The 10-item version of the University of California Los Angeles Loneliness Scale version 3 (The UCLA-LS3-J SF-10 ) was used to evaluate loneliness. Each of the 10 items of the UCLA-LS3-J SF-10 [20, 28] has 4 choices: (1) never, (2) rarely, (3) sometimes, and (4) always. The total score ranges from 10 to 40, with higher scores indicating a higher level of loneliness. The reliability and validity of this scale have been established by the authors  after the second author (ET) obtained permission to translate the UCLA-LS3-J from its original author (Dr. Russell). The Cronbach’s alpha of the Japanese translated version of the scale was 0.888 .
Social isolation (social network size)
The Japanese version of the Lubben Social Network Scale (LSNS-6 [29, 30]) was used to evaluate social networks. This scale was chosen because it allows for comparisons between the size and quality of social networks. The LSNS-6 comprises 6 items each with 6 response options from 0 (nobody) to 5 (more than 9 persons) that evaluated the mother’s social network in her family (3 items) and among her friends (3 items). The total score ranges from 0 to 30, with higher scores indicating a larger social network. A score of less than 12 marks the cutoff point for social isolation. Based on the LSNS-6 cut-off value of 12 points, the LSNS-6 was converted to a binary variable where 1 = isolated and 0 = non-isolated. The reliability and validity of the Japanese version have been established [29, 30]. Cronbach’s alpha for the LSNS-6 was 0.82 in the Japanese translated version . This scale is not under license but permission has been obtained for its use.
Several independent variables were explored as being factors that were potentially related to loneliness from an ecological systems model . These variables were selected based on literature reviews of previous studies [4,5,6,7,8,9, 15,16,17,18,19,20,21] and 3 focused group interviews with 41 professionals or community volunteers, and 31 one-to-one interviews with mothers raising children less than 3 years of age.
Demographic data, such as age, family structure (nuclear family; parents and children, extended family, and single-parent family), employment status (housewife, full-time office worker, part-time worker, contract/temporary worker, and self-employed), number of years of residence (less than 1 year, more than 1 year and less than 5 years, more than 5 years and less than 10 years, over 10 years), age of children eligible for health screening, number of children, and employment status of spouse/partner were collected.
Individual and family factors comprised of parenting concerns, and the health behaviors of caregivers. The number of parenting concerns was created from 12 dichotomously rated items (1 = yes and 0 = no), reflecting stressors on parenting and life (e.g., children’s toileting problems, how to play with and pamper their child, financial worries, and balancing work and childcare). The health behaviors of mothers were measured using the Good Health Habits Scale developed by Breslow  that were dichotomously rated items (1 = yes and 0 = no) regarding the respondent’s health behaviors (e.g., sleep, smoking, diet, exercise, alcohol use).
Partners’ supportive behaviors for household and childcare were assessed using the partners’ supportive behaviors for household and childcare scale . This scale comprises 29 items that were rated on a 4-point scale from 1 (do not at all) to 4 (do often) that evaluated the supportive behaviors of the mothers’ partners in terms of emotional support (14 items), and household and childcare (15 items). The scale score is calculated by dividing the sum of the crude scores of all items by the number of items, with higher scores indicating more supportive behaviors by partners. The reliability and validity of this scale have been established . This scale is not under license but permission has been obtained for its use.
Community factors comprised of people consulted about parenting, interest in and desire to interact with the neighbors, use of community childcare resources, and recognition of the community (community commitment). The number of people that the mother consulted about parenting was created using 12 dichotomous items (1 = yes and 0 = no) reflecting parenting social support (e.g., consulted with parents, friends, family doctor, child welfare commissioner). One dichotomous item was used to measure interest in neighbors and one to assess the desire to interact with neighbors. The use of community childcare resources was created from 5 dichotomous items (1 = yes and 0 = no) that measured resources (e.g., community child rearing support center, comprehensive community support center, volunteer-sponsored park play, parent–child group).
The Community Commitment Scale  was used to evaluate recognition of the community. This scale comprises 8 items rated on a 4-point Likert-type scale from 0 (strongly disagree) to 3 (strongly agree), each evaluating the recognition of the community as an opportunity for socializing (4 items) and belonging (4 items). The total score ranges from 0 to 24, with higher scores indicating a higher community commitment. The reliability and validity of this scale have been established . The Cronbach’s alpha for the CCS was 0.78 in the original scale . This scale is not under license but permission has been obtained for its use.
Based on the LSNS-6 cut-off value of 12 points, the participants were classified into two groups, which were then compared. After calculating the descriptive statistics for each item for all the participants and comparing the two groups using chi-square tests and independent t-tests, we analyzed the data for differences in loneliness between the demographic variables using t-tests and one-way analysis of variance (ANOVA). Following the ANOVA, Tukey’s multiple comparisons were performed.
Next, correlations between loneliness and items that had ordinal and continuous scores were examined using Pearson’s correlation coefficients. Among the items with significant correlations (p < 0.05), multiple linear regression analyses were performed with loneliness as the dependent variable and the selected items as independent variables, while taking into account multicollinearity with social isolation as a mediator. Age was set as the adjustment variable. The significance level was set at less than 5% on both sides. IBM SPSS version 26 statistical software (SPSS Inc., Chicago, IL, USA) was used to perform all statistical analyses.