Examining biases in measures of firm innovation
Melbourne Institute Working Paper No. 10/04
Date: May 2004
A challenge for applied studies of innovation is to find quality-adjusted and unbiased measures for the extent and type of innovative activity in firms. This paper considers, in the light of the uses for these measures, how well we expect these common indicators quantify innovative services and how well they actually correlate using data from a sample of 641 companies. The results suggest that while the correlations between indicators are positive, they are quite small. In addition, our findings show that compared with survey measures, the accounting and administrative measures have different biases with respect to firm size, industry sector and innovation type. This should be born in mind when using these measures to depict trends in innovative activity and reveal relationships between firm activities.