1500.0 - A guide for using statistics for evidence based policy, 2010
Latest ISSUE Released at 11:30 AM (CANBERRA TIME) 20/10/2010  First Issue
Page tools: Print All
Contents >> Analyse, interpret and evaluate statistical information

ANALYSE, INTERPRET AND EVALUATE STATISTICAL INFORMATION

The process of data analysis, is the process of turning data into meaningful information. Although there are no hard and fast rules for how to analyse statistical data, ensuring that you have a methodical and systematic approach is vital to ensuring your analysis is accurate. Poor quality analysis can lead you to draw incorrect and inappropriate conclusions.

 STEP 1 Identify the issues or questions you require information about, specify objectives, and formulate expectations To ensure the analysis conducted is appropriate for addressing the underlying objective, it is vital that you understand the issue you are investigating. It is also necessary to understand the interrelationships that exist between relevant social, economic and/ or environmental factors relating to the issue. You can then formulate a set of questions which you’re seeking answers and specify your objectives for analysing the data. You may like to consider: • what is the topic or issue? • what is the context in which to understand this issue? • how will the analysis be used? For example: • the issue is the increase in students leaving school before completing year 12 • the context might include, what is the economic, social and demographic characteristics of students who are and aren’t leaving school • the analysis will throw light on the circumstances between those staying and those leaving school, and help introduce a program aimed at encouraging students to remain in schools until year 12. It is also a good idea to formulate a set of expectations for what the data might reveal. Developing an understanding of why certain patterns might emerge in the data and what it might mean for your analysis, will help you analyse the data and draw conclusions.

 STEP 2 Determine appropriate analytical techniques and undertake data analysis Determining which analytical techniques are appropriate for investigating the data is necessary before any analysis takes place. The different analytical tools and techniques available range from simple (e.g. measures of spread) to quite complex (e.g. modelling). Keep in mind that some analytical techniques are not always appropriate for all sets of data. It is important to ensure that appropriate techniques are used in order to avoid misinterpretation or misleading results. Most of these statistical measures can be calculated automatically in spreadsheets. The different analytical techniques can be broadly broken down into summary statistical measures and graphical analysis, however these are often used in combination. Graphical analysis Graphical analysis is a useful way to gain an instant picture of the distribution of the data and identifying any relationships in the data that require further investigation. Patterns in data can be more easily discernible when displayed in graphs. A range of graphical techniques can be used to present data in a pictorial format. For example, column graphs, row graphs, dot graphs and line graphs. One way of summarising data is to produce a frequency distribution table or graph. A frequency table is a grouping of data into categories showing the number of observations in each category. These categories are referred to as classes. Once the class frequencies have been produced, the distribution can be represented graphically by column, row, dot or line graph. It may also be appropriate to plot relative frequencies to show the percentage of the population within each class interval – which enables the different sizes to be directly compared. Summary statistical measures Calculating summary statistics will assist you to understand the distribution of the data. These summary measures are useful for comparing information and are more precise than graphical analysis. Summary statistics assist you to develop an understanding of: • the centre of a set of data. This is important as we often want to know what the central value is for the sample or population. The mean, median and mode are useful measures of central location. However, these measures of location can’t tell the whole story about the distribution of the data. It is possible for two data sets to have the same mean but vastly different distributions. Therefore, you should also analyse the amount of variability in the data • the variability or the spread of the data. The range, inter-quartile range, standard deviation, and variance are useful measures of variability or the spread of the data. There are also a range of analytical techniques that can enable you to gain a deeper understanding of the data. This can involve analysing the data to determine change over time; comparison between groups; comparing like with like; and relationships between variables. Modelling techniques such as linear regression, logistic regression, and time series analysis are some ways to explore these relationships. Assistance can be sought from experienced analysts when undertaking complex statistical analysis.

 STEP 3 Assess the results of analysis against the objectives and expectations Once you have analysed, and computed some statistics from the data and feel that you have a good grasp for what the data is saying, you can then look at drawing appropriate conclusions about the data. This process can be quite complex depending on the questions you are seeking answers for and in some instances, the answers will not be clear cut. Your analysis may provide you with the basis for describing what happened but there may be many possible reasons for why this has occurred. It is important not to consider the issue in isolation, but to think about the interrelationships between social, economic and environmental factors. You may need to seek clarification through further analysis and research to ensure the conclusions you draw are accurate. Some things to consider when drawing conclusions may be: • what are the main results or conclusions that can be drawn? • what other interpretations could there be? • can the results or conclusions be supported statistically? • do the conclusions make sense? • do the results differ from initial expectations?

 STEP 4 Review the objectives, recommence data analysis cycle as appropriate If there are still questions unanswered, you may need to begin the data analysis cycle again. If you’re interested in learning more about data analysis, the ABS delivers a specialised training course on ‘Analysing Survey Data Made Simple’. The course covers these basic principles for analysing data as well as exploring more complex data analysis techniques.

 Previous Page Next Page