RELATIONSHIPS AND POTENTIAL MECHANISMS BETWEEN PARENTAL SOCIODEMOGRAPHIC CHARACTERISTICS, CONCERNS, AND CHILDREN’S PHYSICAL ACTIVITY DURING THE COVID-19 PANDEMIC
Received: 2024-03-08 Revised: 2024-03-21 ; Accepted: 2024-04-20
Published Online: 2024-08-22
Abstract
Physical activity (PA) is essential for children’s health, yet the majority of children in Canada do not meet the recommended levels, a situation worsened by the COVID-19 pandemic. This study examined the relationships and potential mechanisms between parental sociodemographic characteristics, parental concerns, and children's PA. The study utilized proxy-reported data from the 2020 Statistics Canada survey, Impacts of COVID-19 on Canadians - Parenting during the Pandemic (n = 5,863 parents of children aged 6-14 years). Multiple logistic regressions, Spearman correlations, and mediation analyses were employed. Nearly half of the parents expressed high concerns about their children’s PA (47.7%) and reported that their children had low to moderate PA levels during the pandemic (46.4%). Parental concerns regarding the amount of children’s PA was the key predictor of children’s PA with large effect size. Parental concerns also partially explained the relationship between parental visible minority status and children’s PA (47% of variance explained) as well as between parental immigrant status and children's PA (33% of variance explained). These results suggest that while the pandemic's impact might have been uniform across families, parental concerns associated with visible minority and immigrant statuses likely contributed to lower levels of PA among their children. As we continue to navigate the recovery process from the pandemic, future strategies promoting children’s PA could prioritize supporting parents from marginalized communities with increased concerns regarding their children’s PA.
본문
1. INTRODUCTION
The health benefits of regular physical activity (PA) among children are well-established (Poitras et al., 2016). In 2016, Canadian researchers developed the 24-Hour Movement Guidelines, which provide behavioural recommendations for PA, sedentary behaviour, and sleep for children and youth aged 5-17 years (Tremblay et al., 2016). Specifically, the guidelines recommend that children and youth engage in at least 60 minutes of moderate- to vigorous-intensity PA (MVPA) daily (Tremblay et al., 2016). Despite these efforts, most recent comprehensive data indicated that 39% of Canadian children and youth met the PA recommendation, with significant disparities observed by gender and immigrant status (ParticipACTION, 2024).
The COVID-19 pandemic, which was declared a public health emergency in early 2020, likely exacerbated already low levels of PA among children in Canada (Guerrero et al., 2023; Moore et al., 2021). Community-wide lockdowns and stay-home orders, implemented to curb the spread of COVID-19, further contributed to this decline. According to a ParticipACTION report (2022), only 37% of children and youth aged 5-17 years met the PA recommendations in the fall of 2020, a 14% decrease from the pre-pandemic level of 51%. Reviews conducted during the pandemic revealed a significant reduction in children's PA just one year into the pandemic (Paterson et al., 2021; Rossi et al., 2021). Reduced PA levels due to the COVID-19 pandemic appear to persist until today, compounded by other public health challenges such as climate change (ParticipACTION, 2024) and may continue well into the future. This highlights the importance of promoting and fostering PA participation among children as we move forward in the post-COVID-19 era.
Socioecological modelling of physical activity posits that health behaviour is influenced by multiple levels of correlates (Spence & Lee, 2003). Among these layers, interpersonal level correlates such as parental age, gender, race, ethnicity, immigrant status, and socioeconomic status (SES) have all been identified as key correlates (Rhodes et al., 2020). Additionally, parental concerns about their children's PA may also play a key role in their children's PA. For example, children of parents who are more concerned about their PA levels tend to be less physically active compared to those whose parents are less concerned (Jackson et al., 2008). This trend persisted even during challenging times such as the COVID-19 pandemic, where heightened parental anxiety about the virus was linked to reduced visits to parks and increased screen time among children (McCormack et al., 2020). As the COVID-19 pandemic has transitioned into the endemic phase, incorporating learnings from the pandemic phase could be beneficial in promoting PA among children more effectively.
While previous research has explored the individual links between parental sociodemographic characteristics, parental concerns, and children’s PA, the underlying mechanisms connecting these variables have remained unexamined. Given the potential interplay of parental factors influencing children’s PA within a complex socioecological framework (Spence & Lee, 2003), and the established association between parental sociodemographic characteristics and concerns regarding children’s PA (Hornby-Turner et al., 2014; Trigwell et al., 2015), this study postulated that parental concerns could elucidate the relationship between specific parental sociodemographic characteristics and children’s PA. For instance, parents with immigrant status might harbor heightened concerns about their children’s PA during the COVID-19 pandemic due to fears of infection (McCormack et al., 2020) or anti-immigrant sentiment that were prevalent during the pandemic ( Esses & Hamilton, 2021).
The hypothesized pathway explaining the relationship between parental sociodemographic characteristics, parental concerns for children’s PA, and children’s PA is indicated in Figure 1. Specifically, it was hypothesized that parental concerns for PA would mediate the relationships between parental sociodemographic characteristics and children’s PA. The overall objective of the study was to examine the relationships and potential mechanisms between parental sociodemographic characteristics, parental concerns for children’s PA, and children’s PA among a sample of parents in Canada during the COVID-19 pandemic.
2. METHODOLOGY
This study used the data from the Impacts of COVID-19 on Canadians - Parenting during the Pandemic 2020, which is a cross-sectional crowdsourcing survey consisting of parents with children aged less than 15 years old from various regions of Canada (Statistics Canada, 2020). The aim of the survey was to explore the impact of the COVID-19 pandemic on parental concerns and experiences related to their children's health and social life. Data were collected online from June 9th to 22nd, 2020, using a self-administered, proxy-reported questionnaire (http://www.statcan.gc.ca/COVIDparenting-questionnaire). Participation in the survey was voluntary and facilitated through diverse platforms such as social media, agencies, public/private organizations, and news channels. Participants needed to meet two inclusion criteria to participate in the survey: 1) residing in one of the 10 eligible provinces of Canada and 2) being able to read in either English or French. A non-probabilistic recruitment approach was employed, and the questionnaire included various items related to family composition, health and social impacts, labor market impacts, and sociodemographic characteristics. A total of 32,228 parents completed the questionnaire, but only parents of children aged 6-14 years (n = 5,863) were included in the analysis. The study incorporated variables indicating parental sociodemographic characteristics, parental concerns, and their children’s PA. Implicit consent was obtained when participants chose to complete the questionnaire. Ethics approval was not required as the study relied solely on anonymized secondary data, ensuring the protection of participants' privacy and confidentiality.
Parental sex (male, female), education, visible minority status, immigrant status, employment status, and working arrangement during the pandemic were used to indicate parental sociodemographic characteristics. Parents were asked to indicate their highest degree obtained with the following response options: “Less than high school diploma or its equivalent,” “High school diploma or a high school equivalency certificate,” “Trade certificate or diploma,” “College/CEGEP/other non-university certificate or diploma,” “University certificate or diploma below the bachelor’s,” “Bachelor’s degree university certificate,” “Diploma,” or “Degree above the BA level.” These were subsequently grouped into two categories: < university education (College/CEGEP/other non-university certificate or diploma” and below) and ≥ university education (“University certificate or diploma below the bachelor’s” and above). For visible minority status, parents were given the following response options: “White,” “South Asian,” “Chinese,” “Black,” “Filipino,” “Arab,” “Latin American,” “Southeast Asian,” “West Asian,” “Korean,” “Japanese,” or “Other.” These were subsequently grouped into two categories: visible minority (“South Asian,” “Chinese,” “Black,” “Filipino,” “Arab,” “Latin American,” “Southeast Asian,” “West Asian,” “Korean,” “Japanese,” and “Other”) and non-visible minority (“white”). To determine the immigrant status of parents, three questions were asked: “Where were you born?” with response options “Born in Canada” or “Born outside Canada”; “Are you a Canadian citizen?” with response options “Yes, a Canadian citizen”, “Yes, a Canadian citizen by naturalization” or “No, not a Canadian citizen.”; and “Are you a landed immigrant or permanent resident?” with response options “Yes” (landed immigrant or permanent resident) or “No.” These responses were used to categorize parents as either immigrants or Canadians. Parents were also asked to indicate changes in employment status in their household due to the COVID-19 pandemic with the following question: “Someone in my family lost their job, was laid off, or has reduced work hours due to COVID-19,” with response options of “Yes” or “No.” For working arrangements during the pandemic, parents were asked the following question: “During the COVID-19 pandemic, which of the following statements apply to family members living in your home? 1) “Someone in my family is working at a fixed location outside the home”; 2) “Someone in my family is working outside the home with no fixed location”; and 3) “Someone in my family is working from home,” with response options of “Yes” or “No.” For statistical purposes, the responses were reclassified as working from home, working outside, and mixed. Responses marked as “Not Stated” in the above items were recoded as missing and excluded from the analysis.
Parental concerns were operationalized as concerns about the amount of PA their child was obtaining during the COVID-19 pandemic. The survey asked parents, “Due to the COVID-19 pandemic, how concerned are you about the following for your child or children aged 0 to 14 years? – Amount of physical activity.” Response options included “Not at all,” “Somewhat,” “Very,” or “Extremely.” Parental concerns were categorized into two groups: low concern (“Not at all” or “Somewhat”) and high concern (“Very” or “Extremely”). Responses marked as “Not Stated” were recoded as missing and excluded from the analysis.
To indicate the levels of children’s PA during the pandemic, parents were asked the following question: “On average, during the COVID-19 pandemic, how often has your child or have your children engaged in any of the following activities at home? - Physical activities.” Response options included “Never,” “1 to 2 times per week,” “3 to 5 times per week,” and “Daily or almost every day.” Children’s PA was categorized into two groups: low to moderate PA (“Never,” “1-2 times per week,” or “3 to 5 times per week”) and high PA (“Almost every day” or “Daily”). “Not Stated” and “Valid skip” responses were recoded as missing and were excluded from the analysis.
Parental sociodemographic variables were simultaneously used as covariates in all analyses given their established relationships with children’s PA (Rhodes et al., 20200).
IBM SPSS Version 28.0 (IBM Corp) was used for all statistical analyses. To account for the non-randomized data collection techniques employed in the survey, the Complex Samples (CS) plan procedure was utilized. CS frequency statistics and corresponding 95% confidence intervals (95% Cl) were calculated to describe the sample characteristics. CS multivariate logistic regressions were also utilized to explore the associations between the key variables and children’s PA.
To explore the potential role of parental concerns in the relationship between parental sociodemographic characteristics and children’s PA, correlations among parental sociodemographic characteristics, concerns, and children’s PA were first tested using Spearman correlations. To be eligible for mediation analysis, each parental sociodemographic variable must show a statistically significant correlation with both parental concerns and children's PA. Variables demonstrating either no correlations (r = 0.00) or very high correlations (r > 0.90) were removed from the mediation analyses (data-original-title=""Tabachnick et al., 2013).
If a variable met the criteria for further mediation analysis, we conducted a mediation analysis using the PROCESS macro for SPSS, utilizing the bootstrap method to assess the indirect effects of parental sociodemographic characteristics on children’s PA through parental concerns. We examined the total effect (c), direct effect (c’), and indirect effect (ab) based on unstandardized beta coefficients and 95% CIs for the indirect effect exclusively, as suggested when dealing with binary variables (Hayes, 2017). This is because standardized coefficients are influenced by the distribution of cases across categories and differences in group means. Furthermore, standardized beta coefficients do not facilitate comparison of average values and relative indirect effects between groups. Thus, we relied on unstandardized coefficients (B), which represent the mean difference between each path in mediation models (Hayes, 2017).
To assess the magnitude of the mediation effect, Hair Jr and colleagues (2019) proposed using Variance Accounted For (VAF) values calculated by dividing the indirect effect by the total effect. A VAF exceeding 80% indicates complete mediation, while a value between 20% and 80% suggests partial mediation. A VAF below 20% indicates no mediation (Hair Jr et al., 2010). We ensured that assumptions for multiple logistical regression and mediation analysis were met to avoid any violations. The threshold for statistical significance was set at p < .001 for all analyses, given the large sample size.
3. RESULTS
Table 1 describes the participant characteristics consisting of 5,863 parents. A majority of parents who completed the survey identified as female (91.2%), with a significant proportion holding a university education level or higher (74.3%). Among family members previously active in the workforce pre-pandemic, 67.9% reported no job loss or reduction in hours, while 53.4% indicated working from home, and 30.9% reported a mixed working arrangement. Regarding parental demographic attributes, 91.9% identified as white, and 89.3% reported Canadian citizenship. Concerning parental concerns about their children's PA during the COVID-19 pandemic, 47.7% expressed high levels of concern. Regarding children's PA levels during the pandemic, it was noted that 53.6% engaged in high levels of PA.
| Key variables | % (95% CI) |
| n | 5,863 |
| Parental characteristics | |
| Sex Female Male |
91.2 (90.2-92.0) 8.8 (8.0-9.8) |
| Education levels < University degree ≥ University degree |
25.7 (24.4-7.1) 74.3 (72.9-75.6) |
| Work arrangement of family members Working from home Working outside Mixed |
53.4 (51.9-54.9) 15.7 (14.6-16.8) 30.9 (29.5-32.3) |
| Employment status (Lost job/reduced hours) Yes No |
32.1 (30.7-33.6) 67.9 (66.4-69.3) |
| Visible minority status Visible minority Non-visible minority (white) |
8.1 (7.4-9.0) 91.9 (91.0-92.6) |
| Immigrant status Immigrant or non-permanent resident Canadian/permanent resident |
10.7 (9.8-11.7) 89.3 (88.3-90.2) |
| Concerns for their children’s PA during the COVID-19 pandemic Low High |
53.3 (50.8-53.9) 47.7 (46.1-49.2) |
| Children’s characteristics | |
| Physical activity (PA) Low to moderate PA (0-5 times a week) High PA (6-7 times per week) |
46.4 (44.9-48.0) 53.6 (52.0-55.1) |
Table 2 presents the results of the associations between parental sociodemographic characteristics, parental concerns, and children’s PA. Parents without a university education were more likely to have children engaged in low to moderate PA/week compared to those with a university education or higher (OR 1.38, 95% CI 1.18,1.62). Additionally, parents who expressed high concerns for their children's PA engagement during the COVID-19 pandemic were more likely to report their children participating in low to moderate PA/week compared to parents with low concerns (OR 3.91, 95% CI 3.44, 4.46).
| Parental sociodemographic characteristics and concerns | Children’s PA |
|
OR (95%) |
||
Sex |
|
|
Education levels |
|
|
Work arrangement of family members |
|
|
Employment status (Lost job/reduced hours) |
|
|
Visible minority status |
|
|
Immigrant status |
|
|
Concerns for PA during COVID-19 |
|
|
|
|
|
Table 3 presents the correlations among parental sociodemographic characteristics, parental concerns, and children’s PA. The analysis indicated that parental immigrant and visible minority statuses were the only parental sociodemographic characteristics exhibiting statistically significant correlations with parental concerns and their children’s PA. Specifically, visible minority status (rho = 0.08, p < .001) and immigrant status (rho = 0.06, p < .001) demonstrated positive correlations with concerns related to PA. Furthermore, parental visible minority status (rho = 0.04, p < .001) and immigrant status (rho = 0.04, p < .001) were positively correlated with children’s PA. Consequently, two separate mediation models were subsequently conducted. The first model examined parental visible minority status as the exogenous variable, with parental concerns considered as a potential mediator in the relationship between parental visible minority status and children's PA. The second model investigated parental immigrant status as the exogenous variable, with parental concerns examined as a potential mediator in the relationship between parental immigrant status and children's PA.
| Key variables | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
|---|---|---|---|---|---|---|---|
| 1. Sex | 0.04 | -0.07 | 0.05 | 0.00 | -0.02 | -0.01 | -0.03 |
| 2. Education levels | - | -0.17 | 0.15 | -0.09 | -0.10 | -0.01 | 0.06 |
| 3. Work arrangement | -0.17 | - | -0.09 | 0.03 | 0.06 | -0.02 | 0.00 |
| 4. Employment status | 0.15 | -0.09 | - 0.02 | 0.03 | -0.02 | -0.01 | |
| 5. Visible minority status | -0.09 | 0.03 | 0.02 | - 0.40 | 0.08 | 0.04 | |
| 6. Immigrant status | -0.10 | 0.06 | 0.03 | 0.40 | - 0.06 | 0.04 | |
| 7. Parental concerns | -0.01 | -0.02 | -0.02 | 0.08 | 0.06 | - 0.32 | |
| 8. Children’s PA | 0.06 | 0.00 | -0.01 | 0.04 | 0.04 | 0.32 | - |
In Figure 2, the results of a mediation analysis between parental visible minority status, parental concerns, and children’s PA are presented. Path a, representing the relationship between parental visible minority status (1 = visible minority status; 0 = non-visible minority) and parental concerns, suggested that there is a negative relationship between parental visible minority status and parental concerns for their children’s PA (path a: B = -0.25, p < .001). This indicates that parents of visible minority status reported higher concerns. Path b shows the relationship between parental concerns and children's PA, indicating that with higher parental concerns, lower PA among children is observed (path b: B = -0.30, p < .001). Statistically significant path c (B = 0.15, p < .001) indicated the positive relationship between parental visible minority status and children’s PA. Path c’ (B = 0.08, p = .03) was no longer statistically significant, with the inclusion of parental concerns as a mediator. Parental concerns demonstrated a statistically significant indirect effect (path ab: B = 0.07, 95% CI: 0.05, 0.10) on the relationship between parental visible minority status and children's PA, which indicates that parental concerns partially explain the relationship between parental visible minority status and children’s PA (47% variance explained).
In Figure 3, the results of a mediation analysis between parental immigrant status, parental concerns, and children’s PA are presented. Path a, representing the relationship between parental immigrant status and parental concerns, suggested that there is a negative relationship between parental immigrant status and parental concerns for their children’s PA (path a: B = -0.11, p < .001). Path b showed the relationship between parental concerns and children's PA, indicating that with higher parental concerns, lower PA among children is observed (path b: B = -0.30, p < .001). Path c (B = 0.09, p = .007) and Path c’ (B = 0.06, p = .06) were not statistically significant. However, parental concerns showed a statistically significant indirect effect (path ab: B = 0.03, 95% CI 0.01, 0.06) on the relationship between parental immigrant status and their children’s PA, indicating that parental concerns partially explained the relationship between parental immigrant status and children’s PA (33% variance explained).
5. DISCUSSION
This study explored the relationships and potential pathways linking parental sociodemographic characteristics, parental concerns regarding the amount of children’s PA, and children's PA during the COVID-19 pandemic. Our multivariate analyses identified parental education and parental concerns regarding the amount of their children’s PA as significant correlates of children’s PA during the COVID-19 pandemic. Furthermore, mediation analyses indicated that parental concerns for children’s PA appear to partially explain the relationships between parental visible minority and immigrant statuses and children’s PA.
A general decline in PA among children during the pandemic has been consistently reported across independent studies and reviews conducted on children’s PA during the COVID-19 pandemic (Guerrero et al., 2023; Moore et al., 2021; Knight et al., 2022; Moore et al., 2021; ParticipACTION, 2022; Paterson et al., 2021; Rossi et al., 2021). In general, household income (Mitra et al., 2020), parental anxiety (McCormack et al., 2020), and changes in working arrangement due to the pandemic (Pombo et al., 2020) were identified as correlates of children’s PA during the pandemic (Knight et al., 2022; Paterson et al., 2021), while evidence regarding parental education has been largely mixed (Medrano et al., 2021; Mitra et al., 2020; Ruíz-Roso et al., 2020).
In our multivariate regression model, only parental education and concerns regarding the amount of children’s PA emerged as significant factors associated with children’s PA levels. Notably, parental concerns exhibited the most substantial association with children’s PA, with large effect sizes. This finding is consistent with previous research (Jackson et al., 2008; McCormack et al., 2020), suggesting that parental concerns and anxiety are associated with low levels of PA among their children. Combined, it can be speculated that parental concerns, in general, were heightened during the COVID-19 pandemic due to the perceived risk of infection, potentially hindering children from achieving sufficient levels of PA throughout the week.
Although data is largely lacking, pre-pandemic evidence also suggests that parental concerns about children's PA may be influenced by sociocultural factors tied to parental sociodemographic characteristics. For instance, ethnically diverse parents in the UK were reported to have heightened concerns about their children's engagement in PA due to concerns for safety, adverse weather events, lack of resources, and lack of access (Trigwell et al., 2015). These concerns likely influence children’s PA across varying racial/ethnic groups within a society. For example, one study found that British Pakistani girls encounter more barriers to PA than their white British counterparts, partly due to increased parental concerns regarding safety (Hornby-Turner et al., 2014). Nevertheless, this remains speculative due to the lack of corresponding data in Canada. This underscores the need for future research to investigate the influence of visible minority and immigrant statuses on children’s PA.
Canadian data regarding the potential influence of parental sociodemographic characteristics, particularly related to visible minority and immigrant statuses, has been lacking. However, both pandemic and endemic data suggested that PA is generally lower among children and youth with immigrant status (ParticipACTION, 2022, 2022). In our multivariable model, neither visible minority nor immigrant status was associated with children’s PA, suggesting that the COVID-19 pandemic affected children’s PA equally regardless of these parental characteristics. However, both visible minority and immigrant statuses were positively correlated with higher parental concerns regarding the amount of PA among children in our data, prompting subsequent mediation analyses. The observed mediating role of parental concerns in relation to parental visible minority and immigrant statuses in our study could be linked to the increased vulnerability faced by individuals with these memberships during the COVID-19 pandemic, which may have, in turn, heightened their anxiety during this period (McCormack et al., 2020). This vulnerability may have stemmed from both the health threats posed by the virus and the increase in anti-immigrant and anti-Asian sentiment prevalent during this period (Esses & Hamilton, 2021). Another important thing to note is that parental visible minority and immigrant statuses may not have been mutually exclusive as it was shown in the moderate strength correlation coefficient between the two variables (rho = 0.40, p <.001). This suggests that parents who are visible minorities are also likely to be immigrants. The potential overlap between these memberships and the findings of our study, combined, indicates that future research should investigate this through an intersectional lens.
Intersectionality is a theoretical framework coined by Kimberlé Crenshaw, which examines how various social identities such as race, gender, class, and immigration status intersect and create unique modes of discrimination and privilege ( Crenshaw, 1989 ). It acknowledges that individuals can simultaneously belong to multiple marginalized groups, leading to complex and multifaceted experiences of oppression and resilience. Applying intersectionality in future research involves examining how overlapping social identities (e.g., visible minority status and immigrant status) interact to influence parental concerns and children’s PA during crises like the COVID-19 pandemic. This approach would help to uncover the nuanced ways in which these intersecting identities shape vulnerability, coping mechanisms, and access to resources. By incorporating intersectional analysis, researchers can better understand the compounded effects of multiple forms of discrimination and develop more targeted and effective interventions to support these populations (Lee et al., 2023).
This study was conducted during the COVID-19 pandemic to investigate potential mechanisms linking parental characteristics and children's PA using a mediation analysis technique, highlighting the role of parental concerns in children’s PA. The findings suggest that while the risk of COVID-19 infection may have been equal regardless of parental visible minority and immigrant statuses, children’s PA is impacted differentially due to parental concerns tied to these parental characteristics. Nevertheless, the study has several limitations inherent to its survey design and data analysis. The cross-sectional nature of the study prevents making causal inferences because temporality cannot be established. Therefore, the results of mediation analyses should be interpreted with caution. Participants were recruited through convenience sampling in the survey used, which limits its generalizability to the broader Canadian population. Additionally, the study relied on proxy-reported data on all variables including children’s PA, introducing potential recall and social desirability biases. The questions regarding children’s PA were limited to home settings due to nation-wide lockdowns during the data collection period. Future research should aim to address these limitations by employing longitudinal designs, representative sampling methods, and more accurate measures of parental concerns and children’s PA.
6. CONCLUSIONS
This study investigated the relationships and mechanisms between parental sociodemographic characteristics, parental concerns regarding the amount of their children’s PA, and children’s PA during the COVID-19 pandemic in a convenient sample of parents in Canada. The results highlight that the impact of the pandemic may have been equal for all families, regardless of parental visible minority and immigrant statuses; however, higher concerns associated with parents holding these memberships may have contributed to lower levels of PA among their children. To confirm and build on these findings, future research should employ more robust methodologies. Specifically, by using comprehensive data collection, longitudinal studies, and intersectional analysis, researchers can gain a deeper understanding of how parental visible minority and immigrant statuses intersect to influence parental concerns and children's PA levels. This will help in developing more effective strategies to address these concerns and promote healthier lifestyles among children from these communities during times of public health crises.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.