Research and Development in Canadian Industry Intended to Directly Benefit Developing Countries

Detailed information for 2004

Status:

Inactive

Frequency:

One Time

Record number:

5097

The statistical information collected by this survey is useful for better understanding the measurement of the effort invested by Canada in research and development directed toward developing countries. The data collected by this survey contribute to the federal research and development strategy and are covered in an agreement signed with the Office of the National Science Advisor.

Data release - July 19, 2006

Description

This pilot survey will provide information about industrial research and development (R&D) expenditures towards developing countries by type of relevant activities. The population covered by this survey is the same as the Research and Development in Canadian Industry survey, that is all Canadian R&D business enterprises that perform or fund $1 million or more of R&D.

The data collected from this survey will be used for economic analysis related to research and development.

Reference period: Fiscal period

Collection period: From October to February following the reference year

Subjects

  • Research and development
  • Science and technology

Data sources and methodology

Target population

The target population consists of all business enterprises that perform or fund 1 million dollars or more of research and development.

The target population was 1249 units.

Instrument design

The form was designed and developed according to the Organisation for Economic Co-operation and Development (OECD) guidelines that appear in the Frascati Manual 2002 and takes into account the inputs in the research and development.

The questionnaire was designed and tested with the assistance of the Questionnaire Design Resource Center (QDRC), between July 1, 2005 and September 28, 2005. Subsequent to questionnaire testing in Ottawa and Montréal (between August 15, 2005 and September 2, 2005), revisions were made in order to produce the final version.

Sampling

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.

Data sources

Responding to this survey is mandatory.

Data are collected directly from survey respondents.

This mail out survey targeted all Canadian business enterprises that performed or funded $1 million or more of R&D. The survey was sent out by mail and mailed back within twenty days following receipt.

All non-respondent business enterprises were contacted by telephone from September to November 2005 to negotiate reporting options and make special arrangements, which could include allowing mandatory partial response.

View the Questionnaire(s) and reporting guide(s) .

Error detection

The data are corrected to guarantee their logic and internal consistency and are analysed to determine trends and validity.

We compared the data collected in this survey with data from the Research and Development in Canadian Industry (RDCI) survey in order to ensure the consistency, quality and validity of the data.

Imputation

Detection of errors (e.g., a character in a numeric field) was done primarily in SAS, but also in Microsoft Excel, given the few records that were processed. The strategy for edit and for deterministic imputation (in obvious cases) was developed between methodologists and subject matter experts so as to minimize imputation.

Imputation was done to fill empty fields in cases of partial non-response. Weights were adjusted to compensate for cases of total non-response. Imputation was done mainly using the hot deck and nearest neighbour methods except for some fields where imputation by the mean or ratio proved more appropriate.

Imputation was programmed in SAS. It is not manual in the sense that the values were not imputed purposively but instead randomly (hot deck, nearest neighbour, by the mean). The work of finding values to impute was done in Microsoft Excel to speed up the imputation process.

The variable TIE (total intra-mural expenditures) was used for imputation. Among other things, as an external information source, it served as a measure of distance in the nearest neighbour imputation method. For edit purposes, it served among other things to determine that a unit responded in dollars rather than in thousands of dollars by comparing with the value in Question 2.

Telephone follow-up was used to obtain more respondents and check the status (out of scope, merged, etc.) of some companies that did not return the questionnaire.

Units which answered No on the pre-contact card (no research and development directed toward less fortunate countries) and which did not respond to the questionnaire were imputed No to Question 1 (Did you have research and development expenditures directed towards developing countries?); this seems a valid assumption since developing countries can be interpreted broadly and thus include less fortunate countries.

Estimation

Units were stratified using Generalized Sampling System 4.3, developed at Statistics Canada. Variables from external sources, such as total intramural expenditures (research and development from the RDCI survey) were used with the cumulative function of the root of f to form strata. Two strata were thus created: TIE d $34 million and TIE > $34 million. A unit with a TIE value that was extreme in relation to the other units was then separated into a third stratum.

Two sets of weights were calculated, one for section 1 (research and development expenditures) and one for the rest of the questionnaire. With deterministic imputation, it was possible to use the units that answered No on the pre-contact card (no research and development directed toward less fortunate countries), and hence No to Question 1, for the estimation of Section 1 of the questionnaire. This helped to increase the number of respondents for this section and thus reduce weights and the variance. However, these units cannot be used for estimating the other sections of the questionnaire, since it would have been necessary to do total imputation because they did not receive the questionnaire.

A complete document on the estimation methodology is available in the link below.

Quality evaluation

This is a first experience with this survey. There is very little information from other surveys at Statistics Canada or elsewhere on which we can make comparisons.

The information concerning research and development from the RDCI survey may be compared to Question 2 (Please indicate total research and development expenditures performed in Canada by your enterprise). However, owing to massive imputation of this field for the units that answered No to Question 1 (Did you have research and development expenditures directed towards developing countries?), comparisons should be confined to units that answered Yes to Question 1. The answers reported for Question 2 for units that answered Yes to Question 1 are consistent with those from the RDCI survey.

We are hypothesizing that there is no bias, since we are dealing with a census of the population studied. The results of the survey tell us that units doing research and development directed toward developing countries constitute a rare population. We believe that there is no bias related to the coverage of this population for the overall study population, which consists of all business enterprises that perform or fund $1 million or more of R&D.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects which could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.

In order to prevent disclosure of individual respondents many industries must be grouped together to provide sufficient observations for dissemination.

Revisions and seasonal adjustment

This methodology type does not apply to this statistical program.

Data accuracy

One of the problems in a survey of this type is to ensure that the quality of the data is satisfactory. It cannot be expected that all business enterprises performing R&D expenditures directed towards developing countries will be covered by the target sample. Business enterprises which perform or fund less than $1 million of R&D annually were deliberately excluded to reduce respondent burden.

In addition, developing countries is a term subject to inconsistencies. Thus, the data, although reasonably accurate, cannot be regarded as precise.

The response rate for the Research and Development in Canadian Industry Intended to Directly Benefit Developing Countries Survey was 63%.

Date modified: