1. Earth Sciences: This is a very easy-to-navigate unit that covers many topics. With regards to mathematics, the material is useful if somewhat limited in scope. However, there is a section that commences with looking at graphical representation, gradients and variables. There follows a straightforward piece concerned with interpolation and extrapolation of data. Log plots are explained well and there is a clear explanation of triangular plots and why and when they are appropriate. Each page neatly incorporates animated graphics where appropriate and there are interactive maths problems to solve before moving on to the next page. This section should take about one hour to complete and would be time well spent.
2. Archstats; Also very easy to navigate, good graphics and a practical approach to the subject. There are 10 sections, each covering one aspect of elementary statistics. Each topic is interspersed with both examples and a few interactive exercises. There are no pretensions at depth here but simply to offer the reader an insight into the area where statistical analysis and archaeology overlap. In fact, the unit does not really go beyond hypothesis testing. The work concludes with a 'dig' in which the reader is asked to complete a series of analyses on the material being uncovered on the 'dig'.

There is much more to this site than is at first apparent, because it is very colourful and carriesminimal text, there is the assumption that it must be 'lightweight' in some way. This is certainly not the case. All archaeology students would do well to spend some time on this unit, it will certainlyhelp them with whatever statistics they are required to learn.

A cautionary note:

Electronic sources should be the ideal way to acquire very 'up-to-date' data for research, assignment or dissertation purposes but be aware that web sites tend to be neglected once they are launched, even by large institutions. Similarly, small websites are often 'abandoned' by their authors. If you do decide to utilise web-based data in your submitted work make very sure that it can be authenticated and that you reference it very carefully so that others may also check its validity.

Thus below is a carefully chosen selection of what seems to be the 'best on the web' at the present time (Sept 2002), however it is very likely that any such list will need to heavily overhauled each year.

Also, be prepared to explore links from these sites for yourself .

http://www.statsoft/textbook/contents

This is possibly the most comprehensive website that is aimed at students, trainees and researchers alike. The comprehensive nature of the material is impressive. Some of the topics are over-condensed and thus difficult to follow. Nevertheless, the all-important elementary material is excellent.

Firstly go to the CONTENTS page and open ELEMENTARY CONCEPTS. There are 22 sub-headings here and each topic has been condensed to just 2 or 3 paragraphs, e.g. "What is Statistical Significance? (p-values). Whilst the pieces of text are decidedly brief and to the point, they do give clear definitions of all the required concepts in a very short space of time.

There are numerous animated graphics scattered throughout the site that illustrate how (statistical) values can change as parameters etc change.

The BASIC STATISTISTICS section is comprehensive and describes the theoretical basis of the commonly used tests (e.g. Students t-test) very well. Do note however, that few worked examples are actually given and the aim of the text is to explain the philosophy behind each procedure rather than quantifying the maths involved.

Look also at the sections covering Linear Regression, ANOVA, Cluster Analysis and Discriminant Analysis.

The comprehensive nature of this site is most impressive and the screen layout is easy to read.

There are also a number of sections that cover more 'specialist' topics that enterprising final year students may wish to explore with the aim of using appropriate methods in their final dissertations.

There is a comprehensive Glossary of Terms and an 'Advisor' section to guide the reader to the most appropriate tests for the data analysis that he / she may need to carry out.

This site should be used as a 'first port of call' for all Conservation Science Students who are attempting to understand how and what techniques exist to analyse their experimental data.

A small site that concentrates on two topics in particular; Linear Regression and ANOVA. The ideas surrounding statistical modelling are outlined clearly. Useful site to visit once the idea of Analysis of Variance is generally understood.

There is a data set archive on this website that relates to ANOVA, Linear and non-linear regression. Perhaps of limited value to students but does give worked examples of a number of test procedures.

A complicated site but there are over 300 links to other sites. Of particular note are the links to the on-line material available from other Universities around the world.

"Seeing Statistics" is an online textbook that uses Java applets to visually demonstrate certain statistical concepts, Normal Probabilities for example. Enter your own data and see the effect upon the area under the graph. Unfortunately, only a small portion of this site is available without online registration.

This is simply a library document but it lists a complete range of other sources of data for and about the UK

This site is included because it demonstrates how statistical analysis has become an essential research tool…in this case to examine, record and unravel the mysteries of the world's climate. By developing statistical models based on observed values the researchers hope to improve the predictability of future events and climate change indicators. The ability to forecast hurricanes or El Nino's for example could save lives, crops, livestock, property and infrastructure. The process of observation>>> analysis>>>modelling>>>prediction is clearly seen. The site is very well built, easy to navigate and above all offers fascinating, up to date and pertinent use of statistics.

http://www.sjsu.edu/faculty/gerstman/StatPrimer/

Do read the essay at the beginning of Part 2 "The Median isn't the message", it will only take a few minutes but gives a clear insight into the problems of trying to interpret 'statistics' if you don’t have a clear understanding of their true derivation.

A particular feature of this site is that each topic is accompanied by a set of exercise question but they are somewhat unusual in that they test comprehension rather than simply running through a maths procedure.

There are 20 topics covered, all in a straightforward manner. Worked examples are plentiful albeit that they seem to be directed towards the medical and behavioural sciences e.g. Section 13 'Cohort Studies and Risk factors'. The site is written in PDF and can be slow to load.

There is a useful supplemental topics "How to know what to use" and this section attempts to answer the often repeated question "Which test do I use". Here, the authors emphasise the idea that one should firstly decide what output is required.

In conclusion, this is a no-nonsense, clearly defined site with no frills but it is easy to follow and easy to access and should therefore give the reader a readily absorbed field of solid, basic information.

Whilst "the lives of the great statisticians" may seem irrelevant, it will do no harm to read a little of the work of such men as Christiaan Huygens and his Essay on the value of Chance" 1697 and perhaps one of the first deliberately compiled data sets: The General Bill of Mortality 1665.

The particular interest on this site has to be the wide range of interactive on-screen calculators. Initially, it is easier to use than SPSS or Excel because each calculator is specific to a particular type of test. The screen prompts will ask you to enter data in the required format prior to analysis. Besides the statistical output, the graphs that can be constructed are particularly clear and well displayed. Try for example the Histogram maker and the Box-Plot maker. The correlation and Regression calculator is a particularly interesting tool and easily allows the reader to familiarise oneself with the behaviour of a scatterplot and what the generated statistical values actually mean.

The page illustrating the "Normal Curve in action" demonstrates the relationship between the mean, standard deviation ,the area under the graph and Probability.

This site did prove difficult to load on occasions but is well worth the effort.

http://europa.eu.int/comm/eurostat/Public/datashop

This site is primarily designed for commercial use and therefore much of the information is only available on subscription. However, there is still a large array of free data ranging from energy consumption, atmospheric pollutant levels and agricultural production figures. The site is very large but is not difficult to navigate. A fruitful source of meaningful data that should encourage the reader to think about how all this information can be fully digested to make it useful in the real world.

http://www.shef.ac.uk/uni/projects/wrp/mjcdisp.ppt

A very quick and simple site containing two Powerpoint presentations aimed at first year Biology / Medical students. The aim of the presentations is to take some of the 'fear' out of working with numerical data and covers summary statistics and graphical representations only. The site is worth 20 minutes of your time.

This site from the Irish Statistics Office is exceptionally well presented and could be considered as a 'benchmark' of good practice with regards to the presentation of complex numerical data in a fashion that it comprehensible to most people. There is an excellent Links page to many other National Statistics sites from many countries around the world. The site is well worth a short visit for this reason alone. Look particularly at the Agriculture statistics and the Tourism statistics.

http://www.statistics.com/

A very comprehensive site with a large variety of datasets to explore. Particularly useful for comparative purposes. There is a large amount of demographic material (facts and figures etc) and they are mainly US orientated. This is not a site to learn how to deal with data but it is a good indicator of the vast array of formats that you may encounter data in.

This is a vast site and gives a wide selection (4155 at the time of writing) of databases ranging from fish stocks to air pollution. Each set is available in Excel and can be downloaded if required. Navigation is easy and many of the data sets are ideal for experimentation and gaining practice in handling large data sets ….and these are not hypothetical and so may add a valuable element of realism to a practice session.

This is the US equivalent of the above site and performs much the same task. The site did prove to load on occasions. The most interesting datasets tended to be concerned with finance, both personal and governmental.

The Department for Environment, Food and Rural Affairs has produced this broad but competent website and they have included a mass of information concerning the current state of the environment of the UK. Those interested in trends and trend analysis will find much to interest them here.

The concept of 'Quality Indicators' both regional and national are introduced. These datasets are primarily non-parametric and thus can be ranked. This immediately suggests that one could use them for a wide range of non-parametric tests. Most of the datasets relate to recycling and waste management issues 1996-2001.

There is also a 'Digest of Environmental Statistics'; all of which may be downloaded in Excel, PDF or .csv formats.

Another very large site but easy to navigate. There are plenty of relevant sections to investigate such as River Quality statistics, Pollution levels, Effluent discharges etc. Check out the Bathing Water Directive pages, there are some fine examples of the good use of graphics to display complex data in a comprehensible fashion. It is also possible to locate usable sets of local data using the 'My backyard' facility and it is here that the Meteorology pages might prove useful.

A particularly well designed site with a large number of links to other sites. The graphical content is good but is limited only to energy production /consumption statistics. If 'trends' data is required then this is a good source because records are complete for the past 33 years.

Explore the book review page also