squares methods
To understand simple linear regression, multiple linear regression and resulting equations
To understand how non-linear data can be transformed to fit a linear model
To understand binomial logistic regression, and how statistical outputs are intrepreted
To introduce multinomial logistic regression methods
Content:
This session will enable participants to understand the principles behind linear and logistic regression. The different methods will be discussed as well as data limitations. Practical examples of how regression models can be applied will be used, and the potentil for using regression techniques as predictive models will be explored.
Duration:
1 Day
Prerequisites:
Basic numeracy
Completion of “Introduction to Descriptive Statistics” session or similar understanding
Completion of “Introduction to Inferential Statistics” session or similar understanding
Introduction to Time Series
Aims:
Knowing the areas to consider and clarify before using any time series method
Setting up and using simple forecasting methods
Confidently using the functions FORECAST and GROWTH within MS Excel
Knowing the limitations of these functions
Identifying whether data has trend and / or seasonality
Using regression functions in MS Excel to analyse data with seasonality
Understanding the advantages of exponential smoothing models over regression models
Identifying the correct type of exponential smoothing for the data, identifying whether it is an additive or multiplicative model
Setting up and / or using exponential smoothing spreadsheets
Understanding the limitations of any forecast produced, including interpreting confidence intervals when produced
Understanding some of the practical issues with data used for time forecasting
Content:
This session will look at how we deal with time related data, and the different types of variation within it. Different methods will be discussed to identify and smooth out the affects of variation, with a view to being able to use data to predict future activity. A particular focus on practical application and the use of MS Excel will be explored, with the main emphasis on ETS models.
Duration:
1 Day
Pre-requisites
Basic Numeracy
Time Series and Forecasting (1 Day)
Aims:
To understand time series data and how it is defined
To understand the relationship with process control and regression models
To understand and identify the different elements of time series data including secular trend, cyclical variation, seasonal variation and residual variation.
To be able to identify the difference between linear and non-linear trends.
To understand when to transform data.
To understand the concept of rolling averages, weighting data and smoothing techniques including exponential smoothing.
To be able to create and use exponential smoothing techniques including those incorporating seasonality
To understand the difference between additive and multiplicative models
To understand the importance of stationarity in data for ARIMA models, and to be able to make data stationary
To know how to interpret ARIMA models and outputs, including auto-correlation and partial auto-correlation functions
To give an overview of BATS and TBATS models within R.
Content:
This session will look at how we deal with time related data, and the different types of variation within it. Different methods will be discussed to identify and smooth out the affects of variation, with a view to being able to use data to predict future activity.
Duration:
1 Day
Prerequisites:
Basic numeracy
Completion of “Introduction to Descriptive Statistics” session or similar understanding
Completion of “Introduction to Inferential Statistics” session or similar understanding