We compare the performances of balances of opinion to those of indicators introduced by Mitchell, Smith and Weale for the one-quarter forecasting of the manufacturing production growth rate. The sources used are the Business Tendency Survey in industry and the quarterly accounts published by INSEE. The indicators relate to answers given by the production units questioned about their past and expected production Unlike the balances of opinion, the indicators proposed by Mitchell, Smith and Weale take into account the heterogeneity of the response behaviours of the entrepreneurs taking part in the Business Tendency Survey. The responses of entrepreneurs which are the most tightly linked to the overall fluctuations of manufactured production contribute to the variability of these indicators to a larger extent than the responses of the other surveyed. Does this specific feature of these indicators enable the latter to perform better in short-term forecasting than more classic indicators, such as the balance of opinion? The application of Mitchell, Smith and Weale to British and German data seems to suggest that this is the case, but their application to Swedish and Portuguese data suggests not. In our study carried out using French data, the predictive performances of the Mitchell, Smith and Weale indicators turn out to be inferior or, at best, equivalent to those of the balances of opinion, depending on the models used. This result seems robust due to both the large size of the French panel and the evaluation method used, the predictive qualities of the indicators being tested outside of their estimation period.