Dec
2018

Making the Most of Budgets for Public Health Interventions

Quality-adjusted life years (QALYs) are a measure of the life expectancy of a patient, weighted by a quality of life score. Generally QALYs are calculated using a naive method which is inefficient and can introduce bias. An improved method uses the statistical technique of joint modelling.

read more
Apr
2016

Habitat Modelling for Protected Species

States are required to monitor and protect plants and animals listed as European Protected Species. However, species data are often ad hoc and vary considerably in their quality and coverage. In these cases, species distribution models can be applied to obtain useful estimates from the limited data available.

read more
Dec
2015

Saving to Prevent Problem Debt

Following on from our previous investigation into the link between cash savings and problem debt, StepChange Debt Charity asked Select to conduct an analysis to understand which types of households are failing to save, and also to identify attitudinal factors that might influence their decision not to save.

read more
Nov
2015

Optimising Stock Control

One of the major challenges all retailers face is how to manage their operations to maintain efficient stock control. Using a statistical approach, historic sales figures combined with data on relevant external drivers can provide accurate forecasts of future demand.

read more
Nov
2015

Modelling Visitor Rates for Special Protected Areas

When local planning authorities are planning a new development, they must consider whether it will affect a Special Protected Area (SPA) and how it could impact visitor numbers. Using data collected by visitor surveys, it is possible to model SPA visitor rates, identify the key drivers and predict how the rate may change for different scenarios.

read more
Nov
2015

Understanding Customer Retention

Customer retention is one of the biggest challenges facing businesses in a wide range of industries, from subscription media and mobile phone operators to fitness clubs and insurance providers. Our client, a national UK business with a customer base of over 800,000, wanted to understand what drives customer retention and attrition.

read more
Feb
2015

Action Plan on Problem Debt

StepChange are the UK’s leading debt charity. As part of their Action Plan on Problem Debt, Select were asked to investigate the potential statistical link between cash savings and problem debt.

read more
Jan
2015

Predicting Energy Efficient Households

The Energy Saving Trust has developed a data set on Scottish housing stock that provides information on property characteristics and the potential for energy efficiency measures. To deliver a complete record of all properties, Select developed a statistical model to predict property characteristics where no data are available.

read more
May
2013

Assessing the Accuracy of Tumour Measurements

Currently the effectiveness of anticancer therapies is primarily evaluated using the length of tumour. A study was performed to identify whether tumour volume measurements were a more accurate alternative in assessing therapy response with the overall aim of informing imaging protocols for use in clinical drug trials and clinical care.

read more
Apr
2013

Using Statistical Sampling for Monitoring Call Centre Quality

Companies who operate call centres such as telesales, market research or utility companies all monitor the calls that their staff make to ensure that they are complying with the relevant protocols. Using statistical techniques, these companies can sample a subset of calls that will accurately monitor call quality whilst minimising cost.

read more
Feb
2013

Developing Medical Decision Tools

The decision to extubate intensive care patients is critical to the long-term recovery of patients. To assist clinicians in their decision, Select developed a statistical model that predicts the probability of a successful extubation based on information about the patient, such as the patient's ability to cough.

read more
Feb
2013

A/B Split Testing and Identifying Target Segments for Marketing Strategies

Many businesses use online marketing strategies to increase the number of visitors to their websites. To test the effectiveness of these strategies, A/B split testing can be used to identify whether they increase visitor conversion to sales. Careful design of the test allows a company to minimise any potential costs and disruptions as well as better understand the demographics of their customer base.

read more
Jan
2013

Assessing the Impact of Climate Change on the UK’s Electricity Networks

The electricity network is susceptible to power outages caused by weather such as severe winds, lightning and snow. In addition to the short-term implications, network operators must consider the long-term implications of climate change. This study uses future projections of climate to model how weather faults may change over time.

read more
Dec
2012

Modelling Ecosystems to Help Inform Habitat Management

Population or community ecology is the study of how population sizes change as species interact both with each other and with their environment through time. Individual based models (IBMs) can be used to model these complex systems that often arise in ecology.

read more
Nov
2012

Matching SMEs with Countries

When choosing to establish a new subsidiary in a foreign country, firms must consider many factors both short-term (such as labour and material costs) and long-term (economic stability). Using a statistical model that considers both the profile of the firm and the political, economic and social characteristics of each country, it is possible to choose a destination that minimises risks and maximises profit.

read more
Nov
2012

How to Measure Effectiveness of a Novel Device

To test the effectiveness of a new device designed to eliminate smoke in the surgical field, an experiment was designed that collected information on the clarity of view during the surgery. By analysing these data it was possible to detect a positive effect of the new device and compare the effectiveness relative to alternative methods.

read more
Oct
2012

Understanding Fuel Poverty

Energy suppliers have a regulatory duty to help combat fuel poverty through the take-up of energy efficiency practices. However, information on an individual's energy use is not always available. By combining a variety of Official Statistics in a statistical model, it's possible to identify areas in the country that are at greatest risk of fuel poverty and where the take-up of efficiency measures is likely to be low.

read more
Oct
2012

Assessing Wild Animal Population Size

Estimating population sizes for wild animals and how they are changing is critical for the conservation of biodiversity. The difficulty is that it's often impossible to physically count them. Data from capture-recapture methods can be used with a statistical model to estimate population sizes and, by including environmental information, it is possible to understand what factors influence changes in population.

read more
Oct
2012

Assessing Cancer Risk

Mutations in the BRCA1 and BRCA2 genes are known to lead to an increased risk of cancer in humans. Elective surgery can reduce the cancer risk, but is an incredibly difficult and personal decision for healthy patients to take. Statistical techniques can be used to produce a personalised risk profile with likely timescales of disease onset allowing a patient to make a more informed decision.

read more
Oct
2012

Minimising Customer Waiting Times

A bank would like to keep customer waiting times below 5 minutes in order to ensure an appropriate level of customer service. The wait time depends on two main factors: the time of day and the number of tellers of operating. By carrying out an operational analysis, it is possible to calculate the optimal number of tellers required throughout the day.

read more
Oct
2012

Modelling Extreme Sea-Levels

When planners are developing coastal flood defences, they have to consider to what height they should build a sea wall so that the probability of flooding over a pre-specified future period is sufficiently small. Reliable estimates can be obtained using a statistical model that incorporates all the processes that affect extreme sea-levels.

read more
Oct
2012

Accounting for Self-Selection Biases in Customer Satisfaction Surveys

Customer satisfaction surveys are a useful tool for identifying strengths and weaknesses in a business. However, these types of surveys can suffer from a self-selection bias reducing their reliability. Using customer information (such as nationality, age or gender), it is possible to model the probability of a customer responding to a survey and to correct the bias caused by self-selection allowing businesses to better understand customer needs.

read more