Most projects can be broken down into four broad steps, which can then be broken into smaller steps. The four main steps are:
Data are used to answer questions. The question and the data need to match. There is no point trying to determine how far green frogs will jump if you only have data on red frogs’ jumps. If you plan to collect your own data, you should design the data collection around the question of interest. If you want to make use of existing data, you have to choose the question appropriate for the available data.
To properly plan a project, the following questions should be asked:
What exactly is/are the question(s) of interest?
It is important to know and understand what project you are doing, why you are doing it, and what you hope to learn from the project. Be clear in your objectives, and break it down into smaller parts if you can. Don’t try to do too much all at the same time. Trying to report too much at once spoils some good Projects.
What data did I use to answer the question of interest?
Finding an answer to your question is not the end of the process. You would normally want to tell others the answer – and they will be interested in seeing how you arrived at that answer. The role of your Project is to show your question and answers to the readers and to convince them that your conclusions are correct. One rule of thumb is to give sceptical readers sufficient information about your methods, so that they could run your study themselves, if they wanted to, to check your results.
Carefully consider and document how you obtain your data.
Collect and compile the data
Collect your data carefully. Errors in data collection or data selection could invalidate your results. Look at the data you have and ask yourself whether it looks sensible. If not, double check for any errors. It is not unusual to find that there have been errors in recording data; try to find these errors, and either fix them if you can, or just delete these observations.
Consider the completeness and adequacy of your data.
Do you really have enough observations to answer the question of interest? Do you have all the measurements which may be relevant to the question of interest? Does the data set contain any missing data and, if so, how many? Identifying and acknowledging weaknesses in your data proves your understanding of statistical analysis to the reader.
When analysing the data, you seek to identify the main patterns within the numbers you have collected.
Use statistics to analyse your data
Use some simple statistics to find out what your data is saying. You may use pictures and graphs (‘a picture tells a thousand words’) such as histograms, barcharts, boxplots, scatterplots and/or quantities such as means, standard deviations, medians, two-way tables etc.
When you report your findings, you must be clear and concise about what you did and how the data answers the question of interest.
Interpret and make conclusions
One of the goals of statistics is to gather data and turn it into information. Look at your data: what does it tell you? How does it answer your original questions? Try to critically review your data, analyses and conclusions. Be honest and open with any potential weaknesses – prove that you understand the strengths and weaknesses of what you have done. Suggest how you would do it better next time.
Avoid over-interpreting the results
Even if three students at your school have reported near misses when crossing a particular road this may not necessarily be sufficient justification for recommending the building of a $1million pedestrian overpass. We don’t actually have proof that an overpass would correct the problem. Keep the conclusions and recommendation in context with the results.
Check your Project
You may like to read the competition checklist for what to present on the Project. Check your Project for errors. In the past, good Projects have been spoilt by incorrect wording, typing errors and calculation errors. In at least two of those cases, the Projects missed out on prizes because of these simple errors.