ICS 03.120.30                          MSD 03

Superseding IS 8900 : 1978

Reaffirmed 2021

NATIONAL FOREWORD

This Indian Standard (Part 4) which is identical with ISO 16269-4 : 2010 ‘Statistical interpretation of data — Part 4: Detection and treatment of outliers’ issued by the International Organization for Standardization (ISO) was adopted by the Bureau of Indian Standards on the recommendation of the Statistical Methods for Quality and Reliability Sectional Committee and approval of the Management and Systems Division Council.

This standard (Part 4) is a part of IS/ISO 16269 under the general title ‘Statistical interpretation of data’. Other parts in this series are:

Part 6 Determination of statistical tolerance intervals

Part 7 Median — Estimation and confidence intervals

Part 8 Determination of prediction intervals

This part of the standard was originally published as IS 8900 : 1978 ‘Criteria for the rejection of outlying observations’. It has been decided to adopt the corresponding International Standard in order to harmonize it with the latest developments that have taken place at international level. This part of the standard contains all the provisions of and will supersede IS 8900 : 1978.

An outlying observation or an ‘outlier’ is one that appears to deviate markedly from the other observations of the sample in which it occurs. Identification of outliers is one of the oldest problems in interpreting data. An outlier may arise merely because of an extreme manifestation of the random variability inherent in the data or because of the non-random errors, such as, measurement error, sampling error, intentional under- or over-reporting of sampling results, incorrect recording, incorrect distributional or model assumptions of the data set, and rare observations, etc. If it is known that a mistake has occurred, the outlying observation must be rejected irrespective of its magnitude. If, however, only a suspicion exists, it may be desirable to determine whether such an observation may be rejected or whether it may be accepted as part of the normal variation expected.

Outliers can distort and reduce the information contained in the data source or generating mechanism. In the manufacturing industry, the existence of outliers will undermine the effectiveness of any process/ product design and quality control procedures. Possible outliers are not necessarily bad or erroneous. In some situations, an outlier may carry essential information and thus it should be identified for further study.

The study and detection of outliers from measurement processes leads to better understanding of the processes and proper data analysis that subsequently results in improved inferences.

This standard provides certain statistical criteria which would be useful for the identification of the outlier(s) in data from measurement processes. When so identified, necessary investigations may also be initiated wherever possible to find out the assignable causes which gave rise to the outlier(s). The other objectives of the statistical tests for locating the outlier(s) may be to:

a) Screen the data before statistical analysis;

b) Sound an alarm that outliers are present, thereby indicating the need for a closer study of the data generating process; and

c) Pinpoint the observations which may be of special interest first because they are extremes. The text of ISO Standard has been approved as suitable for publication as an Indian Standard without deviations. Certain conventions are, however, not identical to those used in Indian Standards. Attention is particularly drawn to the following:

a) Wherever the words ‘International Standard’ appear referring to this standard, they should be read as ‘Indian Standard’.

b) Comma (,) has been used as a decimal marker while in Indian Standards, the current practice is to use a point (.) as the decimal marker.

Annexes B, C, D and E form an integral part of this adopted standard. Annexes A and F are for information only.

In reporting the result of a test or analysis made in accordance with this standard, if the final value, observed or calculated, is to be rounded off, it shall be done in accordance with IS 2 : 1960 ‘Rules for rounding off numerical values (revised)’.