Methods for trend analysis: Examples with problem/failure data

Statistics are emphasized as an important role in quality control and reliability. Consequently, Trend Analysis Techniques recommended a variety of statistical methodologies that could be applied to time series data. The major goal of the working handbook, using data from the MSFC Problem Assessment System, is to illustrate some of the techniques in the NASA standard, some different techniques, and to notice patterns of data. Techniques for trend estimation used are: regression (exponential, power, reciprocal, straight line) and Kendall's rank correlation coefficient. The important details of a statistical strategy for estimating a trend component are covered in the examples. However, careful analysis and interpretation is necessary because of small samples and frequent zero problem reports in a given time period. Further investigations to deal with these issues are being conducted