Data quality, concepts and methodology: Data quality, concepts and methodology

View the most recent version.

The universe is comprised of all members of the Association of Universities and Colleges of Canada (AUCC), as well as the university-affiliated research hospitals. The latter includes some members of the Association of Canadian Teaching Hospitals (ACTH) and some other hospitals reporting R&D activity on the Annual Hospital Survey.

This survey is a census with a cross-sectional design. Data are collected for all units of the target population, therefore no sampling is done.

Surveys are subject to certain types of errors: coverage, non-response, interpretation and processing errors. The methodology of this survey has been designed to minimize errors and to reduce their potential impact.

Because of significant variations observed between universities and hospitals on intellectual property issues, data from this survey are not weighted or imputed. Results are presented as gross results and only cover the respondent population.

Both micro and macro-editing are done. As questionnaires are returned, the information is captured onto a screen containing the previous response. This forces a comparison of the previous and current responses. As well, internal inconsistencies are noted and followed up by telephone.

The data are also compared against external public sources of information, such as university websites, the Association of University Technology Managers (AUTM) survey, the Canadian Association of University Business offices (CAUBO), annual reports, press releases and conference presentations.

There is no imputation of missing data for this survey except when additional information is available as a basis for imputation. One of the most common cases of missing data is where information is provided in aggregate form only and not broken down into the categories requested. This was only observed for a few questions. For some questions, historical data was used to break down the reported totals, while for others, the total was placed in the “other” category or in a new “unclassified” category. If there is absolutely no information available, the field is left blank and no imputation is done.

There were minor revisions to the survey questionnaire in 2008 in order to address data quality issues and response burden. Some of the detail requested was dropped while additional clarifications and improvements to the flow of certain questions were made where needed.

An analysis was carried out to ensure the quality of the data obtained in 2008. This analysis was based mainly on a comparison of the total estimates of certain variables of interest between 2006, 2007 and 2008. It also compared the total estimates for all respondents with those for regular respondents, i.e., those who responded the last three times. In short, the analysis provided for ensuring that the most important contributors to the previous surveys responded to the 2008 survey.

Response rates for 2008:

1. 142 questionnaires mailed out
2. 125 responding institutions (this includes combined reports)

Further details on the methodology of the survey can be found at: http://www.statcan.ca/english/sdds/index.htm (Survey record no. 4222).