Analytical methods in Statistics New Zealand encompass a range of tasks and methodologies.
The majority of Statistics New Zealand's outputs consists of tables and graphs. Analytical methods require that those tables and graphs meet the analytical needs of users and to that end Statistics New Zealand has developed protocols for their production and presentation.
Statistics New Zealand is obligated to provide outputs that do not compromise the confidentiality of respondent data, and analytical methods include confidentiality techniques used for this purpose.
Statistical analysis includes:
- Exploratory Data Analysis (EDA), an important part of any preliminary analysis of data.
- Time series analysis, mainly seasonal adjustment and trend estimation.
- Multivariate analysis, such as multivariate regression and principal component analysis.
- Longitudinal data analysis, which has become important with current development of longitudinal surveys.
Statistics New Zealand recognises that all users of its data require statistical knowledge so part of analytical methods includes training, both internal and external.
To ensure that its analytical methods are statistically defensible, Statistics New Zealand releases research reports and makes presentations on its methods at conferences and meetings.