Chapter 4 Data Collection and Normalization Flashcards
___________, which is unexplained variability, always exists in data.
Statistical noise
_______________, or the conversion of raw data sets into normalized data sets
functional transformation
______________ is captured in a specified type of dollars and hours and is associated with a particular activity (labor) or product (material). Resources associated with some activities are captured in cost pools such as Overhead (OH) or General and Administrative (G&A).
Cost data (Resource data)
_____________ derives from requirements or physical characteristics of systems and may drive cost.
Technical Data
________________ includes the program parameters that explain and drive cost.
Programmatic data
____________ is obtained from the original source. It is data collected from the contractor facility on-site collections, government reports, test centers, it is unaltered or unchanged and represents actual, historical data. Examples include Bills Of Materials (BOM), documented test results, and documented resource hours to accomplish a task.
Primary Data
___________ is based on primary data. It is derived, sanitized for classification and proprietary purposes, or changed in some way from the original source data. examples include documented cost estimates, factors and factor books, studies and white papers, contractor cost report summaries
Secondary Data
____________ is preferred for data collection. It is usually objective and is measurable in either the physical or cost accounting sense. An example of subjective _____________ data is numeric risk scores provided by engineers.
Quantitative, or numeric, data
_________ represents the rank order of the data (e.g., 1st, 2nd, 3rd). The only valid comparisons are greater than or less than in magnitude or scope.
Ordinal
________ represents relative scaling (e.g., the year, so that from 1995 to 2005 is twice as long as from 1995 to 2000).
Interval
________ supports absolute comparisons (e.g., two million dollars ($2M) is twice as much money as ($1M) – assuming the amounts have been normalized to the same year and type of currency).
Ratio
______________ describes a quality of the program or system, usually in categorical terms referred to as nominal.
Qualitative data
____________ is typically subjective but can be objective (externally verifiable). If the data requires expert judgment and all of the experts would make the same determination, then the data is objective. If those experts disagree, then the data is subjective.
Qualitative data
__________ is usually quantitative and, like primary data, preferred. It includes actual counts collected through a formal data collection process or derived from other quantitative data (e.g., staff hours, Extended Source Lines of Code (ESLOC), Function Points (FPs), test items, documented errors, or end items). It can sometimes be qualitative (e.g., U.S. vs. Soviet fighters).
Objective Data
______________ is based on individual or group feelings or understanding about a particular condition or characteristic pertinent to a system. It data tends to be qualitative and typically provides information needed to interpret or validate objective data. It is valuable in helping cost estimators extrapolate information (i.e., data) from one generation of equipment to the next. Examples include complexity, requirements stability, level of difficulty, and degree of new technology involved.
Subjective Data
