Automated scoring of polygraph data

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

The objective of automated scoring algorithms for polygraph data is to create reliable and statistically valid classification schemes minimizing both false positive and false negative rates. With increasing computing power and well developed statistical methods for classification we often launch analyses without much consideration for the quality of the datasets and the underlying assumptions of the data collection. In this paper we try to assess the validity of logistic regression when faced with a highly variable but small dataset of 149 real-life specific incident polygraph cases. The data exhibit enormous variability in the subject of investigation, format, structure, and administration, making them hard to standardize within an individual and across individuals. This makes it difficult to develop generalizable statistical procedures. We outline steps and detailed decisions required for the conversion of continuous polygraph readings into a set of features. With a relatively simple approach we obtain accuracy rates comparable to those reported by other more complex algorithms and manual scoring. Complexity underlying assessment and classification of examinee’s deceptiveness is evident in a number of models that account for different predictors giving similar results, typically “overfitting” with the increasing number of features. While computerized systems have the potential to reduce examiner variability and bias, the evidence that they have achieved this potential is meager at best.

Original languageEnglish (US)
Title of host publicationStatistical Data Mining and Knowledge Discovery
PublisherCRC Press
Pages135-154
Number of pages20
ISBN (Electronic)9780203497159
ISBN (Print)9781584883449
StatePublished - Jan 1 2003

Fingerprint

Logistics
Statistical methods
Scoring
Data collection
Incidents
Predictors
Classification schemes
Logistic regression
Overfitting

All Science Journal Classification (ASJC) codes

  • Economics, Econometrics and Finance(all)
  • Business, Management and Accounting(all)
  • Computer Science(all)

Cite this

Slavkovic, A. B. (2003). Automated scoring of polygraph data. In Statistical Data Mining and Knowledge Discovery (pp. 135-154). CRC Press.
Slavkovic, Aleksandra B. / Automated scoring of polygraph data. Statistical Data Mining and Knowledge Discovery. CRC Press, 2003. pp. 135-154
@inbook{61c32690fbd64efabfe9bf15c483149e,
title = "Automated scoring of polygraph data",
abstract = "The objective of automated scoring algorithms for polygraph data is to create reliable and statistically valid classification schemes minimizing both false positive and false negative rates. With increasing computing power and well developed statistical methods for classification we often launch analyses without much consideration for the quality of the datasets and the underlying assumptions of the data collection. In this paper we try to assess the validity of logistic regression when faced with a highly variable but small dataset of 149 real-life specific incident polygraph cases. The data exhibit enormous variability in the subject of investigation, format, structure, and administration, making them hard to standardize within an individual and across individuals. This makes it difficult to develop generalizable statistical procedures. We outline steps and detailed decisions required for the conversion of continuous polygraph readings into a set of features. With a relatively simple approach we obtain accuracy rates comparable to those reported by other more complex algorithms and manual scoring. Complexity underlying assessment and classification of examinee’s deceptiveness is evident in a number of models that account for different predictors giving similar results, typically “overfitting” with the increasing number of features. While computerized systems have the potential to reduce examiner variability and bias, the evidence that they have achieved this potential is meager at best.",
author = "Slavkovic, {Aleksandra B.}",
year = "2003",
month = "1",
day = "1",
language = "English (US)",
isbn = "9781584883449",
pages = "135--154",
booktitle = "Statistical Data Mining and Knowledge Discovery",
publisher = "CRC Press",

}

Slavkovic, AB 2003, Automated scoring of polygraph data. in Statistical Data Mining and Knowledge Discovery. CRC Press, pp. 135-154.

Automated scoring of polygraph data. / Slavkovic, Aleksandra B.

Statistical Data Mining and Knowledge Discovery. CRC Press, 2003. p. 135-154.

Research output: Chapter in Book/Report/Conference proceedingChapter

TY - CHAP

T1 - Automated scoring of polygraph data

AU - Slavkovic, Aleksandra B.

PY - 2003/1/1

Y1 - 2003/1/1

N2 - The objective of automated scoring algorithms for polygraph data is to create reliable and statistically valid classification schemes minimizing both false positive and false negative rates. With increasing computing power and well developed statistical methods for classification we often launch analyses without much consideration for the quality of the datasets and the underlying assumptions of the data collection. In this paper we try to assess the validity of logistic regression when faced with a highly variable but small dataset of 149 real-life specific incident polygraph cases. The data exhibit enormous variability in the subject of investigation, format, structure, and administration, making them hard to standardize within an individual and across individuals. This makes it difficult to develop generalizable statistical procedures. We outline steps and detailed decisions required for the conversion of continuous polygraph readings into a set of features. With a relatively simple approach we obtain accuracy rates comparable to those reported by other more complex algorithms and manual scoring. Complexity underlying assessment and classification of examinee’s deceptiveness is evident in a number of models that account for different predictors giving similar results, typically “overfitting” with the increasing number of features. While computerized systems have the potential to reduce examiner variability and bias, the evidence that they have achieved this potential is meager at best.

AB - The objective of automated scoring algorithms for polygraph data is to create reliable and statistically valid classification schemes minimizing both false positive and false negative rates. With increasing computing power and well developed statistical methods for classification we often launch analyses without much consideration for the quality of the datasets and the underlying assumptions of the data collection. In this paper we try to assess the validity of logistic regression when faced with a highly variable but small dataset of 149 real-life specific incident polygraph cases. The data exhibit enormous variability in the subject of investigation, format, structure, and administration, making them hard to standardize within an individual and across individuals. This makes it difficult to develop generalizable statistical procedures. We outline steps and detailed decisions required for the conversion of continuous polygraph readings into a set of features. With a relatively simple approach we obtain accuracy rates comparable to those reported by other more complex algorithms and manual scoring. Complexity underlying assessment and classification of examinee’s deceptiveness is evident in a number of models that account for different predictors giving similar results, typically “overfitting” with the increasing number of features. While computerized systems have the potential to reduce examiner variability and bias, the evidence that they have achieved this potential is meager at best.

UR - http://www.scopus.com/inward/record.url?scp=85057414894&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057414894&partnerID=8YFLogxK

M3 - Chapter

AN - SCOPUS:85057414894

SN - 9781584883449

SP - 135

EP - 154

BT - Statistical Data Mining and Knowledge Discovery

PB - CRC Press

ER -

Slavkovic AB. Automated scoring of polygraph data. In Statistical Data Mining and Knowledge Discovery. CRC Press. 2003. p. 135-154