On-line statistics

Introduction to Statistical Analyses

Feeling out of your depth and a little out of sorts?

Q: Why are statistics so difficult for Biology undergraduates?

A: They are not, it is all in the imagination.

Biology undergraduates generally treat statistics like bubonic plague. But you will be surprised to learn how familiar statistics are to us. Without mental statistics we would not know when to leave our house in the morning and how much money we need for a night out. They do not possess sharp fangs, though they may indeed penetrate your nightmares in your third year. This web-site will allay your fears and provide help on those statistical tests commonly used in the life-sciences. There is no pressure on anyone visiting it, and you can visit time and again to soak up the life enriching wisdom found hereafter.

Statistics are mathematical methods of interrogating data. As such you must know the questions you wish to ask before you collect your subjects and instruments of torture. If not, the questions may be irrelevant to your victim or your instruments may not be sharp enough to extract the required response. Read Chapter 2 of Fowler and Cohen before you start!

If you already have your data and are searching this web page to find out what to do with it, you are using statistics in the wrong way. That said, you may be able to redeem your efforts by following the instructions on this web site. Ensure for your own sake that you get it right in future.

Do not get hung-up on the mathematics, instead concentrate on the purpose of each test and the type of data they must be used on. Until you become more familiar with the test and acquire experience, treat them as tools. To use the tools properly you must know their applications and limitations but not necessarily their internal mechanics.

Read the whole of this page before you proceed!

New Users: Take a bit of time to glance around this web site to familiarise yourself with what it contains. Thereafter follow the guidance from A to G below. Once you have found out what type of data you will acquire from your experiment or survey, you will need to decide what type of question you are going to ask of your data. Five categories of questions covered by this site are listed on the Categories page. On selecting a test type, you will be taken to a key. Use the keys to decide on a suitable test. There will be some uncertainty regarding the final test you use because of some unknown characteristics of your data (i.e. normal distribution). However, the key will steer you in the right direction until you have some data to analyse. Each test then has its own page for a fuller description. Links are provided from every page to other relevant tests.

Previous Users: If you are familiar with this site or are confident in the test you require you can use the "Quick Jump" list near the bottom of this page.

Once you are familiar with what this site has to offer use the relevant area based on the following steps:

1. Make an initial appraisal of your data (Data types and initial appraisal)

 

2. Select the type of test you require based on the question you are asking (see Categories)

 

3. Select the actual test you need to use from the appropriate key

 

4. Determine any preliminary tests you need to carry out prior to performing the statistical test

 

5. If your data are suitable for the test chosen based on the results from 4 proceed to the test

 

6. If your data do not meet the demands of the chosen test go back to 3 and choose the non-parametric equivalent.

 

7. It may be that your data are still not suitable in which case you need to search wider than this web site or get more data or discard them (one of the problems you may face if you have not planned properly)

There are three sections that are generally applicable:

Glossary

Data types and initial appraisal

Overriding limitations of statistics

Use the first two pages to find out more about statistical terminology and the type of data you are expecting. The third page is essential reading to ensure that you use statistics and the results obtained properly.

---

Descriptive Stats

Diversity Indices

Comparisons

Correlations

Regression