Are we measuring loneliness in the same way in men and women in the general population and in the older population? Two studies of measurement equivalence

PLOS ONE RESEARCH ARTICLE Are we measuring loneliness in the same way in men and women in the general population and in the older population? Two studies of measurement equivalence Thomas V. Pollet ID1*, Alexandra Thompson1, Connor Malcolm1, Kristofor McCarty1, Tamsin K. Saxton1, Sam G. B. Roberts2 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 1 Dept. of Psychology, Northumbria University, Newcastle upon Tyne, United Kingdom, 2 Dept. of Psychology, Liverpool John Moores University, Liverpool, United Kingdom * Abstract OPEN ACCESS Citation: Pollet TV, Thompson A, Malcolm C, McCarty K, Saxton TK, Roberts SGB (2022) Are we measuring loneliness in the same way in men and women in the general population and in the older population? Two studies of measurement equivalence. PLoS ONE 17(12): e0266167. https:// doi.org/10.1371/journal.pone.0266167 Editor: Ghaffar Ali, Shenzhen University, CHINA Received: June 29, 2021 Accepted: March 16, 2022 Published: December 29, 2022 Copyright: © 2022 Pollet et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: https://osf.io/3puht/. Funding: Sample 1 reported in the paper was supported by Facebook Research through an Instagram Research Award to TVP and SGBR (https://research.fb.com/blog/2018/12/ announcing-the-winners-of-the-instagramresearch-awards-for-social-technologies/). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Sample 2 reported in the paper was supported by the Dept. of Psychology at Background High levels of loneliness are associated with negative health outcomes and there are several different types of interventions targeted at reducing feelings of loneliness. It is therefore important to accurately measure loneliness. A key unresolved debate in the conceptualisation and measurement of loneliness is whether it has a unidimensional or multidimensional structure. The aim of this study was to examine the dimensional structure of the widely used UCLA Loneliness Scale and establish whether this factorial structure is equivalent in men and women. Methods and sample Two online UK-based samples were recruited using Prolific. The participants in Study 1 were 492 adults, selected to be nationally representative by age and gender, whilst the participants in Study 2 were 290 older adults aged over 64. In both studies, participants completed the UCLA Loneliness Scale (Version 3) as part of a larger project. Results In both studies, the best fitting model was one with three factors corresponding to ‘Isolation,’ ‘Relational Connectedness,’ and ‘Collective Connectedness.’ A unidimensional single factor model was a substantially worse fit in both studies. In both studies, there were no meaningful differences between men and women in any of the three factors, suggesting measurement invariance across genders. Conclusion These results are consistent with previous research in supporting a multidimensional, three factor structure to the UCLA scale, rather than a unidimensional structure. Further, the measurement invariance across genders suggests that the UCLA scale can be used to compare PLOS ONE | https://doi.org/10.1371/journal.pone.0266167 December 29, 2022 1 / 16 PLOS ONE Northumbria University to AT and TVP. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. There was no additional external funding received for this study. Competing interests: We have read the journal’s policy and the authors of this manuscript have the following competing interests: TVP/SGBR have been supported by Facebook Research via an Instagram Research Award. This does not alter our adherence to PLOS ONE policies on sharing data and materials. Loneliness equivalence levels of loneliness across men and women. Overall the results suggest that loneliness has different facets and thus future research should consider treating the UCLA loneliness scale as a multidimensional scale, or using other scales which are designed to measure the different aspects of loneliness. Introduction Throughout their evolutionary history, humans have lived in social groups and depended on forming long-term relationships with others for survival [1, 2]. Thus, humans have a basic and universal need to form strong, stable interpersonal relationships with others—a ‘need to belong’ [3]. When this need is unmet and people feel disconnected from others, this lack of meaningful social relationships has a profound impact on physical and mental health [4]. Loneliness is defined as an unpleasant subjective state arising from a mismatch between the quantity and quality of social relationships we have and those we would like to have [5]. A large body of research has demonstrated that high levels of loneliness are associated with negative health outcomes in relation to both morbidity and mortality (reviews in [6–11]). Loneliness also has a key place on the social and political agenda in countries such as the United Kingdom [12], and the pandemic has further exacerbated the need for policy intervention on this front [13]. It is thus important that we can reliably measure loneliness, in order to accurately measure its prevalence over time, in different parts of the population and to evaluate whether interventions to combat loneliness are effective [14, 15]. Over the past five decades, many scales have been developed to measure loneliness, including: the Differential Loneliness Scale [16], the Loneliness Rating Scale [17], the De Jong-Gierveld Loneliness scale [18], and the Social and Emotional Loneliness Scale for Adults (SELSA, [19]). One of the most commonly used measures is the UCLA Loneliness Scale, which has appeared in first [20], second [21] and third [22] versions, and its short form adaptations (e.g., [23–25]). The UK Office for National Statistics has recommended that future UK national surveys of loneliness use three items from the UCLA scale [26]. The scale has been translated into many languages (e.g., Russian: [27]) and validated in many countries (e.g., Italy: [28]; Zimbabwe: [29]). UCLA loneliness factor structure: One, two, or three factors? A key unresolved debate in the conceptualisation and measurement of loneliness is whether it has a unidimensional or multidimensional structure [20–22, 30–32]. From its inception, the UCLA Loneliness Scale was argued to tap into a unidimensional construct [20–22], with deficits in a variety of relationships producing the same underlying state. Indeed, many studies have found evidence for a unidimensional structure (e.g., [33, 34]), or for a unidimensional structure with a subsidiary factor accounting for methodological effects due to wording [35]. Some such stud (...truncated)