Statistical Advisory Service

Overview of the service

The Department of Mathematics and Statistics offers a free of charge Statistical Advisory Service (including statistical computing advice) to all Reading University registered postgraduate students who engage in research (undergraduates should instead visit the Mathematics Support Centre's). It is also open to staff whose research has no external funding. For more details about funded projects please scroll down this page.

The Statistical Advisory Service runs as an appointment service only. Appointments are scheduled according to the availability of an appropriate advisor. Please note that in some instances we may not be able to provide advice, e.g. if your query is in a specialised area of statistics or about a non-University supported software package where we have no-one with the relevant expertise. Please note the following as well: Appointments are usually for 30 minutes duration, and will be for a maximum of 45 minutes. If it is not possible to finish the problem within this time frame then further appointments are possible, through resubmission of the booking form in the normal way. However, please note that the usual situation is to limit the number of appointments to a maximum of three per person within a four month period.

If you are completely new to statistics computing packages, then please try one of our statistics packages tutorials before visiting us.

Booking an appointment

Appointments must be booked by completing the Statistical Advisory Service Booking Form: appointments cannot be made in any other way. Note that advice is not given by phone or email, only by face-to-face consultation.

When you complete the form be sure to state clearly whether you are a student or member of staff, and describe your problem carefully. Having a clear idea of your problem in advance will enable the advisor to prepare a more detailed response, helping you to get more out of the service. Thinking carefully in advance, and writing down the nature of your statistical problem may also help you to decide on the type of advice that you're seeking, and is particularly beneficial for students.

You should receive a reply to your submission of the web form within one week, and usually an appointment will be made for some time either the same week or following week, depending on availability. In busy periods or vacations there may be a longer wait, sometimes up to a few weeks. Please do not resubmit your form unless you have heard nothing after two weeks. Note that the Advisory Service is not suitable for last-minute advice!

Repeat visits

Where a follow-up visit is required, this will need to be booked through the web form again. You should indicate the name of the Advisor you have seen before, and whether your problem is a continuation or a new one. Usually you should see the same Advisor for follow-up advice about the same problem.

What the Advisory Service can do for you

Advice at the planning stage of research can help to:

• ensure that the proposed research investigation is designed to address the research objectives within resource limitations as efficiently as possible;
• identify best practices for data management prior to the start of data collection;
• check that the computational facilities necessary to analyse the results are available or can be developed.

Input at the data analysis stage can help to:

• Identify suitable statistical methods that will address study objectives;
• Ensure that reliable and evidence based conclusions can be drawn from the analysis;
• Ensure best-practice presentation of results.

Advice is also given on how to implement the analysis using appropriate statistical software.

Even if you have attended a statistics course before and are familiar with some basic statistical ideas, you may find that a visit to the Advisory Service helps to relate these ideas to your particular problem.

It should be remembered that all University-based human subject research must be considered by the University's Ethics and Research Committee.If relevant, the Statistical Advisor will assume you have already gained approval from this Committee, and it is not their role to check that this has been done. Further information on this can be obtained from supervisors or Heads of School.

Type of advice given

For students and researchers who have no external funding, the Advisory Service is for providing statistics and statistical computing advice. It aims to encourage staff and students to learn to deal with the statistical and data management aspects of their research effectively and efficiently. Students should consider this learning process as part of their degree programme experience.

Support offered by the Statistical Advisory Service includes:

• input to the design of your research study, in terms of resources, and how best to plan the study to address the objectives most efficiently and avoid bias in the results;
• advice on the design of your data collection instrument (questionnaire etc.);
• advice on how to approach the analysis of data collected in your research investigation;
• assistance with the meaningful interpretation of the statistical results of your analysis;
• guidance on how best to enter, store and manage your data on the computer;
• assistance in solving more complex statistical problems connected with your research.

Provision of statistics advice at a professional standard is based on the range of expertise of Statistics staff. This includes for example, experimental design and analysis, survey and sampling methodology, statistical modelling, epidemiology, clinical trials and intervention studies, genetics, forensics and multivariate analysis.

What the Service does NOT provide

The Statistical Advisory Service cannot take responsibility for the outcomes of the research. It has no obligation to monitor or evaluate work undertaken as a result of an advisory session. As such, projects cannot be signed-off as being statistically acceptable unless the statistician has been included as part of the research team and has undertaken the analysis. Short advisory sessions will not generally qualify for letters by the statistician confirming the validity of the proposed research and you should not include the name of the statistician whom you have seen on any Ethics application or similar document without their approval. Statistical advice for grant applications is provided in a different manner, details of which can be found below.

The service also cannot provide tuition that substitutes for proper statistical education that the advisee should have acquired, e.g. as part of a degree programme. Basic information about a statistical method is best acquired by visiting the Mathematics Support Centre's Statistics Tutor. A basic knowledge of the main statistical computing packages (Minitab, SAS, SPSS, Genstat) can be obtained by working through our statistics packages tutorials in your own time. We have also produced some statistical computing tip sheets on the use of Minitab and PASW (SPSS) for certain statistical analyses. You may also wish to take advantage of the following training opportunities

1. The RRDP at Reading (Reading Researcher Development Programme) which is free to all PhD students.
2. The Statistical Service Centre's annual short course programme: they give a 50% discount to internal staff and students (conditions apply). 

Externally funded projects and grant applications

The support offered by the basic Statistical Advisory Service does not cover any consultancy inputs to externally funded projects; yet statistical input can help to ensure that planned studies are both resource-efficient and produce outcomes that are based on solid and well-defensible evidence. Assistance with statistical analysis can also be highly beneficial.

For staff submitting research proposals and running externally-funded projects, there are two possibilities to explore.

Individual contact

Our Statistics staff are experienced in external consultancy and, workload and expertise permitting, may be able to provide one-to-one advice on analysis of data from existing externally funded projects. Depending on the extent of work, co-authorship on any resulting article may then be appropriate. We would encourage you to contact individual members of staff in this instance to see what they can do for you (statistics staff who may be able to help are: Ayres, Baksh, Todd, Zhou). Staff contact details and links to webpages outlining areas of expertise are available under the About Us section of this website.

For input to grant proposals, and for projects requiring a significant amount of advice or analysis, it is most appropriate to include a statistician as a co-investigator. The statistician is then able to provide, free of charge, input into the writing of the grant proposal, including for example sample size calculations and a description of the relevant statistical analysis methods to be employed.

As stated, depending on available expertise and workloads, we may not be able to assist in these areas. In such cases we would suggest you explore the service offered by the Statistical Services Centre, described below.

Statistical Services Centre parallel service

The Statistical Services Centre (SSC) is a self-financing not-for-profit unit within the University.  It has been providing paid-for statistical consultancy and training to a wide range of clients, who are external to the University, for over 25 years.  
If inclusion of a statistician as a co-investigator is not suitable for your grant, then you could seek statistical support from the SSC.  They can offer a free-of-charge initial discussion to identify the statistical requirements of your grant application. Where they are able to provide the necessary statistical support to your proposed project, their inputs will need to be fully costed into the proposal.  SSC staff are happy to discuss and supply this information during preparation of funding applications.

The SSC can also provide statistical analysis support to already established funded project.  This support is not free of charge.  You will need to have funds to pay for their inputs.  If you are interested in this service the SSC would be happy to discuss your needs.  If they have the relevant expertise to help with your project they will give you a quote for the work and a timeframe within which it can be done.

Please e-mail statistics@lists.reading.ac.uk if you wish to make use of this service (do not submit the web form in this instance).

Making the service work for you

Successful statistical consultancy requires mutual respect and co-operation on the parts of both the researcher and the statistics advisor. The advisor needs to develop an adequate understanding of the background to the statistical problem for which advice is sought. The researcher therefore needs to describe the study objectives and expected outcomes quite thoroughly to get the best advice from the statistics advisor.

The following is a list of things that each can expect and provide.

 

What you should, and should not, expect from the statistics advisor

 

1. You should expect the advisor to be considerate, attentive and constructive. However, the advisor may not always have an immediate answer and may request to see you at a later time by mutual appointment. Alternatively, you may be referred to another statistical advisor for specialist help. For assistance with general computing aspects of your research, you should consult IT Services.
2. You should expect advice from the statistician on how to design your study, manage your data or carry out a statistical analysis. It is not the advisor's role to do the work for you, unless this has been agreed in advance under the extended Statistical Advisory Service, or to take responsibility for monitoring your work. Hence letters confirming the suitability of your study approaches and analysis cannot be provided unless the statistician's involvement has been substantial.
3. Advisees cannot necessarily expect support on any software package which is not centrally licensed by the University of Reading or whose main use is not statistical.
4. The role of the statistician is to work closely with you - and your supervisor, if you are a student - in order to achieve aims which have been agreed in advance about a project. We strongly advise that the supervisor accompanies the student for at least the initial consultation with the statistics advisor.
   

What the statistics advisor expects from you

 

1. Recognition that statistical inputs to a project at its planning stage can help to lead to better designed studies, as well as saving time and resources. This is as important as, if not more important than collaborating in data analysis.
2. It takes time to set up an advisory session, as well as to provide the advice and to act on it. Thus, if you have a deadline to meet and need prior statistical advice, we urge you to seek the advice well in advance.
3. Please try to explain the whole problem to the statistician at the initial meeting, rather than presenting different components of your problem at different times. This will make for greater efficiency and understanding on the part of the advisor. It will also help to ensure you are given the correct advice first time round. If you can bring along clearly stated protocols, and any necessary diagrams or output, that will be useful.
4. Try to write a detailed description of your statistical problem on the web form when requesting an appointment. It will be helpful in identifying an appropriate advisor, and will allow them to prepare for the meeting.
5. When the statistical input to a research paper being submitted for publication has been substantial, it is right to consider co-authorship with the statistics advisor. Such cases include those where the statistician has done the analysis for you, or where they have driven the direction of the quantitative research. For cases where only advice was provided, co-authorship would not usually be appropriate. However, please do not make the statistician a co-author or acknowledge his/her inputs to your project without his/her agreement.