Using the K-Means Node Clustering Method and ROC Curve Analysis to Define Cut-Off Scores for the Caregiving System Scale

Daiana Colledani; Mario Mikulincer; Phillip R. Shaver; Anna Maria Meneghini

doi:10.1002/ijop.70038

Using the K-Means Node Clustering Method and ROC Curve Analysis to Define Cut-Off Scores for the Caregiving System Scale

Daiana Colledani^*
, Mario Mikulincer
, Phillip R. Shaver
, Anna Maria Meneghini

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

This study was conducted to establish cut-off scores for the subscales of the Caregiving System Scale (CSS). Two samples of Italian adults (N's = 682 and 227) completed the CSS. In the first sample, K-means node clustering and ROC curve analyses were conducted. Four caregiving profiles were identified and cut-off scores were calculated for classifying participants into these profiles. In the second sample, participants completed the CSS and the Attachment Style Questionnaire. Findings supported the presence of unique CSS profiles and meaningful connections between them and attachment orientations. This work offers a method for determining cut-off scores when gold-standard measures needed to run ROC curve analyses are unavailable.

Original language	English
Article number	e70038
Journal	International Journal of Psychology
Volume	60
Issue number	2
DOIs	https://doi.org/10.1002/ijop.70038
State	Published - Apr 2025
Externally published	Yes

Bibliographical note

Publisher Copyright:
© 2025 The Author(s). International Journal of Psychology published by John Wiley & Sons Ltd on behalf of International Union of Psychological Science.

Keywords

CSS
K-means node
ROC curves
caregiving profiles
cut-off scores

Access to Document

10.1002/ijop.70038

Cite this

@article{20a31098e4714ab5ae61145ff03af607,

title = "Using the K-Means Node Clustering Method and ROC Curve Analysis to Define Cut-Off Scores for the Caregiving System Scale",

abstract = "This study was conducted to establish cut-off scores for the subscales of the Caregiving System Scale (CSS). Two samples of Italian adults (N's = 682 and 227) completed the CSS. In the first sample, K-means node clustering and ROC curve analyses were conducted. Four caregiving profiles were identified and cut-off scores were calculated for classifying participants into these profiles. In the second sample, participants completed the CSS and the Attachment Style Questionnaire. Findings supported the presence of unique CSS profiles and meaningful connections between them and attachment orientations. This work offers a method for determining cut-off scores when gold-standard measures needed to run ROC curve analyses are unavailable.",

keywords = "CSS, K-means node, ROC curves, caregiving profiles, cut-off scores",

author = "Daiana Colledani and Mario Mikulincer and Shaver, \{Phillip R.\} and Meneghini, \{Anna Maria\}",

note = "Publisher Copyright: {\textcopyright} 2025 The Author(s). International Journal of Psychology published by John Wiley \& Sons Ltd on behalf of International Union of Psychological Science.",

year = "2025",

month = apr,

doi = "10.1002/ijop.70038",

language = "אנגלית",

volume = "60",

journal = "International Journal of Psychology",

issn = "0020-7594",

publisher = "John Wiley and Sons Inc.",

number = "2",

}

TY - JOUR

T1 - Using the K-Means Node Clustering Method and ROC Curve Analysis to Define Cut-Off Scores for the Caregiving System Scale

AU - Colledani, Daiana

AU - Mikulincer, Mario

AU - Shaver, Phillip R.

AU - Meneghini, Anna Maria

PY - 2025/4

Y1 - 2025/4

N2 - This study was conducted to establish cut-off scores for the subscales of the Caregiving System Scale (CSS). Two samples of Italian adults (N's = 682 and 227) completed the CSS. In the first sample, K-means node clustering and ROC curve analyses were conducted. Four caregiving profiles were identified and cut-off scores were calculated for classifying participants into these profiles. In the second sample, participants completed the CSS and the Attachment Style Questionnaire. Findings supported the presence of unique CSS profiles and meaningful connections between them and attachment orientations. This work offers a method for determining cut-off scores when gold-standard measures needed to run ROC curve analyses are unavailable.

AB - This study was conducted to establish cut-off scores for the subscales of the Caregiving System Scale (CSS). Two samples of Italian adults (N's = 682 and 227) completed the CSS. In the first sample, K-means node clustering and ROC curve analyses were conducted. Four caregiving profiles were identified and cut-off scores were calculated for classifying participants into these profiles. In the second sample, participants completed the CSS and the Attachment Style Questionnaire. Findings supported the presence of unique CSS profiles and meaningful connections between them and attachment orientations. This work offers a method for determining cut-off scores when gold-standard measures needed to run ROC curve analyses are unavailable.

KW - CSS

KW - K-means node

KW - ROC curves

KW - caregiving profiles

KW - cut-off scores

UR - https://www.scopus.com/pages/publications/105000000365

U2 - 10.1002/ijop.70038

DO - 10.1002/ijop.70038

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???

C2 - 40059112

AN - SCOPUS:105000000365

SN - 0020-7594

VL - 60

JO - International Journal of Psychology

JF - International Journal of Psychology

IS - 2

M1 - e70038

ER -

Using the K-Means Node Clustering Method and ROC Curve Analysis to Define Cut-Off Scores for the Caregiving System Scale

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this