Cap and trade program design options f9
Though obtaining clean data from multiple organizations with appropriate measurement controls is difficult, research that pools data from multiple sites with adequate controls for various factors discussed above is needed to further validate our results. Our field study is based on primary data collected on commercial software projects of a leading vendor. A larger software product is likely to have several modules, leading to many possible interactions between the modules.
Hence, for a poor quality product, software vendors may incur substantial support costs to fix the problems reported by the customers. We find that adherence to the practices of certain development process areas as specified in the CMM framework is associated with higher conformance quality in software products. This indicates the direct negative association of size and quality not normalized for size.
Broadman and Johnson  report results on the perceived effect of the CMM process areas on the cost and quality outcomes in software projects. The SUR estimates of the parameters of the two equations are presented in fourth columns of Table 2A cap and trade program design options f9 2B respectively. As noted earlier, life-cycle productivity in our model aggregates effort in development and maintenance phases. Since the small sample properties of the endogeneity tests are not certain, in order to correct for any bias in the OLS estimates due to potential endogeneity of quality variable in the productivity model, we also treated the two equations 3 and 4 as simultaneous equations and estimated the parameters using 2SLS two stage least squares.
We estimated the model without normalizing the dependent variable quality for size i. Likewise, should the manager hire new programmers or invest in process improvements? The results of the quality equation identify several drivers of product quality. A proper control of these work products would lead to reduction in software errors due to incorrect versions of source files and documents. What are the effects of development resources on productivity and quality?
In other cap and trade program design options f9, it can be claimed that quality affects productivity and productivity affects quality. The details of our measurement of these constructs and the specific CMM process areas included are provided in the Appendix. However, since these variables are not dependent variables in our models, the OLS estimates of the model parameters are not affected. The inter-rater reliability index for this measure was 0. However, the specific impacts of these factors on life-cycle productivity and quality are still not clear.
Thus our quality measure is normalized for product size. Our empirical analysis is described in Section 6, the results in Section 7. Second, our study identifies several quality drivers in software products.
For ease cap and trade program design options f9 interpretation, we define our quality measure as thousand lines of code per customer reported defect. We also verified through confirmatory factor analysis that these sets of four process areas loaded on one factor. Since the software products in our sample were developed over a span of five years, we normalize the cost figures to constant dollars using an appropriate normalization table [PRICE, ]. The model addresses the research questions related to tradeoffs between life-cycle productivity and quality and the effects of resource deployment and process design. However, pooling data from multiple firms would need control for accounting standards for cost data and consistent measurement of product size, cost and quality.