Technical Notes

Population

From 2008, the Government Use of ICT Survey population comprises all State sector organisations in the Public Sector Directory (http://www.psd.govt.nz/) except those classified as Conservation Sector organisations (Fish and Game Council, Reserve Boards etc), School Boards of Trustees, Tertiary Education Institutions, and Trusts.

In 2008 this resulted in a total of 163 organisations in the population, comprising:

• Central government
  • Public Service Departments
  • Non Public Service Departments
  • Offices of Parliament
  • Autonomous Crown entities
  • Independent Crown entities
  • Other Crown entity companies
  • Crown agents
  • District Health Boards
  • Other Public Finance Act 4th schedule organisations
  • Reserve Bank
  • Crown research institutes
• State-owned enterprises.

All organisations were invited to participate in the survey with the exception of five very small (less than about 5 employees) organisations which were excluded for respondent burden reasons.

Response rate

Particular effort was made to obtain responses from larger organisations in the population, resulting in a response rate varying from 22 percent for organisations with less than 50 employees, up to 95 percent for organisations with over 1000 employees. The overall response rate was 64 percent.

Data collection

The survey comprised an online questionnaire constructed using the open source software LimeSurvey (http://www.limesurvey.org). A printed version of the questionnaire was supplied to the few respondents who requested this method.

Collection frequency

From 2008, the Government Use of ICT Survey is expected to be an annual collection.

Imputation

Imputation was used to obtain data in cases of non-response for the questions regarding ICT operating expenditure and ICT capital expenditure only. All other data is simply reported for the responding 105 organisations.

For expenditure, four imputation methods were used depending on the organisation type, to take advantage of the different expenditure characteristics of each organisation type. For each of operating and capital expenditure:

• Non-responding District Health Boards (DHBs): Linear regression fit to responses versus population served.
• Non responding Crown Research Institutes (CRIs): Weighted mean (weight = number of employees determined a priori through published figures).
• Non-responding energy companies: Weighted mean (weight=number of employees).
• Other non-responding organisations (all less than 1000 employees): After removal of outliers, CRIs, DHBs, energy companies, and organisations with more than 1500 employees, linear regression fit to responses versus number of employees.

Non-sample errors

These cannot be quantified and could include mistakes by respondents, variation in interpretation of questions, and data processing errors. Additionally, survey respondents were CIOs or the equivalent role in the organisation who may not have complete knowledge of all IT and business matters explored in the questionnaire. In particular, in many organisations much IT activity occurs within individual business units and may not be visible to the CIO role.

Comparison with Government Use of ICT Survey 2006

This survey has some questions in common with Statistics New Zealand's Government Use of ICT Survey 2006. Nevertheless, the populations are different and (with the exception of two expenditure questions) imputation for non-response has not been performed in this survey yet was in 2006. Therefore, caution should be taken in comparing and inferring change over time. The major differences are shown in this table:

Note that it is possible to perform some limited longitudinal comparison of operating and capital expenditure for the core population in common with Statistics New Zealand's Government Use of ICT Survey 2006. The combined population for the categories 'Central government' plus 'Crown research institutes' in the 2006 survey correspond approximately to the population for 'Central government' in this work.