3分钟弄懂CMMI2.0基准评估的抽样规则

Summary of CMMI 2.0 vs CMMI 1.3 Assessment Approach Changes

The transition from CMMI 1.3 to CMMI 2.0 brought significant changes to the assessment methodology, particularly in sampling rules. The new sampling rules in CMMI 2.0 demand wider project coverage without necessitating full coverage, distributing the assessment workload across multiple projects. This ensures that an enterprise's process system is genuinely implemented throughout the company. Unlike previous versions where assessors and sponsors determined the projects to be evaluated, CMMI 2.0 employs a random sampling system to minimize human intervention and ensure representativeness.

Eligible Projects for Evaluation and Sampling

Projects that can be included in the evaluation and sampling process extend to, but are not limited to, those that are currently ongoing, as well as those completed or shelved within a reasonable timeframe before the assessment phase, provided they utilized current processes and have verifiable records. Any excluded projects must be explicitly listed with reasons for their exclusion, and the assessment results will not apply to these excluded projects.

Determination of Sample Size

The number of projects sampled is determined by the total number of projects available for sampling within the organization. The minimum number of sampled projects per Process Area (PA) ranges from one for 1-10 available projects, to an additional sampled project for every 200 additional projects beyond 600. The exact number of projects covered is determined by the system, and the specific algorithm is not disclosed.

Project Classification for Sampling

Projects are classified for sampling to reflect differences in process implementation. At least five sampling factors should be considered: location, client type, size, organizational structure, and work type. Projects using identical processes are merged into a single category and are not considered distinct types.

Timing for Sampling

Sampling should occur no earlier than 60 days before the assessment commences. If the assessment doesn't start within 60 days, an application must be made to the CMMI Institute for an extension of up to an additional 30 days; otherwise, resampling is required.

Expressing Sampling Needs by Sponsor and LA

The sponsor and Lead Appraiser (LA) have limited influence on the random sampling process. They may designate one project that must be included before random sampling, request resampling if dissatisfied with the outcome (subject to CMMI Institute approval and time constraints), substitute a project within the same subgroup, and propose additional samples over the minimum required. Once the Random Sampling System (RGS) is accepted, the sponsor cannot request new samples unless a new assessment is applied for, while the LA can suggest additional sampling in agreement with the sponsor.