Analysis of Large Datasets
Analysing existing data can help us to gain new insights to illuminate policy and practice questions. At NFER, we are adept at applying secondary data analysis methods to key education datasets to do just that.
There are a number of large datasets in the education sector which contain a rich set of information that we can use to answer important policy and practice questions. NFER’s professional economists and statisticians are ideally placed to explore the possibilities with you as we have a deep understanding of key data sets such as the National Pupil Database, the School Workforce Census, Understanding Society, and the Labour Force Survey. We also have extensive knowledge of education policy, excellent technical skills and a wealth of practical experience of using quantitative methods to analyse large datasets.
We offer a full quantitative analysis research service, which includes:
- advising clients about the most appropriate data sources to use to answer key questions, and their strengths and limitations
- the use of robust methodologies to conduct rigourous analysis of different kinds of data
- applying to data owners to get secure access to sensitive datasets
- presenting highly technical information in accessible ways
- helping our clients interpret what the data means for them and translating the analysis into insights and recommendations that they can take forward.
We use a range of data analysis and quantitative techniques to do secondary data analysis and undertake policy evaluations, including:
- descriptive analysis, which is useful for deriving insights about what the current policy landscape looks like and developing hypotheses which can later be tested using more sophisticated quantitative techniques
- matched comparison designs (for example, propensity score matching), which we use for evaluating the impact of a policy change
- regression methods, which we use to identify the association and relationship between different variables
- cluster analysis (e.g. latent class methods), which we use to identify distinct subgroups within a population
- value for money evaluations, including cost / benefit analysis and the cost effectiveness of a range of education policies and initiatives.
Using existing data, rather than collecting new data, can be a very cost-effective option. It is also the ethical choice where the data needed already exists, as it avoids burdening research participants with a new data collection exercise. We are happy to provide advice on which design and methods will work best for each client’s particular study, participant group, timescale and budget.