How to write a statistics dissertation
If you are lucky you will have a clear idea of what research you want to carry out and how you are going to go about it. Most people, however, need to examine the subject and read around the issues to first have an idea of what would make an interesting and important dissertation. One way to do this is to consider the type of evidence you require. A statistics dissertation is rarely written on the abstract and therefore you should consider the various sources of data. A statistics dissertation is essentially asking questions of data and applying different methods to answer the questions. A body of quantitative data is thus required and this should be your starting point. Some dissertations will require collecting this yourself, and others will permit the use of pre-existing datasets. One way to get an idea for a statistics dissertation is to look at what data is available and consider the questions that can be asked and some ways in which you want to answer them. Another possibility is to consider the methods you want to use. For example, if you want to explore logistical regression, to find a dataset that allows the use of such techniques. This will typically consist of both discrete and continuous data, and you will want to examine how one affects the other. You should play around with data for a while to see what can be done before formulating a research question. An initial idea for a dissertation will often comprise of a subject area and a set of issues you are interested in. It is important to attempt to pare this down to avoid being consumed by detail, and trying to think of your dissertation subject as a simple question provides a good way to focus. When deciding your subject area it is usually important to find something you are interested in which you think you will enjoy investigating. However, it is also important to make sure you have the skill set required for the dissertation. This is possibly more important in statistics than in other disciplines as it is not just attempting to read a number of books, it also includes access to and expertise in software, mathematical ability and an understanding of the issues. In statistics it is vitally important to understand the techniques fully, otherwise your analysis can appear shaky and tenuous. It is not a good idea to decide to find a research question in order to learn a set of skills, you should make sure you understand what you are setting up for yourself before you start writing.
One you have formulated a research question, have selected a dataset and are aware of the various statistical methods you wish to use for the dissertation, it is vitally important to research your dissertation. A common problem with statistics dissertations is that they can run out of control and a student can become buried under a vast number of data readouts. You therefore must plan each chapter carefully before you have begun. This way, if you get sidetracked during the analysis and discover new and interesting questions to ask of the data, you can make a more prescient decision of whether it conforms to the research question. Serendipity plays a part in dissertation writing and often the joy of doing research is coming up with results you were not expecting. However, you should consider the flow of your dissertation as taking the reader from the first chapter to the end of the conclusion in a straight line; if you change your research half way through the resulting dissertation will become disjointed, it will read badly and the conclusion will not match the introduction. It is vitally important to avoid this, and if you discover something really interesting that you believe will improve your dissertation marks you can always put it in an appendix. The number of chapters and level of detail necessarily depends on the word count of your dissertation. Be aware of how much word count is affected by tables and graphs and check the rules and regulations to make sure you do not get penalised. Thorough and careful planning is vital to writing a successful statistics dissertation and you should devote a significant amount of attention to this stage of the process.
If you have a good research question and planned your statistics dissertation well, the writing should be almost a matter of course. You should always appraise what has been written before. In statistics this can involve using similar techniques to look at different datasets, which allows you to see how effective your chosen techniques can be, and also get some ideas of what to do differently. Alternatively, you can find similar evidence and see what statistical methods have been used and challenge their techniques and conclusions using a superior method. To research this section of your dissertation, it is useful to examine online journal stores such as Jstor and search the abstracts by keywords. The abstract will usually have a brief mention of statistical methods or the dataset and you can use this as a starting point. Once you have found articles that you find useful, you should consult their bibliography to see where their source material is gleaned. Of course, your supervisor will also provide guidance but it is useful to use initiative and you will be more likely to be rewarded with higher marks. Remember newer is not necessarily better, and although you may want to be on the cutting edge some great research has always been conducted by taking forgotten ideas and making them current again. Do not get bogged down in the research and make sure you leave a substantial amount of time for the analysis. Remember in statistics, analysis can take up more time then you might realistically estimate and unless you are very advanced in the software you are using it is best to play safe. Pay attention to formatting and make it consistent throughout: presentation will affect your marks, and even small and insignificant mistakes can make work look unprofessional. Remember you want to get the highest mark possible for your statistics dissertation, and this means following the stages through with the greatest attention to details. This means formulating a good research question. Discuss it with your supervisor and modify it if necessary: a poor dissertation can be written with a great research question, but it is more unlikely to have a great dissertation based on a poor research question. Research your dataset and techniques. Plan your dissertation very carefully. When writing pay great attention to format and presentation issues as this will pay off. If you follow these pointers closely, you will be much more likely to write a successful statistics dissertation.
Other dissertation help pages in this category
- Accounting Dissertation
- Animation Dissertation
- Anthropology Dissertation
- Archaeology Dissertation
- Architecture Dissertation
- Art Dissertation
- Biology Dissertation
- Business Dissertation
- Chemistry Dissertation
- Criminology Dissertation
- Economics Dissertation
- English Language Dissertation
- English Literature Dissertation
- Film Studies Dissertation
- Finance Dissertation
- Law Dissertation
- Management Dissertation
- Marketing Dissertation
- Media Dissertation
- Nursing Dissertation
- Science Dissertation
- Social Work Dissertation
- Sociology Dissertation
- Sports Dissertation
- Statistics Dissertation
- Theatre Dissertation
- Tourism Dissertation
Need help writing a dissertation?
