LXEdge Scoring Rules

Question Last Updated on: 2/5/2025

Scoring Rule

Grand Total Rule

The Grand Total Rule (GTR) is used with single-issue exams. RQs for single issue exams are assumed to have a shared source, or cause, of response variance. The GTR is the simplest of all decision rules. However, accuracy effect sizes using the GTR have been among the highest in published polygraph studies. The GTR uses all available recorded data to produce a single classification of the test as a whole. Procedurally, the GTR involves the summation of a grand total using all scores for all RQs and all recording sensors, and the comparison of this grand total with the numerical cutscores for deceptive and truthful classifications. Each RQ inherits the same result as the test as a whole. In other words, RQs are not interpreted or classified differently than the test as a whole.

Subtotal Score Rule

The Subtotal Score Rule (SSR) is a commonly used decision rule for multiple-issue polygraph exams. RQs for multiple-issue polygraph exams are assumed to have independent criterion variance. In practical terms, this means that different RQs describe different behaviors, and it is conceivable that a person may be involved in none, some, or all the behavioral target issues. In scientific terms, an assumption of independence means that there is no shared basis of response among any of the RQs ? that each response to each of the RQs is completely unaffected by any of the responses to all of the other RQs. Independence is a rigorous concept that can be difficult to attain. However, an assumption of independence, declared prior to data acquisition and data analysis, can broaden and increase the sensitivity of the polygraph test. However, the SSR, the use of multiple statistical classifications introduces multiplicity effects that are known to reduce the precision of the test result. Use of statistical corrections ? statistically corrected cut scores, or statistically corrected outcome probabilities ? can reduce, though does not completely eliminate, the effects of statistical multiplicity. Traditional cutscores do not involve the use of a statistical correction. ESS cutscores are obtained with a statistical correction for truthful classifications. This reduces the occurrence of inconclusive results among innocent/truthful persons. The ESS does not use a statistical Procedurally, the SSR requires the identification of the RQ with lowest numerical subtotal score, and the comparison of this score with numerical cutscores for deceptive and truthful classifications. The overall test result is then inherited from the RQ with the lowest numerical subtotal score. Next, the categorical results are obtained for the other RQ by comparing each remaining subtotal to the numerical cutscores for deception and truth-telling. Importantly, it is not permissible to classify both deception and truth-telling within a multiple-issue polygraph. Whenever any of the RQ subtotals is classified as deceptive all other RQ subtotals that cannot be classified as deceptive are classified as inconclusive or no-opinion. The GTR can be used with either statistical or traditional cutscores.

Two-stage Rule

The Two-stage Rule (TSR) can be thought of as the sequential use of both the GTR and SSR concepts. Stage 1 of the TSR is procedurally like the GTR. Subtotal scores are not used during Stage 1. The TSR is complete at Stage 1 if the test result can be classified as deceptive or truthful by comparing the grand total score to the numerical cutscores. Stage 2 is used only when a test cannot be classified as either deceptive or truthful during Stage 1. Stage 2 is somewhat like the SSR with the exception that truthful classifications are not made during this stage. Stage 2 requires the comparison of the lowest RQ subtotal score with the numerical cutscore for deceptive classifications using subtotal scores. A test result is inconclusive/no-opinion if the result cannot be classified as either deceptive or truthful at Stage 1 and cannot be classified as deceptive at Stage 2. Because of the sequential use of the grand total and subtotal scores, the GTR provides a procedural protection against multiplicity effects (inflated alpha) that can occur when using subtotal scores. Multiplicity effects with TSR are most obvious when using traditional cutscores. ESS cutscores are obtained using a statistical correction which also reduces statistical multiplicity error. Grandre occasions, when using the TSR, a grand total score may support a truthful classification while the lowest subtotal score supports a deceptive classification. In such cases, when using ESS cutscores it is permissible to classify the test result as deceptive based on the lowest subtotal score regardless of the grand total score, but ony if the subtotal cutscore includes a statistical correction for multiplicity. This approach is justified by the presence of unexpected or excessive variability in the test data, which undermines the reliability of conclusions based on the grand total score. When using the TSR, each RQ inherits the same result as the test as a whole. In other words, RQs are not interpreted or classified differently than the test as a whole. The TSR can be used with either traditional or statistical cutscores.

Federal Zone Rule

The Federal Zone Rule (FZR) is traditionally used with the Federal ZCT and Federal YouPhase test formats using traditional cutscores. RQs for these formats have a shared source of response variance ? a single known or alleged incident ? and therefore function as single-issue formats. The FZR involves several comparisons of the grand total and subtotal scores with the numerical cutscores for deception and truthtelling. Using the Federal ZCT format the cutscores are [-6, -3, +6/+1]. A test result is classified as deceptive if the grand total score is less than or equal to -6 or if any subtotal score is less than or equal to -3. A test result is classified as truthful if the grand total score is greater than or equal to +6 if all subtotal scores are equal to or greater than +1. Using the Federal YouPhase format the cutscores are [-4, -3, +4/+1]. A test result is classified as deceptive if the grand total score is less than or equal to -4 or if any subtotal score is less than or equal to -3. A test result is classified as truthful if the grand total score is greater than or equal to +4 if all subtotal scores are equal to or greater than +1. The FZR provides increased sensitivity to deception, comp though without management of statistical multiplicity error. When using the FZR each RQ inherits the same result as the test as a whole. In other words, RQs are not interpreted or classified differently than the test as a whole.

TES/DLST Rule

The TES/DLST Rule (TES) is described in the published studies on TES format that is used in a variety of multiple issue screening contexts. The TES rule uses traditional cutscores and applies the procedures from the FZR to the TES/DLST format. TES cutscores are [-4, -3, +4/+1]. A test result is classified as deceptive if the grand total score is less than or equal to -4 or if any subtotal score is less than or equal to -3. A test result is classified as truthful if the grand total score is greater than or equal to +4 but if all subtotal scores are equal to or greater than +1. Each RQ inherits the same result as the test as a whole. In other words, RQs are not interpreted or classified differently than the test as a whole.

Utah 4-Question Rule: The Utah 4-Question Rule (UT4) is used with the Utah 7-Position scoring method and the Utah 4-Question test format using UT4 cutscores [-6/-1, +6/+1, -3, +3]. A test result is classified as deceptive if the grand total score is less than or equal to -6 but only if all subtotal scores are less than or equal to -1. A test result is classified as truthful if the grand total score is greater than or equal to +6 but only if all subtotal scores are greater than or equal to +1. Subtotal scores are then used, and results are reported only for the individual RQs, if the test result is not yet classified as either deceptive or truthful using the grand total score. RQ subtotal scores are classified as deceptive if they are less than or equal to -3. If none of the RQ subtotals is classified as deceptive they may be classified as truth if they are greater than or equal to +3. Importantly, it is not permissible to classify both deception and truth-telling within a multiple-issue polygraph. Whenever any of the RQ subtotals is classified as deceptive all other RQ subtotals that cannot be classified as deceptive are classified as inconclusive or no-opinion.

Auto-Select

The Auto-Select option will select a polygraph decision rule ? either the TSR or the SSR ? depending on whether the RQs appear to be exemplary of a single-issue or multiple-issue test. The deduction of single-issue or multiple-issue RQs, is based on the observed variation among the RQ subtotal scores. When there is statistically significant variation between the RQ subtotal scores it is assumed that they may have independent criterion variance and the SSR is then selected. When the observed variation among the RQ subtotals scores is not statistically significant it is assumed that criterion variance of the RQs may be non-independent and the TSR is selected. Importantly, the Auto-Select rule does not select a decision rule based on the name or type of the CQT format and does not attempt to read the RQ question text. The Auto-Select rule is intended to select the TSR for single issue exams and is intended to select the SSR for multiple issue exams. However, the auto select rule can potentially select the SSR for a single issue exam if there is significant variation among the RQ subtotal scores. Selection of the SSR in this situation may be acceptable to when considering that the significant variation may signal a need for added caution or risk aversion. Also, the Auto-Select rule can potentially select the TSR for a multiple-issue exam if the RQ subtotal scores do not vary significantly ? such as when an examinee is actually truthful to all RQs, or when an examinee is actually deceptive to all RQs. Selection of the TSR in these situations may be acceptable as it will most often give the same result as the SSR. The Auto-Select rule is intended for use with ESS cutscores that include the use of statistical corrections to reduce multiplicity error.

Decide by Question List

This option will select a decision rule based on the name of the CQT test format (template). Question templates, consisting of multiple lists that all have exactly the same questions arranged in varying sequence, can be assigned a default decision rule upon creation or when editing/saving a template.

A potential problem here is that some test formats, such as the AFMGQT family of test formats, can be used as either single-issue or multiple-issue techniques. For this reason, each question template should be linked via meta-data to a default decision rule.

Categories

Polygraph (6)

Software (2)

Orders / Customer Service (2)

Website (6)

LXEdge (11)

LXSoftware (38)

Polygraph Pro Suite Software (1)

LXEdge Release Notes (7)

Polygraph Pro Suite Release Notes (4)

LXSoftware Release Notes (1)

Submit Question

Unable find an answer to the question that you looking for in our FAQ section? Submit your question directly to our customer service representatives.