Caregivers of 732 school-age children (6-12 years) from urban metropolitan areas in Canada (Halifax n = 300), Iceland (Reykjavik n = 252), and Thailand (Khon Kaen n = 180) responded to an invitation to participate. Of the 732 caregivers, 548 completed the survey. Of this, 183 were from Canada, 185 were from Iceland, and 180 were from Thailand. One participant was excluded from the Icelandic sample because of an excessive amount of missing data.
2.1.1. Recruitment and data collection
This study used samples from the general population. In Canada, participants were recruited using posters at various locations (eg, libraries, grocery stores, and web sites). A small number of participants were obtained through word of mouth. In Iceland and Thailand, participants were recruited through public elementary schools.
Convenience sampling was used to select participants at all sites. To decrease potential sample bias, a set of inclusion criteria of individuals was created and enforced at all sites.31 The inclusion criteria were (1) a child was between the ages of 6 and 12 years, (2) the caregiver was a resident of the chosen area in each country, (3) the caregiver understood the local language/dialect, and (4) only one caregiver per household could participate. Furthermore, the participants were explicitly informed that to participate, he or she needed to be a parent or a legal guardian of the child.
In Canada and Iceland, participants first read the study information letter and then indicated on the online survey “accept/consent” if they wished to continue. In Thailand, research assistants obtained written informed consent from the parent/legal guardians before they participated in the study.
All participants completed a survey package that consisted of 3 main questionnaires and demographic questions. Parents were asked to respond to the survey by reference to only one child. At all 3 sites, data were accrued through self-administered questionnaires. Participants in Canada and Iceland used online questionnaires, and Thai participants used a paper-based version. The decision to use the paper-based questionnaire in Thailand was due to local Khon Kaen researchers' assessment of the availability of the Internet.
Ethical approval was sought and received from all 3 study sites. In Canada, ethical approval was obtained through the IWK Health Centre Research Ethics Board. In Iceland, ethical approval was obtained from the National Bioethics Committee of Iceland. In Thailand, ethical approval was obtained from the Khon Kaen University Ethics Committee for Human Research.
2.2. Ecosocial contexts
The selection of ecosocial contexts was based on geography, economy, education levels, and cultural values (ie, individualism–collectivism), as well as authors' preestablished connections with local research communities.107
Halifax (population over 390,000) is the capital city of Nova Scotia (population over 921,000), one of Canada's (population over 33 million) 10 provinces. The majority of the Halifax population is Canadian-born (90.6%), speaks primarily English (90%), and affiliates as Christian (74%).95 On the individualism–collectivism spectrum, scholars describe and rate Canada as an individualistic-focused society.38,39,71,98 When comparing Canadian parents with parents from collectivistic cultures, Canadian parents consistently score higher on authoritative-parenting (ie, child-centered).14,59,84,97
Reykjavik (population over 213,000) is the capital region of Iceland (population over 330,000), where the vast majority of Iceland's population lives.96 The Icelandic nation is homogenous and cohesive; 93% of the population is Icelandic citizens who speak Icelandic; and 75% belong to the Evangelical Lutheran Church of Iceland (G. Hauksson, Statistics Iceland, personal communication, September 13, 2014). Iceland is a society with a unique combination of collectivistic and individualistic cultural values,69 but rated more individualistic than collectivistic.98 With respect to parenting styles, studies show that authoritative-parenting and neglectful-parenting are the most commonly used styles by Icelandic parents.1,5
2.2.3. Thailand-Khon Kaen
The city of Khon Kaen (population over 140,000; KhonKaen.com, 2013), in the Isan region, is located in the northeastern part of Thailand (population about 68 million). The Thai population is homogeneous, and as of the 2000 census, 99.8% of the population in Khon Kaen was of Thai nationality, with 99.4% of Buddhist religious affiliation.66 Thailand is described and ranked as a collectivistic society that builds on tradition and hierarchy.6,38,39,41,91,98 Parenting in collectivistic countries, including Khon Kaen,90 is traditionally described as being more authoritarian, intrusive, and restrictive than in individualistic societies.13,85
2.3.1. Translation of study measures
The translation process was an integral part of the study design, where the translation, review, adjudication, pretesting, and documentation translation model, or TRAPD, were applied to translate the main study instruments from English (source language) to Icelandic and Thai (target language).33,99 This translation model starts with a source-language instrument, and through its 5 interconnected steps, ends with a target-language instrument. In the translation model, an interdisciplinary group of individuals with widespread knowledge of the local language and culture gathered and used numerous procedures to maximize equivalence. For example, in the pretesting step, back-translation, qualitative interviews, and pilot testing techniques were used.35,93 The goal was to develop translations that “ask-the-same-questions.” This meant that, for the instrument to be culturally appropriate and relevant, in some cases it was necessary to adapt items to the specific context.34,99 If items could not be translated word-for-word, the reason was reported and coded using a coding system (language = 1, culture = 2, concept = 3, and measurement = 4). The coding system built on van de Vijver's108 adaptation categories: (1) language-driven adaptation, which involves the accommodation for differences in language structures (eg, gender-specific sentences) and principles (eg, directness); (2) culture-driven adaptation, which involves accommodation for differences in cultural norms, customs and practices of communication (eg, acceptability of emotional expression), and terminology characteristic (eg, temperature); (3) concept-driven adaptation, which involves accommodations for differences in familiarity and sameness of concepts across cultures (eg, changing the names of institutions); and (4) measurement-driven adaptation, which involves accommodations for differences in familiarity with stimulus (eg, pain faces) and the formatting of stimuli (eg, if language is read from left to right or vice versa).
2.3.2. Parent response to child's pain behaviors
The Inventory of Parent/Caregiver Responses to the Children's Pain Experience (IRPEDNA)42 includes 37 self-reported items that capture 3 interrelated scales related to parental responses to their child's pain behavior: solicitousness, discouragement, and promotion of well-behavior and coping. The solicitousness scale (15 items) measures parents' positive reinforcement (eg, “I will get home as early as I can”) and negative reinforcement (eg, “We take care of all his/her obligations and chores at home while he/she is in pain and discomfort”) of a child's pain behavior. The discouragement scale (10 items) captures parental responses that ignore/discourage the child's pain (eg, “I ignore him/her”) and criticize the child's pain behavior (eg, “I tell him/her not to complain so much”). The promotion of well-behavior and coping scale (12 items) captures parental responses that promote adaptive behaviors (eg, “I tell him/her to … listen to music or watch television”) and coping (eg, “I tell him/her that he/she can cope with the problem by saying things like “You're strong”). The IRPEDNA asks participants to rate statements on a 5-point scale as follows: 1 (“Never”), 2 (“Very Occasionally”), 3 (“Sometimes”), 4 (“Quite Often”), and 5 (“Always”). The scores indicate the degree to which a parent endorses a particular response. The scale with the highest overall mean indicates the preferred response for that parent. The English version is a translation of the original Catalan version. It has also been translated into Dutch110 and German.37 IRPEDNA has shown good construct (criterion) validity with Catalan samples and shown good Cronbach's alpha (0.80-0.89) in European samples (Catalan and Dutch).42,110 Furthermore, the IRPEDNA scale closely matches the German child-and-parent versions of the Pain-Related Parent Behavior Inventory at item level and scalar structure.37
The translation process resulted in no deletion of items in neither the Icelandic IRPEDNA version or in the Thai IRPEDNA version. In the Icelandic version, all items except #11, 20, and 25 needed a language-driven adaptation and 1 item (#19) needed a culture-driven adaptation. For the Thai version, all items except #1, 11, and 37 needed a culture-related adaptation. The Cronbach's α-coefficient values for the solicitousness scale (Canada 0.88; Iceland 0.91; and Thailand 0.85), promotion of well-behavior and coping scale (Canada 0.81; Iceland 0.83; and Thailand 0.88), and the discouraging scale (Canada 0.83; Iceland 0.75; and Thailand 0.73) were adequate in all 3 samples. Although the Cronbach's alpha was sufficient for the promotion of well-behavior and coping scales, it correlated highly with the solicitousness scale, especially for the Thai (r = 0.77) sample. Given that the solicitousness scale rendered to the study's main hypotheses, the promotion of well-behavior and coping scale was not used, but the solicitousness scale was.
2.3.3. Parenting styles
The Parenting Styles and Dimensions Questionnaire–Short Form (PSDQ-SF)81 consists of 32 self-report items listing 3 parenting style scales: authoritative, authoritarian, and permissive. The authoritative scale consists of 15 items measuring the dimensions of warmth, support, regulation, and autonomy granting (eg, “I emphasize the reasons for rules”). The authoritarian scale consists of 12 items measuring the dimensions of verbal hostility, physical coercion, and nonreasoning/punitive (eg, “I spank when my child is disobedient”). The permissive scale consists of 5 items measuring indulgence (eg, “I spoil our child”). The PSDQ-SF asks parents to rate the frequency of each type of behavior on a 5-point Likert-type scale (1 = “never” to 5 = “always”). A mean score is computed for each scale, with higher scores indicating higher levels of the construct. The PSDQ-SF questionnaire has shown to have concurrent validity,80 face validity,70 criterion validity (long version),86 and adequate psychometric properties.60 Many researchers have tested this instrument cross-culturally. Concept equivalence between the Chinese and the American versions of the authoritative scale and of the authoritarian scale is reported.114 Measurement invariance of the scales has been observed in Lithuania,49 Turkey,70 and Japan.52
The translation process resulted in no deletion of items in the Icelandic PSDQ-SF version or in the Thai PSDQ-SF version. In the Icelandic version, all items needed a language-driven adaptation. Also, in the Thai version, one item (#10) needed a culture-driven adaptation. The Cronbach's α-coefficient values for the authoritative-parenting scale (Canada 0.85; Iceland 0.89; and Thailand 0.90) and the authoritarian-parenting scale (Canada 0.81; Iceland 0.72; and Thailand 0.74) were adequate in all 3 samples. The Cronbach's α-coefficient values for the permissive scale were low in all 3 groups (Canada 0.68; Iceland 0.67; and Thailand 0.55). Thus, this study did not use the permissive scale, only the authoritative and authoritarian scales were used.
2.3.4. Cultural values
The Individualism/Collectivism Scale (INDCOL89) is a 32-item self-report measure listing cultural value dimensions describing how individuals perceive themselves and interact with others (Table 2). It consists of 2 main scales and 4 subscales. The 2 main scales are: (1) the collectivism scale (16 items), which represents individuals who emphasize interdependence between the individual and the group. Collective goals, norms, traditions, and authority figures are valued, and regulate individuals' behaviors and communications; and (2) the individualism scale (16 items), which represents individuals who emphasize independence of individuals from a group. Here, the value is on personal freedom and fulfillment of personal goals, where individual attitudes, rather than group norms, influence behavior and social communication. Thus, full emotional expression is expected and even necessary for personal well-being (Table 2).103,104,106
The collectivism scale consists of 2 subscales: the horizontal collectivism scale (8 items), which represents individuals who see themselves as interdependent and similar to others in a nonhierarchical manner (ie, expects equality), and the vertical collectivism scale (8 items), which characterizes individuals as interdependent, but gives the goals of the group a higher priority than personal individual goals (ie, inequality is acceptable). Similarly, the individualism scale consists of 2 subscales: the horizontal individualism scale (8 items) aligns with individuals who see themselves as independent, self-reliant, and unique, but not competitive (ie, expects equality), and the vertical individualism scale (8 items), which describes individuals who emphasize their independence, self-reliance, uniqueness, and competition. The INDCOL asks participants to rate statements on a 9-point scale (1 = “strongly disagree” to 9 = “strongly agree”). A mean score is computed for each scale. The INDCOL has good construct (divergent and convergent) validity.89,106
INDCOL is one of the most respected and commonly used tools to measure collectivism and individualism71,72 and has been translated into multiple languages. The coefficient alphas for vertical individualism, horizontal individualism, vertical collectivism, and horizontal collectivism, with North American samples ranging between α = 0.47 and α = 0.83; European samples ranging between α = 0.53 and α = 0.77; and Asian samples ranging between α = 0.46 and α = 0.81.29,30,48,79 Factor analyses89 and confirmatory factor analyses17,79 have demonstrated that the defined subscales are relatively constant.
The translation process resulted in no deletion of items in neither the Icelandic INDCOL version or in the Thai INDCOL version. In the Icelandic version, 20 items (#1, 5, 9, 10, 11, 12, 14, 15, 16, 17, 18, 20, 22, 23, 24, 25, 26, 28, 29, and 31) needed a language-related adaptation, and 2 items (#2, 7) needed a culture-driven adaptation. For Thai version, 3 items (#21, 29, and 31) needed a language-related adaptation. In this study, the collectivism scale and the horizontal/vertical individualism scales were deemed to have construct validity, but not the horizontal/vertical collectivism scales. Intersubscale Pearson correlations showed that the horizontal collectivism factor and vertical collectivism factor were highly correlated, especially for the Thai (r = 0.81) sample. Also, the internal-consistency reliability of the vertical collectivism scale was low in the Canadian (α = 0.63) and Icelandic (α = 0.61) samples. On the basis of these findings, the collectivism scale was used, rather than the vertical–horizontal collectivism subscales, in all further analyses. In this study, the Cronbach's α-coefficient values for the collectivism scale (Canada 0.76; Iceland 0.82; and Thailand 0.91), the horizontal individualism scale (Canada 0.78; Iceland 0.75; and Thailand 0.75), and the vertical individualism scale (Canada 0.82; Iceland 0.81; and Thailand 0.69) were adequate in all 3 samples.
2.3.5. Other variables
All data concerning the demographics of the child and the caregiver were gathered using caregiver self-report. The respondent to the survey had to be either a parent or a legal guardian of the child. The relationship of the respondent to the child was coded as mother, father, stepmother, stepfather, legal guardian, sister, aunt, uncle, grandmother, and grandfather. Respondents indicated their age in the categories of 20 to 24 years, 25 to 29 years, 30 to 34 years, 35 to 39 years, 40 to 44 years, 45 to 49 years, 50 to 54 years, 55 to 59 years, 60 to 64 years, 65 to 69 years, and 70+ years. Respondents' marital status was coded in the categories of married, common law, divorced/separated, remarried, widowed, and never married. In addition, socioeconomic information was collected, including education and occupation of the respondents and their spouses. Occupation was collected as an open-ended question. However, it was not possible to code those responses, particularly in the Thai sample, so that variable was dropped. Education level was assessed using an eight-level ordinal scale: less than seventh grade, junior high school graduate, some high school, high school graduate, trade school or community college, partial university (at least 1 year), bachelor's degree, and graduate school (including professional training such as doctor, dentist, or lawyer). An open-ended category called “Other” was included.
Apart from the child's age and sex, information about the child's pain frequency was collected by asking caregivers how often the child had experienced headache, stomachache, backache, joint pain, and other pain complaints in the past 3 months. Each of these was scored on a 5-point scale, with 1 being “Seldom or never,” 2 “About once a month,” 3 “About once a week,” 4 “More than once a week,” and 5 being “Most days.”94 Furthermore, participants were asked how many times in the past 3 months their child had stayed home from school due to pain using a 4-point scale: 1 = “Never,” 2 = “One to three times,” 3 = “Four to six times,” and 4 = “More than seven times.” Finally, data pertaining to chronic illness and to previous hospitalizations were collected as binary variables (yes/no).
2.4. Statistical analysis
Statistical analyses were conducted with IBM SPSS 20.0 statistical software and analysis. Structural equation modeling (SEM) was conducted with the lavaan package in R statistical software, version 126.96.36.199,82 Statistical significance was set at P < 0.05 unless otherwise noted. Frequency counts and percentages were used to describe the demographic backgrounds across the 3 country groups. For assessing the similarity and differences of demographic backgrounds across 3 country groups, the analysis of variance was used for numeric variables, and the χ2 test of independence for categorical variables. If main effects were significant, this was followed by pairwise comparisons.
Structural equation modeling was used to understand how pain-related parental responses are affected by culture and to ensure that the measurement tools used were cross-culturally invariant (comparable). As such, the SEM analysis consisted of 2 models: (1) the measurement model, which tested and ensured that the Canadian, Icelandic, and Thai versions of the IRPEDNA, INDCOL, and PSDQ-SF scales yielded cross-cultural measurement invariance; and (2) the structural model, which examined how cultural parenting models affect discouraging and solicitousness responses across Canadian, Icelandic, and Thai samples. Because of multivariate normality violation, models were tested with robust maximum likelihood estimation.53 Goodness-of-fit indices used included: the root-mean-square error of approximation (RMSEA; ≤0.06 indicated adequate fit), the standardized root-mean-square residual (SRMR; <0.08 indicated adequate fit), the comparative fit index (CFI; >0.90 indicated adequate fit), and the Tucker–Lewis Index (TLI; >0.90 indicated adequate fit).40,50 Models were deemed acceptable if at least 3 of these 4 indices showed adequate fit. Also, item parceling was used to increase model parsimony and stability of all the latent constructs (factor). When using item parceling, 2 or more items are combined to create parcels, which are then used as observed indicators of the latent constructs in the SEM.32,57 No item was deleted and all relevant items were used when creating the parcels. To create the parcels, recommendations from Little57,58 and Matsunaga63 were applied. Finally, bootstrapping (with 5000 resamples) was also used to address non-normality distribution in data. The bootstrapping method is a nonparametric approach to effect size estimation and hypothesis testing.74
2.4.1. Performing cross-cultural measurement invariances (testing the measurement model)
A single-group confirmatory factor analysis (CFA) and multigroup CFA was performed to examine cross-cultural measurement invariance. Measurement-invariance testing ensured that the constructs measured had the same meaning across the 3 country groups. This involves the comparisons of series of measurement models with gradual restrictive constraints between the 3 groups. Cheung and Rensvold's15 rule of CFI-difference (ΔCFI ≥0.01) was used to examine a significant increase between a pair of comparisons between 2 nested model specifications. Finding improvement in CFI of 0.01 or more when freeing a parameter across multiple groups indicates that the parameter value is different across the groups tested.
Two levels of invariance testing were performed.109 First, the configural invariance was tested to see whether the participants from the 3 country groups conceptualized the subscale constructs similarly. Second, metric invariance was performed to check whether each item contributed to the latent factor in a roughly equivalent way across all 3 countries. This study did not compare mean of scale scores across Canadian, Icelandic, and Thai samples. Therefore, it did not test for invariance of mean of scale score (scalar invariance).
2.4.2. Hypotheses testing: moderation and mediation analysis (testing the structural model)
188.8.131.52. Moderation analysis (hypotheses 1 and 2)
The main analysis was multigroup SEM (MG-SEM) that tested whether the direct and indirect pathways in the hypothesized structural model were moderated by ecosocial context. The aim of the moderation was to specify whether the ecosocial context in the 3 countries increases or decreases the strength of the effect of cultural values on pain-related caregiver behaviors. Specifically, we tested whether an individualistic ecosocial context (ie, Canada and Iceland) creates a horizontal/vertical individualism-authoritative-solicitous pathway, and whether a collectivistic ecosocial context (ie, Thailand) would create a collectivism-authoritarian-discouraging pathway. The results of this analysis determined whether single-group or MG-SEM would be used to test whether parenting styles mediated the effects of cultural values on pain-related caregivers' behaviors.
184.108.40.206. Mediation analysis (hypotheses 3-10)
The aim of the mediation was to specify how the predictor variables (vertical individualism, horizontal individualism, and collectivist values) affected outcome variables (solicitous and discouraging behavior) through mediators (authoritative- and authoritarian-parenting style). Specifically, we tested whether horizontal/vertical individualism would have a positive effect on solicitousness, through authoritative-parenting style. We also tested whether collectivism would have a positive effect on discouraging, through authoritarian-parenting style.
When describing mediation pathways in the structural model, the terms a, b, c, and c′ paths were used, to align with the nomenclature used by Baron and Kenny.2 The a paths represent the effect of the predictor variables (cultural values) on the mediator variables (parenting styles) (ie, hypotheses 3-4). The b paths represent the effect of the mediator (parenting styles) variables on the outcome variables (pain-related parental behaviors) (ie, hypotheses 5-6). The c′ paths (often called direct effects) represent any remaining link between predictor and outcome, after controlling for mediators. The c paths (often called total effects) represent the total effect of predictor on outcome, before adding the mediating variables into the model (ie, hypotheses 7-8). The ab paths represent the indirect effect of the predictor on the outcome, through mediator variables (ie, hypotheses 9-10). The ab paths are calculated by multiplying the a path and b path together and are mathematically equivalent to the difference of the c and c′ paths (c−c′74). The mediation analysis was based on Preacher and Hayes's74 framework, where the total effect (path c) does not need to be significant, as the focus is on the size of the indirect effect, and mediation is only said to occur as long as the indirect effect is statistically significant. Statistical significance of the indirect effect was calculated using bias-corrected bootstrapping with 5000 resamples.74 The ratio of indirect to total effect (percent mediation, PM) was calculated to assess the effect size of the indirect effect, using the formula PM = ab/ab +c′.61,74,75
2.4.3. Testing covariate effect
The final part of the MG-SEM involved the controlling for potential confounding variables in the final structural model. The covariates were of secondary concern, and thus included at the later stages of the model-building process. Covariates were selected based on their theoretically meaningful potential influence on the modeled relationships. Given that SEM is a theory-driven method (and to reduce model complexity given the sample size), only one covariate was added at a time. The covariates were added to see whether the effects would change, once any potential covariate's influence was controlled.56
3.1. Demographic background of samples
The characteristics of the 3 country groups are provided in Table 3. For the analysis, female participants who identified as the mother, stepmother, or female legal guardian of the target child were categorized as mothers. Similarly, male participants who identified as the father, stepfather, or male legal guardian were categorized as fathers. A chi-square test showed that more caregivers, or 425 (80%) of participants, were mothers than fathers, χ2(2) = 26.982, P < 0.0001. Compared with the Canadian and Icelandic participants, more Thais used the “Other” category (ie, not “mother” or “father”) to describe their relationship with the child (Table 3). In the Thai “Other” category (n = 31), a majority (65%) were grandmothers. Overall, most caregivers were married (82.4%), and the pattern for marital status did not differ by country (Table 3). Caregivers' age was collected as a numeric value, ranged from 22 to 67 years, and was significantly different between groups (Table 3). A Tukey post hoc test revealed that Thai caregivers were significantly older than Canadian (P < 0.001) and Icelandic (P < 0.001) caregivers. A χ2 test showed significant group differences on maximum family education levels (Table 3), with Thais reporting lower educational levels than Canadians (P < 0.0001) and Icelanders (P < 0.0001). Most Canadian and Icelandic families had postsecondary education, whereas Thai families tended to have trade school or community college education levels. Although the undergraduate degree was the most frequently reported education level in all countries (ie, Canada, 35%; Iceland, 39%; and Thailand, 24%), the Thai sample had a much higher proportion with primary education only.
Children's ages ranged from 6 to 12 years and were relatively similar across groups. Significant differences between groups (Table 3) on children's age emerged with Canadian children being significantly younger than those in Iceland (P < 0.01) and Thailand (P < 0.01). Although the age difference was relatively small and all children were within the expected age range, age was used as a potential covariate in the main analysis. The proportion of boys and girls did not differ significantly across the 3 samples. No significant differences arose between countries on chronic illness (Table 3). However, significant differences emerged between countries on experience of hospitalization (Table 3), with Thais reporting a rate over 57%, compared with Canadians with around 23%, and Icelanders with around 27%. A variable was created (this was a binary variable, looking only at the chronically ill children, and was coded: “0” = not hospitalized but had a chronic illness and “1” = hospitalized and had a chronic illness) that combined chronic illness with hospitalization in each country, the analysis of which showed no significant differences (χ2(2) = 1.210, P < 0.546). When all caregivers were asked whether the child had missed school because of pain in the past 3 months, 30% answered yes. There was no significant difference found between the groups on days missed at school because of pain (Table 3). Figure 2 shows the prevalence of recurrent pain in the past 3 months, by the type of pain. Overall, stomach pain was the most common type of recurrent pain in Canada (16.9%) and Iceland (15.4%), whereas in Thailand, “other pain” (5.9%) was most frequently reported and included: “toothache,” “sore throat,” “eye pain,” “allergy,” “fever,” and “menstruation.” Furthermore, χ2 analysis indicated group differences for stomachache (χ2(2) = 17.670, P < 0.0001) and headache (χ2(2) = 13.458, P < 0.001), and in both instances, Thai caregivers reported significantly lower frequencies compared with those in the other countries (stomachache: Thailand vs Canada [P < 0.0001]; Thailand vs Iceland [P < 0.0001]; and headache: Thailand vs Canada [P < 0.0001]; Thailand vs Iceland [P < 0.0001]) (Fig. 2).
Please note that the mean score, SDs, and correlation coefficients for all measures (horizontal individualism, vertical individualism, collectivism, authoritative-parenting style, authoritarian-parenting style, solicitousness, and discouraging) for each sample can be found in a supplementary Table (available online at http://links.lww.com/PAIN/A591).
3.2. Cross-cultural measurement invariance of the measurement model
First, single-group CFA was used to define the 7-factor measurement model in each of the 3 counties separately. Results from goodness-of-fit absolute indices (ie, RMSEA ≤0.06 and SRMR ≤0.08) showed that the proposed factor structure was adequate in each country (Table 4). Next, MG-SEM was used to test the cross-cultural invariance of the 7-factor measurement model. The goodness-of-fit indices for the configural-invariance model and the metric invariance yielded adequate fit to the data as their CFI and TLI values were >0.9. Finally, when the change in CFI (comparative fit indices) between the configural and metric models was tested, showing the [INCREMENT]CFI = > 0.950 (CFIconfigural) − 0.943 (CFImetric) = 0.007, which, based on Cheung and Rensvold's15 rule (ie, a change < 0.01 is nonsignificant), indicated that the 2 models were not significantly different. Based on these results, the 7-factor measurement model shows metric invariance across the 3 country samples (Table 4). This multigroup, metric invariance model was incorporated for the latent variables at all subsequent steps.
3.3. Testing the hypothesized structural model
Figure 3 shows the results for the final 7-factor structural model. Unstandardized, rather than standardized, regression coefficients were used, as they are preferred when comparing results for the same predictors across different samples.50 Here, the relationship between cultural values (vertical individualism, horizontal individualism, and collectivism) and pain-related behaviors (solicitousness and discouraging) is mediated by parenting styles (authoritative and authoritarian). That is, cultural values are associated with caregivers' parenting style, which in turn are associated with caregiver pain-related behavior. In this way, cultural values have an indirect effect on caregivers' behaviors through parenting styles.
3.3.1. Cross-cultural comparisons of hypothesized structural model (testing hypotheses 1 and 2)
Multigroup SEM was used to test whether the magnitude of the paths, depicted in Figure 3, differed across countries. The goodness-of-fit indices for the freely estimated structural model, and the structural model with paths (but not covariances) constrained to equality, appear in Table 4. The goodness-of-fit indices for a structural model with all paths and covariances allowed to freely vary across countries showed adequate fit: χ2(532) = 789.658, P = 0.0001, CFI = 0.94, TLI = 0.93, RMSEA = 0.05, and SRMR = 0.06. When all paths were constrained to equality across countries, the model continued to demonstrate adequate fit: χ2(564) = 835.740, P = 0.0001, CFI = 0.94, TLI = 0.93, RMSEA = 0.05, and SRMR = 0.08. In comparing the fit statistics for these 2 models, the outcomes were very similar. The change in CFI was small ([INCREMENT]CFI = 0.003), suggesting that the more parsimonious model (ie, paths constrained to equality) should be preferred.15 These results indicated that country did not change the relationships between the predictor, mediator, and outcome variables in the model. Therefore, moderation did not occur. This meant that all future analysis would use a multigroup model, with factor loadings and regression paths constrained to equality across countries, not single-group SEM.
3.3.2. Relation between cultural values and parenting styles (testing hypotheses 3 and 4)
Table 5 and Figure 3 show the relationships between cultural values and parenting styles.
Authoritative-parenting style was negatively associated with vertical individualism and positively associated with both horizontal individualism and collectivism, while controlling for all other predictor variables. Of these 3 relationships, collectivism emerged as the strongest predictor, and collectively, these variables predicted between 17% and 23% of the variance in authoritative-parenting. However, of these effects, only the positive relationship between horizontal individualism and authoritative-parenting style was hypothesized a priori.
Authoritarian-parenting style was positively associated with vertical individualism, and unrelated to horizontal individualism and collectivism. Collectively, these variables predicted between 9% and 13% of the variance in authoritarian-parenting. That is, participants high on vertical individualism were likely to report using an authoritarian-parenting style.
3.3.3. Relation between parenting styles and outcomes (testing hypotheses 5 and 6)
Both authoritative-parenting style and authoritarian-parenting style positively predicted solicitousness. That is, as both of these parenting styles increased, solicitousness also increased. That means parents who used an authoritarian-parenting style were also likely to use solicitousness. However, the largest relationship was observed for authoritative-parenting. Collectively, all variables explained between 11% and 33% of the variance in solicitousness. By contrast, only authoritarian-parenting style predicted discouraging; authoritative-parenting style had a near-zero, nonsignificant relationship with discouraging. Collectively, all variables predicted between 26% and 31% of the variance in discouraging (Fig. 3 and Table 5).
3.3.4. Direct and total effects on solicitousness and discouraging (testing hypotheses 7 and 8)
There were no statistically significant total or direct effects of cultural values (ie, vertical individualism, horizontal individualism, or collectivism) on solicitousness. So, these variables were generally unrelated to the solicitousness. However, given the significant a paths and b paths observed in the data, mediation might still occur. This is known as “inconsistent mediation” and usually occurs because the mediator is acting as a suppressor variable.61 By contrast, there was a positive total effect for vertical individualism on discouragement. Moreover, there was a significant positive direct effect for vertical individualism on discouragement and a significant negative direct effect for collectivism on discouragement. In general then, the vertical individualism → authoritarian → discouragement mediation path seems to be the strongest candidate for traditional mediation (Fig. 3 and Table 5).
3.3.5. Indirect effects on solicitousness (testing hypothesis 9)
Table 5 shows the 4 specific significant mediations for the effect of cultural values on solicitousness through parenting styles that emerged. Results showed that participants' score on vertical individualism had a significant negative indirect effect on their scores for solicitousness through their scores on authoritative-parenting style (95% confidence interval [CI] [0.062 to −0.015]). That is, increased vertical individualism was associated with a decrease in authoritative-parenting, which in turn was associated with a decrease in solicitousness. The relationship between vertical individualism and authoritative-parenting style accounts for 51% of the negative relationship between vertical individualism and solicitousness. That is, about 51% of vertical individualism's negative effects on solicitousness were due to its negative relation with the authoritative-parenting style (which, in turn, positively related to solicitousness).
Results showed that participants' scores on vertical individualism had a significant positive indirect effect on solicitousness through the authoritarian-parenting style (95% CI [0.003-0.036]). That is, increases in vertical individualism were associated with an increase in authoritarian-parenting, which in turn was associated with an increase in solicitousness. The relationship between vertical individualism and authoritarian-parenting style accounts for 32% of the relationship between vertical individualism and solicitousness. Moreover, 32% of vertical individualism's positive effect on solicitousness was due to its positive effects on authoritarian-parenting style.
Results showed that participants' scores on horizontal individualism had a significant indirect effect on solicitousness through the authoritative-parenting style (95% CI [0.008-0.084]). That is, horizontal individualism was associated with an increase in authoritative-parenting, which in turn was associated with an increase in solicitousness. The relationship between horizontal individualism and authoritative-parenting style accounts for 78% of the relationship between horizontal individualism and solicitousness. Moreover, 78% of horizontal individualism's positive effect on solicitousness was due to its positive effect on authoritative-parenting style.
Collectivism also indirectly predicted solicitousness through authoritative-parenting (95% CI [0.039-0.138]), with the authoritative-parenting style accounting for 61% of the total effect of collectivism on solicitousness. The relationship between collectivism and authoritative-parenting style accounts for 61% of the relationship between collectivism and solicitousness.
3.3.6. Indirect effects on discouraging (testing hypothesis 10)
Table 5 shows the 6 indirect paths to discouraging. Of these, one specific significant mediation for the effect of cultural values on discouraging through parenting styles emerged. Results showed that vertical individualism had a significant indirect effect on discouraging through the authoritarian-parenting style (95% CI [0.015-0.055]). That is, vertical individualism was associated with an increase in authoritarian-parenting, which resulted in an increase in discouraging behavior. The relationship between vertical individualism and authoritarian-parenting style accounts for 41% of the relationship between vertical individualism and discouraging. In other words, vertical individualism had an indirect effect on discouraging through the authoritarian-parenting style, but the authoritarian-parenting style accounted for 41% of the total effect of vertical individualism on discouraging.
3.3.7. Controlling for child sex, child age, and family education
Covariates were entered one at a time, at later stages of the MG-SEM, to ensure that they did not impact the primary results.56 The goodness-of-fit indices for the structural model with paths and covariances constrained to equality for child sex, child age, and family max education separately, appear in Table 4. Results show that for all 3 models, 3 of the 4 goodness-of-fit indices showed adequate fit; only the SRMSR index suggested poorer fit. Importantly, the analysis also revealed the same pattern of results described earlier, even when adding these covariates, indicating that the results hold when controlling for child age, child sex, and family education. However, when each individual covariate was explored and each individual variable in the model, some important results emerged.
220.127.116.11. Child age
Child age negatively and significantly linked with authoritarian-parenting style (B = −0.029, SE = 0.009, P = 0.001), indicating that as children got older, parents tended to be less authoritarian. Furthermore, for the Icelandic sample only, child age strongly-to-moderately and negatively correlated with vertical individualism (B = −0.504, SE = 0.224, P = 0.024), indicating that as children got older, parents reported less vertical individualism; however, given the number of tests conducted, the P value (which is close to the 0.05 cutoff), the large standard error, and the lack of a priori theory predicting these results, it is probable that this finding is spurious. In Canada and Thailand, the correlation between child age and vertical individualism was weak and nonsignificant (B = −0.039, SE = 0.221, P = 0.859; B = 0.03, SE = 0.265, P = 0.911, respectively).
18.104.22.168. Child sex
Child sex was not found to be an important covariate, as it did not significantly predict any exogenous or endogenous variables.
22.214.171.124. Family education
After controlling for other variables in the model, a negative and significant relationship emerged with solicitous pain-related parent responses (B = −0.037, SE = 0.016, P = 0.018), indicating that parents with higher education tended to be less solicitous. Again, this exploratory relationship should be treated cautiously, given the number of tests conducted. Furthermore, for the Canadian sample only, family education moderately and positively correlated with vertical individualism (B = 0.410, SE = 0.146, P = 0.005), but this relationship was nonsignificant in the Icelandic sample (B = −0.043, SE = 0.175, P = 0.805) and the Thai sample (B = −0.369, SE = 0.250, P = 0.141). For the Icelandic sample, however, family education strongly and positively correlated with horizontal individualism (B = 0.536, SE = 0.183, P = 0.003), but was nonsignificant in the Canadian and the Thai samples (B = 0.107, SE = 0.111, P = 0.332; B = 0.108, SE = 0.263, P = 0.681, respectively).
To the best of our knowledge, this was the first study to examine the association between ecosocial context, cultural values, parenting styles, and pain-related parental responses by using samples from 3 countries. Two key findings, summarized in Figure 4, can be extracted from this study. First, the ecosocial context does not moderate the association between cultural values, parenting styles, and pain-related parental behaviors. This is contrary to ecological theories. Maybe the study sites were too similar, as all were in small cities. Yet, another explanation consistent with universal theories7,18,76,83,101 is that this developmental process may be universal in the sense that these behaviors are commonly found in individuals, independent of their ecosocial context, and that these processes rest on general principles, rather than the specific individual attributes studied.
Second, consistent with cultural models of parenting theories,25,45,47 a complex relationship exists between cultural values, parenting styles, and pain-related parental behaviors. This study showed that parenting style consistently mediated the relationship between the cultural values and pain-related behaviors adopted by parents. Overall, these findings suggest that, in a pain context, parents oriented towards collectivism (eg, emphasize sharing, helping, and sacrificing) and horizontal individualism (eg, emphasize equality, fairness, and similarity) are likely to adopt authoritative-parenting style and in turn, use solicitousness pain-related response, but not discouraging. On the other hand, parents oriented towards vertical individualism (eg, emphasize hierarchy and self-reliance) are likely to adopt authoritarian-parenting style and in turn, likely use discouraging pain-related parental behavior, but also solicitousness. Furthermore, when controlling for child sex, age, and family education, overall results did not change. To fully understand the results, one must first understand the pieces contributing to the overall findings.
4.1. Parents' cultural values and parenting styles
Within the structural model, it was identified that parents who oriented towards horizontal individualism adopted authoritative-parenting styles, which is consistent with findings of others,97 showing that when a parents place a value on egalitarian communication and individual needs, it is reflected in their support of child's individuality, and needs for emotional self-expression.3,27,71 The findings of direct relationship between collectivism and authoritative-parenting contradicts some,85,115 but align with others,73,88,97 pointing to what others have suggested which is that collectivistic cultures do encourage “other-focused” emotions (eg, empathy), while discouraging “self-focused” emotions (eg, anger).54,62 In fact, cross-cultural neuroscientists have shown that a collectivistic mindset, but not a individualistic one, increases activation of brain areas responsible for “other-oriented” responses to pain, such as empathy.16,112 The direct relationship found between vertical individualism and authoritarian-parenting style aligns with some findings,23,48 but contradict others.84,97 At first glance, vertical individualism and authoritarian-parenting style may seem incompatible. Both, however, facilitate hierarchical communication, power imbalance, and submission to authority.12,23,48,92
These findings indicate that our conventional way of viewing the relationship between cultural values and parenting styles may not hold. This may be especially true when cultural values are operationalized as multidimensional, as it adds 2 additional dimensions, horizontal and vertical, that are generally overlooked.
4.2. Parents' parenting styles and pain-related parental behaviors
The structural model revealed a strong significant relationship between parenting styles and pain-related parental behaviors, suggesting that parenting styles are a powerful mode of pain socialization of parents. The strongest relation emerged between authoritative-parenting style and solicitousness, and between authoritarian-parenting and discouraging. These results are consistent with the findings of others.10,37 The results also showed a direct, but less strong, link between authoritarian-parenting style and solicitousness. Parenting styles are conveyed through parents' body language and tone of voice.21 As such, these results may link to what others have reported, which is that solicitousness can be delivered with hostility.67 These results also suggest that solicitousness may be an inherent pain-related parental response, which is addressed later.
4.3. Parents' cultural values, parenting styles, and pain-related parental behaviors
We posited on the basis of theory25,45 that cultural values affect pain-related parental behaviors, through mechanisms of parenting styles. Parenting styles were shown to mediate the relationship between parents' cultural value orientation and pain-related parental behaviors. These results align with the sociocommunication model of children's pain,78 which suggest that cultural elements assert their influences indirectly through caregivers, providing caregivers with schemas for culturally appropriate behavior.
Most of the indirect relationships found were unexpected (Fig. 4). Both authoritative- and authoritarian-parenting styles were positive mediators for solicitousness, and all 3 cultural values, including collectivism, indirectly influenced solicitousness. Although most of these cultural models of parenting were unexpected, the literature also suggests that around the same developmental task, there are “infinite combinations along these dimensions” (p. 33).45 What these models seem to provide is variance in the way solicitousness can be expressed. More specifically, it is possible that: (1) when horizontal individualism and an authoritative-parenting style facilitates solicitousness, it may occur in a child-centered atmosphere where the child is encouraged to express pain openly and freely, facilitating self-expression, individuality, and self-efficacy, which are all important characteristics of individualism; (2) when vertical individualism and an authoritarian-parenting style facilitate solicitousness, the atmosphere might be self-oriented and less child friendly, perhaps serving as a way for the parent to vent and express personal distress concerning the child's pain; and (3) when collectivism and an authoritative-parenting style facilitate solicitousness, the atmosphere is other-oriented. Parents show empathy and an understanding toward their child's needs and feel obliged to help their child. These are all speculations built on theory, but consistent with current results.
Another unexpected relationship was that authoritarian-parenting style mediated discouraging through vertical individualism, but not collectivism. According to Goubert et al.,24 parental pain-related discouraging behavior may be aimed at reducing the parent's own distress, rather than the child's. This possibility aligns with the suggestion that self-focusing and self-serving bias are characteristic for individuals who are high on vertical individualism.16,106 This notion aligns also with authoritarian-parenting, which is generally described as demanding unrealistic maturity and responsibility from the child.8 Another explanation for the results reported may relate to the self-reliance component of vertical individualism,106 as individualistic parents may strongly value self-reliance and emotional independence.43,102,113 Thus, discouraging behavior may also be the parent's approach to teach the child self-reliance when in pain.
4.4. Study limitations
The findings must be interpreted in the context of several limitations. A cross-sectional survey design and a convenience sample were used, providing parents' self-reports of interactions with children. This limits the generalizability of the findings to the general population as well as making conclusions of directions of relationships. Sampling was also limited to one delineated geographical region in each of the 3 countries. Therefore, conclusions are limited to the sample, variables, and timeframe presented by the design. Future studies should weigh the benefits and challenges of incorporating random samples into the design. In this study, the horizontal–vertical collectivistic subscales were problematic. The internal consistency of the vertical collectivism subscale was low (Canadian and Icelandic samples only), the correlation between the horizontal and vertical collectivism subscales was high (Thai sample only), and factor analysis showed that the vertical collectivistic items loaded highly on the horizontal collectivism subscale (Thai sample only). These issues indicated a conceptual overlap among the vertical/horizontal collectivism subscales. Similar to others,71 this issue was addressed by combining the 2 collectivism subscales. Future studies should consider using other measurements for individualism and collectivism.
This study suggests that parental behaviors around pain are informed by cultural models of parenting. The developmental literature suggests that some universal parenting behaviors exist. For example, adults' baby talk seems to be triggered by the presence of a baby. This behavior appears in virtually all investigated cultures, assumed to exist without explicit learning.7,46 Likewise, pediatric pain researchers described reassurance (one form of solicitousness) as being an “instinctive,”36 “naturalistic,”55 and “ingrained”65 way parents help their children in acute pain situations. Hence, solicitousness seems to be a universal attempt by parents to limit suffering and promote well-being in their children, which may convey an evolutionary advantage. Although solicitousness may be an “ingrained” behavior, and independent of ecosocial contexts, this behavior seems to be expressed differently depending on the cultural values and parenting styles promoting it. Also, the vertical individualism–authoritarian cultural model process suggests that solicitousness and discouraging behaviors do not necessarily preclude each other, as previously assumed. It is therefore possible that, in a clinical setting, caregivers respond to their child's pain with solicitousness and discouraging in conjunction.
The findings from this study have implications for the theory development about culture and pediatric pain. They support the sociocommunication model of children's pain by showing that cultural context does affect parents' behaviors. Furthermore, they add to this literature by specifying what cultural factors, and how these factors, influence pain-related parental behaviors. This study does not provide specific clinical recommendations. Culture has a complex relationship to pain and simplistic notions of the effect of culture on pediatric pain are not supported by the evidence.
Conflict of interest statement
The authors have no conflict of interest to declare.
This research was funded by the IWK Research Grant A (#1002689), the Canadian Research Chairs, the Icelandic Nurse's Association, the University of Iceland Research Fund, and the International Development Research Centre (Canada) through a Teasdale-Corti Team Grant. This research was conducted as part of the doctoral dissertation of O. Kristjansdottir and completed under the supervision of Dr McGrath. During her doctoral work, O. Kristjansdottir received support from the Izaak Walton Killam Memoria Predoctoral Scholarship, the IWK Health Centre Graduate Studentship Scholarship, the Nova Scotia Health Research Fund Scholarship, the PICH CIHR Strategic Training Grant Stipend, and the Icelandic Research Fund for Graduate Students.
The authors thank the caregivers who participated in the study. For help in the translation process, the authors thank Sóley Bender, Sigrún Júlíusdóttir, Helga Lára Helgadóttir, Rakel Jónsdóttir, Kristín Björnsdóttir, Sigurður Grétarsson, Gylfi Sigurðsson, Laufey Böðvarsdóttir, Guðrún Gestsdóttir, Anna Ólafia Sigurðardóttir, Ricardo Faulk, Aðalheiður Jóhannesdóttir, Þóra Laufey Pétursdóttir, Magnús Halldórsson, Sigrún Kristjánsdóttir, Hallgerður Lind Kristjánsdóttir, Magnús Þórarinnsson, Kristján Pálsson, Sóley Halla Þórhallsdóttir, Halla Hrafnkelsdóttir, Svanhildur Steinarsdóttir, Paula Forgeron, Parinya Santima, Wimonrat Sriraj, Somboon Theinthong, and Darunee Jongudomkarn. For help in the pilot testing in Canada, the authors thank Kelly Kitimira, Sarah Stevens, Nicole Gray, and Patricia Pottie. For help with recruitment in Iceland, the authors thank Grandaskóli, Melaskóli, Smáraskóli, Öldutúnsskóli, and Austurbæjarskóli, as well as the generous support from Sambíóin. For recruitment in Thailand, the authors thank Parinya Santima.
Supplemental digital content
Supplemental digital content associated with this article can be found online at http://links.lww.com/PAIN/A591.
. Adalbjarnardottir S, Hafsteinsson LG. Adolescents' perceived parenting styles
and their substance use: concurrent and longitudinal analyses. J Res Adolescence 2001;11:401–23.
. Baron R, Kenny D. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986;51:1173–82.
. Berns R. Child, family, school, community: socialization and support. Belmont: Wadsworth/Cengage Learning, 2013.
. Birnie K, Boerner K, Chambers C. Families and pain. In: McGrath PJ, Stevens BJ, Walker SM, Zempsky WT, editors. Oxford textbook of paediatric pain. New York: Oxford University Press, 2014. p. 111–18.
. Blondal KS, Adalbjarnardottir S. Parenting practices and school dropout: a longitudinal study. Adolescence 2009;44:729–49.
. Boonmathya RT. A narrative of contested views of development in Thai society: voices of villagers in rural northeastern Thailand. Southeast Asian Stud 2003;41:269–98.
. Bornstein MH. Cultural approaches to parenting. Parenting 2012;12:212–21.
. Bornstein MH, Zlotnik D. Parenting styles
and their effects. In: Marshall MH, Janette BB, editors. Encyclopedia of infant and early childhood development. San Diego: Academic Press, 2008. p. 496–509.
. Bush JP, Cockrell CS. Maternal factors predicting parenting behaviors in the pediatric clinic. J Pediatric Psychol 1987;12:505–518.
. Bush JP, Melamed BG, Sheras PL, Greenbaum PE. Mother-child patterns of coping with anticipatory medical stress. Health Psychol 1986;5:137–57.
. Chan SM. Links between Chinese mothers' parental beliefs and responses to children's expression of negative emotions. Early Child Develop Care 2012;182:723–739.
. Chan SM, Bowes J, Wyver S. Parenting style as a context for emotion socialization. Early Edu Dev 2009;20:631–56.
. Chao R, Tseng V. Parenting of Asians. In: Bornstein MH, editor. Handbook of parenting: social conditions and applied parenting. Mahway: Lawrence Erlbaum Associates 2002. p. 59–93.
. Cheah CSL, Chirkov V. Parents' personal and cultural beliefs regarding young children: a cross-cultural study of aboriginal and Euro-Canadian mothers. J Cross Cult Psychol 2008;39:402–23.
. Cheung GW, Rensvold RB. Evaluating goodness-of-fit indexes for testing measurement invariance. Struct Equ Modeling 2002;9:233–55.
. Chiao JY. Current emotion research in cultural neuroscience. Emot Rev 2015;7:280–93.
. Chiou JS. Horizontal and vertical individualism and collectivism among college students in the United States, Taiwan, and Argentina. J Soc Psychol 2001;141:667–78.
. Corter C, Fleming A. Psychobiology of maternal behavior in human beings. In: Bornstein MH, editor. Handbook of parenting: biology and ecology of parenting. London: Lawrence Erlbaum Associates, 2002. p. 141–82.
. Craig KD, Lilley CM, Gilbert CA. Social barriers to optimal pain management in infants and children. Clin J Pain 1996;12:232–42.
. Craig KD, Riddell RP. Social influences, culture
, and ethnicity. In: McGrath PJ, Finley GA, editors. Pediatric pain: biological and social context. Seattle: IASP Press, 2003. p. 159–82.
. Darling N, Steinberg L. Parenting style as context: an integrative model. Psychol Bull 1993;113:487–96.
. Devins GM, Gupta A, Cameron J, Woodend K, Mah K, Gladman D. Cultural syndromes and age moderate the emotional impact of illness intrusiveness in rheumatoid arthritis. Rehabil Psychol 2009;54:33–44.
. Georgiou SN, Fousiani K, Michaelides M, Stavrinides P. Cultural value orientation and authoritarian parenting as parameters of bullying and victimization at school. Int J Psychol 2013;48:69–78.
. Goubert L, Craig KD, Vervoort T, Morley S, Sullivan MJ, Williams ACdeC, Cano A, Crombez G. Facing others in pain: the effects of empathy. PAIN 2005;118:285–8.
. Greenfield PM. Linking social change and developmental change: shifting pathways of human development. Dev Psychol 2009;45:401–18.
. Greenfield PM, Keller H, Maynard A, Suzuki L, Charles S. Lifespan development and culture
. In: Spielberg CD, editor. Encyclopedia of applied psychology. New York: Elsevier, Inc, 2004. p. 567–574.
. Greenfield PM, Suzuki LK. Culture
and human development: implications for parenting, education, pediatrics, and mental health. In: Sigel E, Renninger KA, editors. Handbook of child psychology: child psychology in practice. Hoboken: John Wiley & Sons 1998. p. 1059–109.
. Greenfiled PM, Suzuki K. Culture
and parenthood. In: Westman JC, editor. Parenthood in America: undervalued, underpaid, under siege. Madison: University of Wisconsin Press, 2001. p. 20–33.
. Gushue GV, Constantine MG. Examining individualism, collectivism, and self-differentiation in African American college women. J Ment Health Couns 2003;25:1–15.
. Guss CD. Fire and ice: testing a model on culture
and complex problem solving. J Cross Cult Psychol 2011;42:1279–98.
. Häder S, Gabler S. Sampling and estimation. In: Harkness JA, van De Vijver FJR, Mohler PPH, editors. Cross cultural survey methods. Hoboken: John Wiley & Sons, 2003. p. 117–34.
. Hall RJ, Snell AF, Foust MS. Item parceling strategies in SEM: investigating the subtle effects of unmodeled secondary constructs. Organ Res Methods 1999;2:233–56.
. Harkness JA. Questionnaire translation. In: Harkness JA, van de Vijver FJR, Mohler PP, editors. Cross-cultural survey methods. Hoboken: John Wiley & Sons, 2003. p. 35–56.
. Harkness JA. Improving the comparability of translations. In: Jowell R, Roberts C, Fitzgerald R, Eva G, editors. Measuring attitudes cross-nationally. London: Sage, 2007. p. 79–93.
. Harkness JA, Villar A, Edwards B. Translation, adaptation, and design. In: Harkness JA, Braun M, Edwards B, Johnson TP, Lyberg L, Mohler PP, Pennell BE, Smith TW, editors. Survey methods in multinational, multiregional, and multicultural context. Hoboken: John Wiley & Sons, 2010. p. 115–40.
. Harrison D, Sampson M, Reszel J, Abdulla K, Barrowman N, Cumber J, Fuller A, Li C, Nicholls S, Pound CM. Too many crying babies: a systematic review of pain management practices during immunizations on YouTube. BMC Pediatr 2014;14:2–8.
. Hermann C, Zohsel K, Hohmeister J, Flor H. Dimensions of pain-related parent behavior: development and psychometric evaluation of a new measure for children and their parents. PAIN 2008;137:689–99.
. Hofstede G. Culture
's consequences: international differences in work-related values. Beverly Hills: Sage, 1980.
. Hofstede G. Culture
's consequences, comparing values, behaviors, institutions, and organizations across nations. Thousand Oaks: Sage, 2001.
. Hooper D, Coughlan J, Mullen M. Structural equation modelling: guidelines for determining model fit. Electron J Business Res Methods 2008;6:53–60.
. Hughes C. The influence of self-concept, parenting style and individualism–collectivism on career maturity in Australia and Thailand. Int J Educ Vocational Guidance 2011;11:197–210.
. Huguet A, Miro J, Nieto R. The inventory of parent/caregiver responses to the children's pain experience (IRPEDNA): development and preliminary validation. PAIN 2008;134:128–39.
. Johnson L, Radesky J, Zuckerman B. Cross-cultural parenting: reflections on autonomy and interdependence. Pediatrics 2013;131:631–3.
. Jongudomkarn D, Forgeron PA, Siripul P, Finley GA. My child you must have patience and Kreng Jai: Thai parents and child pain. J Nurs Scholarsh 2012;44:323–331.
. Keller H. Cultures of infancy. London: Lawrence Erlbaum Associates, 2007.
. Keller H. Culture
and development: developmental pathways to psychological autonomy and hierarchical relatedness (2). Online Readings Psychol Cult 2013;6:1–14.
. Keller H, Kartner J. Development: the cultural solution of universal developmental tasks. In: Gelfand MJ, Chiu CY, Hong YY, editors. Advances in culture
and psychology. New York: Oxford University Press, 2013. p. 63–116.
. Kemmelmeier M, Burnstein E, Krumov K, Genkova P, Kanagawa C, Hirshberg MS, Erb HP, Wieczorkowska G, Noels KA. Individualism, collectivism, and authoritarianism in seven societies. J Cross Cult Psychol 2003;34:304–22.
. Kern RM, Jonyniene J. Psychometric properties of the Lithuanian version of the parenting styles
and dimensions questionnaire (PSDQ): pilot study. Fam J 2012;20:205–14.
. Kline RB. Principles and practice of structural equation modeling
. New York: Guilford Press, 2011.
. Kristjansdottir O, Unruh AM, McAlpine L, McGrath PJ. A systematic review of cross-cultural comparison studies of child, parent, and health professional outcomes associated with pediatric medical procedures. J Pain 2012;13:207–19.
. Lau AS. Japanese mothers' parenting style with preschool-age children. Masters thesis. Salt Lake City: Brigham Young University, 2006.
. Lei PW, Wu Q. Estimation in structural equation modeling
. In: Hoyle RH, editor. Handbook of structural equation modeling
. New York: Guilford Press, 2014. p. 164–80.
. Lim N. Cultural differences in emotion: differences in emotional arousal level between the east and the west. Integr Med Res 2016;5:105–9.
. Lisi D, Campbell L, Pillai Riddell R, Garfield H, Greenberg S. Naturalistic parental pain management during immunizations during the first year of life: observational norms from the OUCH cohort. PAIN 2013;154:1245–53.
. Little TD. Longitudinal structural equation modeling
. New York: Guilford Press, 2013.
. Little TD, Cunningham WA, Shahar G, Widaman KF. To parcel or not to parcel: exploring the question, weighing the merits. Struct Equ Modeling 2002;9:151–73.
. Little TD, Rhemtulla M, Gibson K, Schoemann AM. Why the items versus parcels controversy needn't be one. Psychol Methods 2013;18:285–300.
. Liu MW, Guo F. Parenting practices and their relevance to child behaviors in Canada and China. Scand J Psychol 2010;51:109–14.
. Locke L, Prinz R. Measurement of parental discipline and nurturance. Clin Psychol Rev 2002;22:895–929.
. Mackinnon D, Fairchild A, Fritz M. Mediation
analysis. Annu Rev Psychol 2007;58:593.
. Markus H, Kitayama S. Culture
and the self: implications for cognition, emotion, and motivation. Psychol Rev 1991;98:224–53.
. Matsunaga M. Item parceling in structural equation modeling
: a primer. Commun Methods Measures 2008;2:260–93.
. McGrath PJ, Unruh AM. Pain in children and adolescents. Amsterdam: Elsevier, 1987.
. McMurtry CM, McGrath PJ, Chambers CT. Reassurance can hurt: parental behavior and painful medical procedures. J Pediatr 2006;148:560–1.
. National Statistics Office of Thailand. The 2000 population and housing census (Khon-Kean province). Bangkok: Office of the Prime Minister, 2000.
. Newton-John TR, Williams ACdeC. Chronic pain couples: perceived marital interactions and pain behaviours. PAIN 2006;123:53–63.
. O'Neill MC, Pillai Riddell R, Garfield H, Greenberg S. Does caregiver behavior mediate the relationship between cultural individualism and infant pain at 12 months of age? J Pain 2016;17:1273–80.
. Ólafsson S. Contemporary Icelanders—Scandinavian or American? Scand Rev 2003;91:6–14.
. Önder A, Gülay H. Reliability and validity of parenting styles
and dimensions questionnaire. Procedia Soc Behav Sci 2009;1:508–14.
. Oyserman D, Coon HM, Kemmelmeier M. Rethinking individualism and collectivism: evaluation of theoretical assumptions and meta-analyses. Psychol Bull 2002;128:3–72.
. Oyserman D, Lee SWS. Does culture
influence what and how we think? Effects of priming individualism and collectivism. Psychol Bull 2008;134:311–42.
. Pan Y, Gauvain M, Schwartz SJ. Do parents' collectivistic tendency and attitudes toward filial piety facilitate autonomous motivation among young Chinese adolescents? Motiv Emot 2013;37:701–11.
. Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods 2008;40:879–91.
. Preacher KJ, Kelley K. Effect size measures for mediation
models: quantitative strategies for communicating indirect effects. Psychol Methods 2011;16:93.
. Quinn N. Universals of child rearing. Anthropological Theor 2005;5:477–516.
. R Core Team. R: a language and environment for statistical computing. Vienna: R Foundation for Statistical Computing, 2014.
. Riddell RP, Racine NM, Craig KD, Campbell L. Psychological theories and biopsychosocial models in paediatric pain. In: McGrath PJ, Stevens BJ, Walker SM, Zempsky WT, editors. Oxford textbook of paediatric pain. New York: Oxford University Press, 2014. p. 85–94.
. Robert C, Lee WC, Kim-Yin C. An empirical analysis of measurement equivalence with the INDCOL measure of individualism and collectivism: implications for valid cross-cultural inference. Personnel Psychol 2006;59:65–99.
. Robinson CC, Mandleco B, Olsen FS, Hart CH. Authoritative, authoritarian, and permissive parenting practices: development of a new measure. Psychol Rep 1995;77:819–30.
. Robinson CC, Mandleco B, Olsen FS, Hart CH. The parenting styles
and dimensions questionnaire (PSDQ). In: Perlmutter BF, Touliatos J, Holden GW, editors. Handbook of family measurement techniques. Thousand Oaks: Sage, 2001. p. 319–21.
. Rosseel Y. lavaan: an R package for structural equation modeling
. J Stat Softw 2012;48:1–36.
. Rowe DC, Vazsonyi AT, Flannery DJ. No more than skin deep: ethnic and racial similarity in developmental process. Psychol Rev 1994;101:396–413.
. Rudy D, Grusec JE. Correlates of authoritarian parenting in individualist and collectivist cultures and implications for understanding the transmission of values. J Cross Cult Psychol 2001;32:202–12.
. Rudy D, Grusec JE. Authoritarian parenting in individualist and collectivist groups: associations with maternal emotion and cognition and children's self-esteem. J Fam Psychol 2006;20:68–78.
. Scheidegger LP. Parenting style, assimilation, and social competence in children of southeast Asian immigrant families [Doctoral thesis]. Minneapolis: Capella University, 2007.
. Schultz E, Lavenda R, Rauch M. Cultural anthropology: a perspective on the human condition. New York: Oxford University Press, 2009.
. Shuster MM, Li Y, Shi J. Maternal cultural values
and parenting practices: longitudinal associations with Chinese adolescents' aggression. J Adolescence 2012;35:345–55.
. Singelis TM, Triandis HC, Bhawuk DPS, Gelfand MJ. Horizontal and vertical dimensions of individualism and collectivism: a theoretical and measurement refinement. Cross Cult Res 1995;29:240–75.
. Siripul P, Jongudomkarn D, Sutra P, Sakdisthanont S. Child-rearing in Isan families. Research report. Khon Kaen: Faculty of nursing, Khon Kaen University, 2007.
. Smith HE, Nieminen GS, Win MK. Historical dictionary of Thailand: Scarecrow Press, 2005.
. Smith PB, Bond MH. Social psychology across cultures. Boston: Allyn and Bacon, 1999.
. Smith TW. Developing and evaluating cross-national survey instruments. In: Presser S, Rothgeb JM, Couper MP, Lessler JT, Martin E, Martin J, Singer E, editors. Methods for testing and evaluating survey questionnaires. Hoboken: John Wiley & Sons, 2004. p. 431–52.
. Stanford EA, Chambers CT, Biesanz JC, Chen E. The frequency, trajectories and predictors of adolescent recurrent pain: a population-based approach. PAIN 2008;138:11–21.
. Statistics Canada. Halifax, RGM, Nova Scotia. 2011 National Household Survey (NHS). Ottawa: Statistics Canada, 2013.
. Statistics Iceland. Iceland in figures 2016. Reykjavik: Statistics Iceland, 2016.
. Su C, Hynie M. Effects of life stress, social support, and cultural norms on parenting styles
among Mainland Chinese, European Canadian, and Chinese Canadian immigrant mothers. J Cross Cult Psychol 2011;42:944–62.
. Suh E, Diener E, Oishi S, Triandis HC. The shifting basis of life satisfaction judgments across cultures: emotions versus norms. J Pers Soc Psychol 1998;74:482–93.
. Survey research center. Guidelines for best practice in cross-cultural surveys. Ann Arbor: Survey Research Center, Institute for Social Research, University of Michigan, 2011.
. Taras V, Rowney J, Steel P. Half a century of measuring culture
: review of approaches, challenges, and limitations based on the analysis of 121 instruments for quantifying culture
. J Int Manag 2009;15:357–73.
. Triandis HC. Some universals of social behavior. Pers Soc Psychol Bull 1978;4:1–16.
. Triandis HC. The self and social behavior in differing cultural contexts. Psychol Rev 1989;96:506–20.
. Triandis HC. Individualism and collectivism. Boulder: Westview Press, 1995.
. Triandis HC. The psychological measurement of cultural syndromes. Am Psychol 1996;51:407–15.
. Triandis HC. Culture
and social behavior. Boston: McGraw-Hill Companies, 2004.
. Triandis HC, Gelfand MJ. Converging measurement of horizontal and vertical individualism and collectivism. J Pers Soc Psychol 1998;74:118–28.
. van de Vijver FJR, Leung K. Methods and data analysis for cross-cultural research. Thousand Oaks: Sage, 1997.
. van de Vijver FJR, Leung K. Equivalence and bias: a review of concepts, models, and data analytic procedures. In: Matsumoto D, van de Vijver FJR, editors. Cross-cultural research methods in psychology. Cambridge: Cambridge University Press, 2011. p. 17–45.
. Vandenberg RJ, Lance CE. A Review and synthesis of the measurement invariance literature: suggestions, practices, and recommendations for organizational research. Organ Res Methods 2000;3:4–70.
. Vervoort T, Huguet A, Verhoeven K, Goubert L. Mothers' and fathers' responses to their child's pain moderate the relationship between the child's pain catastrophizing and disability. PAIN 2011;152:786–93.
. Vinall J, Riddell RP, Greenberg S. The influence of culture
on maternal soothing behaviours and infant pain expression in the immunization context. Pain Res Manag 2011;16:234–238.
. Wang C, Ma Y, Han S. Self-construal priming modulates pain perception: event-related potential evidence. Cogn Neurosci 2014;5:3–9.
. Wang S, Tamis-Lemonda CS. Do child-rearing values in Taiwan and the United States reflect cultural values
of collectivism and individualism? J Cross Cult Psychol 2003;34:629–42.
. Wu P, Robinson C, Yang C, Hart C, Olsen S, Porter C, Jin S, Wo J, Wu X. Similarities and differences in mothers' parenting of preschoolers in China and the United States. Int J Behav Dev 2002;26:481–91.
. Yaman A, Mesman J, van Ijzendoorn M, Bakermans-Kranenburg M, Linting M. Parenting in an individualistic culture
with a collectivistic cultural background: the case of Turkish immigrant families with toddlers in the Netherlands. J Child Fam Stud 2010;19:617–28.
Culture; Ecosocial context; Cultural values; Horizontal–vertical individualism–collectivism; Parenting styles; Authoritative-parenting; Authoritarian-parenting; Pain-related parental behaviors; Solicitousness; Discouraging; Structural equation modeling; Mediation; Moderation
Supplemental Digital Content
© 2018 International Association for the Study of Pain