Abstract
PURPOSE
To assess the association of dog walking with adolescents’ moderate-to-vigorous physical activity (MVPA) and BMI, and identify correlates of dog walking.
METHODS/DESIGN
Participants were 12–17 year-olds (n=925) from the Baltimore, MD and Seattle, WA regions. Differences in accelerometer-assessed minutes/day of MVPA and self-reported BMI (percentile) were compared among adolescents (1) without a dog (n=441) and those with a dog who (2) did (≥1 days/week, n=300) or (3) did not (n=184) walk it. Correlates of (1) dog walking (any vs. none) among adolescents with dogs (n=484), and (2) days/week of dog walking among dog walkers (n=300) were investigated. Potential correlates included: demographic, psychosocial, home environment, perceived neighborhood environment, and objective neighborhood environment factors.
RESULTS
52% of adolescents lived in a household with a dog, and 62% of those reported dog walking ≥1 day/week. Dog walkers had 4–5 more minutes/day of MVPA than non-dog-walkers and non-dog-owners. BMI was not associated with dog walking or ownership. Among households with dogs, adolescents who lived in objectively walkable neighborhoods were 12% more likely to walk their dog than those in less walkable neighborhoods. Among dog walkers, having a multi-family home, college-educated parent, lower perceived traffic safety, higher street connectivity and less mixed use were related to more days/week of dog walking.
CONCLUSIONS
Dog walkers had 7–8% more minutes/day of MVPA than non-dog walkers, and correlates of dog walking were found at multiple levels of influence. Results suggest multilevel interventions that include both environmental and psychosocial components to increase dog walking should be evaluated.
Keywords: dog walking, physical activity, walkability, ecological models, built environment, obesity
BACKGROUND
The National Health and Nutrition Examination Survey (NHANES) found that only 8% of US adolescents met the recommended 60 minutes of physical activity a day, based on objective measures.1 While 84% of adolescents reported walking as a source of physical activity,2 GPS-measured minutes of walking in this population appear low.3 Therefore, walking may be a promising approach to increase adolescents’ physical activity.
Because nearly half of US households have a dog,4 dog walking could be an important contributor to physical activity, but many adult and adolescent dog owners report little or no dog walking.5–7 A meta-analysis of 17 studies found that dog ownership and dog walking were associated with greater overall physical activity. Only 4 studies used objective measures of physical activity, and few studied adolescents or children.8 A review of 9 dog walking studies among adults calculated the odds of meeting moderate intensity physical activity guidelines and concluded that dog walkers were 2.5 times more likely to meet the guidelines.9 Identifying factors, like motivators and barriers, related to dog walking is important because results can inform interventions to increase dog walking.10
Multiple levels of correlates should be examined, because ecological models posit that variables at individual, social, community environment, and policy levels influence behaviors.11 Correlates of dog walking in previous studies included those at the individual level (i.e., race/ethnicity, income, illness), social level (i.e., social support, walking as a family, neighborhood social cohesion), perceived environment level (i.e., perceived crime) and objective environment level (i.e., weather and neighborhood walkability).6,12,13 Few studies applied principles of ecological models by examining multiple levels and interactions (i.e. moderators) across levels.
A systematic review of dog walking studies found only 2 studies examined children or adolescents,5 and 18% of youth aged 10–12 years walked their dogs at least 3 times per week.6,7 One of the studies found owning a dog was associated with 29 additional minutes of moderate to vigorous physical activity (MVPA) per day among younger female children, yet no effects for males or older females.6 Children who lived in households with dogs were 49% more likely to achieve physical activity recommendations.14 The current study filled gaps in the literature by quantifying the contribution of dog walking to objectively-measured total physical activity in adolescents and investigating a broader range of correlates of dog walking at multiple levels.
The first objective of the present paper was to quantify the difference in MVPA and weight status (i.e. BMI) between adolescents living in households 1) without dogs, 2) with a dog but did not walk it, and 3) who reported any dog walking. A second objective was to explore the subsample of dog owners to assess ecological correlates of walking the dog at all versus none. A third objective was to assess correlates of dog walking frequency (days/week) among dog walkers. A final objective was to explore cross-level interactions in both dog household subsamples (i.e. dog owners and dog walkers) to identify moderators of associations.
METHODS
Study design and Participants
The present study used data from the Teen Environment and Neighborhood (TEAN) observational study.3,15 Participants were adolescents aged 12–17 living in the Seattle, WA or Baltimore, MD regions in 2009–2011 (n=925). Participants were one adolescent and one parent/guardian selected from neighborhoods (i.e., census block groups) defined by high or low walkability (based on GIS measures of built environment factors) and stratified by high or low income (based on Census 2000 data), similar to methods described previously.16 Households with adolescents in selected block groups were identified from a marketing company and recruited by mail and telephone. Overall participation rate was 36% and did not vary by quadrant. Compared to Census demographics, the study sample had somewhat higher education and household income. Adolescents and parents each completed a survey to assess demographics, psychosocial characteristics and perceived neighborhood environment (available at http://sallis.ucsd.edu/Documents/Measures_documents/TEAN%20Survey%20ADOL%20FINAL%20010509.pdf). Adolescents wore an accelerometer for one week to determine daily minutes of MVPA. The Institutional Review Board of San Diego State University approved the study, parents/guardians signed informed consents, and adolescents signed assent forms.
Measures
Dog ownership and dog walking (survey data)
Adolescents were asked if their family owned a dog (yes/no). If yes, the adolescent was asked how many days a week he/she walked the dog (0 to 7 days).
Psychosocial and perceived environment variables (survey data)
Self-efficacy for physical activity was determined by asking the adolescents 6 items that assessed confidence in doing physical activity despite barriers (e.g., “do physical activity even when the weather is bad, or when sad or stressed”). Response options ranged from 1 = “I’m sure I can’t” to 5 = “I’m sure I can” and were averaged to create a scale (Cronbach’s alpha = .76; test-retest intraclass correlation coefficient (ICC) = .71).17
Decisional balance for physical activity was assessed with 5 “pro” items (Cronbach’s alpha = .81; test-retest ICC = .74) and 5 “con” items (Cronbach’s alpha = .53; test-retest ICC = .86) where each item was rated from 1 = strongly disagree to 4 = strongly agree.17 “Pro” items focused on benefits of physical activity (e.g., would have fun) and the “con” items focused on negatives of physical activity (e.g., time away from being with friends). Decisional balance was measured by subtracting the mean for the 5 “cons” items from the mean of the 5 “pros” items, resulting in a variable ranging from −5 to 5.
Enjoyment was measured with 1 item asking whether the adolescent enjoyed doing physical activity, with response options ranging from 1 = “strongly disagree,” to 5 = “strongly agree.”
Rules were measured by having adolescents report on 13 rules (yes/no) their parent(s) enforced regarding physical activity (e.g. “come in before dark,” “do not go places alone”) (Cronbach’s alpha = .87; test-retest ICC = .68; unpublished data) with items summed to create an index.
Adolescents were asked whether they owned 4 types of portable electronics (e.g., cell phone, iPod/MP3 player), yielding a summed score ranging from 0 to 4. Participants reported which of 6 electronic devices were in their bedroom (e.g. TV, computer), yielding a summed score ranging from 0 to 6 (test-retest ICC = >.60 for both scales).18
A subset of the Neighborhood Environment Walkability Scale for Youth (NEWSY) was completed by both the adolescent and parent. Parent sections included neighborhood aesthetics with 4 items (e.g. interesting things to look at), traffic safety with 3 items (e.g. most drive above the speed limit), pedestrian safety with 3 items (e.g. crosswalks and signals present), crime safety with 1 item (high crime rate), and stranger danger with 4 items (e.g. afraid of my child being taken or hurt by stranger). Response options ranged from 1 (strongly disagree) to 4 (strongly agree) where larger numbers represented more favorable conditions for physical activity. The adolescent sections of the NEWS-Y included traffic safety, pedestrian safety, crime safety and stranger danger. Means of item values were calculated for multiple item sections. Test-retest ICCs for subscores ranged from 0.61 to 0.78 for adolescents and parents.19
Weight Status (survey data)
In the survey, adolescents were provided instructions on how to accurately measure and record their weight and height. BMI percentiles were based on CDC BMI-for-age growth charts.20
Objective built environment (GIS data)
Built environment features were derived from county tax assessor data, regional land use at the parcel level, and street networks and integrated into GIS. Variables were calculated for 1 kilometer street network buffers around participants’ homes.16 A walkability index was created by summing the sample z-scores for each of 4 built environment measure: (1) housing units per residential land area, (2) intersection density, (3) retail floor area ratio, and (4) mixed use including residential, retail, food and entertainment, and office land use.16 Higher index values represented more walkable neighborhoods.
Overall physical activity (accelerometer data)
Enrolled adolescents were mailed an Actigraph accelerometer (models 7164/71256 or GT1M/GT3X with Normal filter) with instructions to wear the device for 1 week. Participants wore the accelerometer on a belt at their left iliac crest with acceleration captured at 30-second epochs. Minutes/day of MVPA were scored using the Freedson 3-MET age-based cut points.21,22 Days were removed from the scoring if the participant did not wear the accelerometer for at least 10 valid hours (range 0–17 valid days in present analyses). Strings of >60 sequential 30-second epochs within an hour with no movement (i.e. count=0) were considered not valid.
Statistical Analyses
All models utilized mixed effects regression in SPSS V.22.0 to adjust for nesting of participants within block groups. Participant characteristics, potential dog walking correlates, and overall daily MVPA and BMI percentile were compared across adolescents in households without dogs (n=441), adolescents who had a dog but did not walk it (n=184), and those who reported any dog walking (≥1 day/week) (n=300). In dog owners, total MVPA was also regressed on days/week of dog walking to assess the contribution of each additional day of dog walking. Models were adjusted for demographic covariates (adolescent age, gender, white Non-Hispanic race/ethnicity, parent married/living with a partner, parent with college degree, house type [e.g. apartment, single family]).
Correlates of dog walking were investigated in (1) all adolescents who lived in households with a dog (n=484) and (2) adolescents who walked their dog at least one day a week (n=300). In the subsample of adolescents in households with dogs, the dependent variable was walking the dog at all vs. no walking. In the smaller subsample of only those who walked their dog (at all), the number of days per week (1–7) the adolescent walked the dog was the dependent variable. Independent variables/correlates were first analyzed in 6 separate models based on levels of ecological models and source of data: (1) demographic, (2) psychosocial, (3) home environment, (4) parent’s perceived neighborhood environment and (5) adolescent’s perceived neighborhood environment, and (6) objective neighborhood built environment. All subsequent models were adjusted for the demographic factors tested in the initial model. Independent variables with P<0.15 were entered into a final, cross-level model. All potential cross-level interactions were tested using a backwards stepwise approach, where interaction terms were removed one at a time until only terms with P<0.1 remained. The independent variables were grand mean centered to create orthogonal interaction terms so the intercept would approximate the sample mean for dog walking days/week in the subsample of dog walkers. Interaction visuals were plotted using one standard deviation above and below the mean for continuous variables to represent high and low supportiveness of dog walking. Unstandardized regression coefficients (B) are reported and can be interpreted as the change in the dependent variable for a 1-unit change in the independent variable. In the final model, standardized regression coefficients are also reported so effects can be compared across variables.
RESULTS
Sample characteristics
The study sample was composed of adolescents with a mean age of 14 and was almost equally split between males and females. About two thirds of participants were White Non-Hispanic, and over 80% lived in a household with a parent who was married or living with a partner and in a single-family home (Table 1). The only significant difference in demographic characteristics was that households with dogs (regardless if they walked the dog or not) had a greater proportion of White Non-Hispanic adolescents (73%) compared to households without dogs (59%) (P<0.001; Table 1).
Table 1.
Overall sample characteristics of adolescents (N=925), living in households without dogs (n=441), households with dogs but do not walk the dogs (n=184) or households with dogs and walk the dogs ≥1 day per week (n=300) in the Seattle and Baltimore regions
| Variables of interest within ecological model levels | Total Sample Characteristics (N=925) Mean (SD) or n (%) |
Households without dogs (n=441) Mean (SD) or n (%) |
Households with dogs non-dog walkers (n=184) Mean (SD) or n (%) |
Households with dogs walkers (n=300) Mean (SD) or n (%) |
|---|---|---|---|---|
| Individual Characteristics | ||||
| Adolescent Age | 14.09 (1.40) | 14.12 (1.37) | 14.16 (1.45) | 14.02 (1.39) |
| Adolescent Gender (Male) | 460 (49.6%) | 230 (52.2%) | 91 (49.5%) | 137 (45.7%) |
| Adolescent White Non-Hispanic | 611 (66.3%) | 258 (58.9%) | 135 (73.8%) | 217 (72.8%) |
| Parent married/living with a partner | 774 (83.9%) | 363 (83.1%) | 159 (86.9%) | 251 (84.0%) |
| Parent with college degree | 695 (75.4%) | 334 (76.3%) | 132 (72.1%) | 227 (75.9%) |
| House type (single family) | 835 (90.5%) | 388 (88.8%) | 171 (93.4%) | 275 (91.7%) |
| Psychosocial Characteristics | ||||
| Confidence in ability to do PA (self-efficacy) [1–5] | 3.53 (1.00) | 3.44 (1.01) | 3.46 (0.97) | 3.72 (0.98) |
| Decisional balance: (pros of PA - cons of PA) [−5 to 5] | 2.01 (0.78) | 1.95 (0.78) | 1.98 (0.76) | 2.12 (0.79) |
| Enjoyment of PA [1–5] | 4.27 (0.98) | 4.22 (1.00) | 4.16 (1.06) | 4.41 (0.87) |
| Home Environment | ||||
| Portable electronics ownership index [0–4] | 2.91 (0.95) | 2.85 (1.00) | 2.85 (0.95) | 3.04 (0.86) |
| Electronic items/things in bedroom index [0–6] | 2.61(1.70) | 2.50 (1.61) | 2.65 (1.70) | 2.75 (1.81) |
| Activity rules index [0–14] | 8.72 (3.10) | 8.73 (3.08) | 8.74 (3.08) | 8.68 (3.12) |
| Perceived neighborhood environment (NEWS) | ||||
| Parents NEWS | ||||
| Aesthetics [1–4] | 3.12 (0.64) | 3.11 (0.62) | 3.03 (0.64) | 3.16 (0.65) |
| Traffic Safety [1–4] | 2.58 (0.88) | 2.59 (0.57) | 2.53 (0.59) | 2.60 (0.60) |
| Pedestrian Safety [1–4] | 2.83 (0.65) | 2.84 (0.65) | 2.79 (0.68) | 2.85 (0.64) |
| Low Crime Risk [1–4] | 3.09 (0.88) | 3.07 (0.87) | 3.14 (0.89) | 3.11 (0.90) |
| Low Stranger Dangers [1–4] | 3.01 (0.73) | 2.96 (0.74) | 3.07 (0.72) | 3.04 (0.72) |
| Adolescent NEWS | ||||
| Traffic Safety [1–4] | 2.73 (0.61) | 2.74 (0.59) | 2.72 (0.67) | 2.73 (0.61) |
| Pedestrian Safety [1–4] | 3.10 (0.52) | 3.11 (0.51) | 3.06 (0.52) | 3.12 (0.53) |
| Low Crime Risk [1–4] | 3.21 (0.91) | 3.21 (0.91) | 3.17 (0.89) | 3.24 (0.92) |
| Low Stranger Dangers [1–4] | 3.38 (0.74) | 3.35 (0.74) | 3.39 (0.76) | 3.41 (0.73) |
| Built environment characteristics | ||||
| Number parks [parks/sq km] | 1.46 (1.65) | 1.50 (0.17) | 1.24 (1.49) | 1.52 (1.72) |
| Residential density [housing units/parcel] | 6.27 (8.68) | 6.99 (11.47) | 5.34 (4.84) | 5.70 (4.86) |
| Street Connectivity [intersections/sq km] | 73.37 (21.57) | 74.60 (20.90) | 70.01 (20.96) | 73.41 (22.64) |
| Retail floor area ratio [building:parcel sq ft] | 0.16 (0.19) | 0.16 (0.19) | 0.13 (0.16) | 0.17 (0.20) |
| Mixed use [0=single 1=mixed] | 0.18 (0.22) | 0.18 (0.22) | 0.13 (0.18) | 0.20 (0.23) |
| Walkability index | −0.03 (2.85) | 0.14 (3.03) | −0.67 (2.52) | 0.07 (2.71) |
Dog walking and overall physical activity and BMI
Table 2 shows that average daily MVPA was about 62 minutes for adolescents who lived in households without a dog as well as those who had a dog but did not walk it. Adolescents who walked their dog at least one day a week performed over 66 minutes of MVPA per day, which was significantly more than those in the other two groups (by 4–5 minutes), for an overall difference of 7–8% in MVPA time (P=0.044 and 0.025). Additionally, each day of dog walking was associated with 1.57 more minutes of total MVPA among dog owners (P=0.005). Dog ownership and dog walking were not associated with the adolescents’ BMI percentiles (Table 2).
Table 2.
Comparing MVPA minutes per day and BMI age adjusted percentiles of adolescents by those who live in households with dogs and walk them ≥1 day/week or don’t walk them or live in households without dogs (N=928) in the Seattle and Baltimore regions
| N | Mean (95% CI) MVPA minutes/daya,b | P value | Mean (95% CI) CDC Age adjusted BMI Percentilesa,b | P value | |
|---|---|---|---|---|---|
| A. Adolescents who don’t own a dog | 441 | 61.99 (59.02, 64.95) | Vs. B, p = .816 Vs. C, p = .025 |
66.53 (63.48, 69.57) | Vs. B, p = .516 Vs. C, p = .481 |
| B. Adolescents who own a dog but do not walk it | 184 | 61.45 (57.76, 65.41) | Vs. A, p = .816 Vs. C, p = .044 |
64.86 (59.15, 68.90) | Vs. A, p = .516 Vs. C, p = .890 |
| C. Adolescents who own a dog and walk it ≥ 1 day per weekc | 300 | 66.41 (62.96, 69.86)) | Vs. A, p = .025 Vs. B, p = .044 |
65.21 (61.64, 68.77) | Vs. A, p = .481 Vs. B, p = .890 |
Separate models were run in SPSS to compare the categorical variables with each other
All models controlled for adolescent age, gender, race/ethnicity, parent marital status, parent education and house type
The p-values are repeated contrasts from above but in different orders (i.e. A vs. C and C vs. A, B vs. C and C vs. B)
Correlates of dog walking at all versus none among adolescents with dogs
No demographic factors were associated with dog walking at all versus none, though all demographic variables were included as covariates in subsequent models (Table 3). In the subsequent 5 models, portable electronics ownership, parent’s perceived neighborhood aesthetics, mixed use, and the walkability index had positive associations with any dog walking at P<0.15, and residential density had a negative association with walking the dog at all at P<0.15, which met the criterion for inclusion in the final cross-level model.
Table 3.
Ecological analyses of correlates of dog walking among adolescents who live in households with dogs and walk the dog ≥1 day/week vs do not walk the dog in the Seattle and Baltimore regions. Ecological analyses of correlates of dog walking among adolescents who walk the dog ≥1 day a week
| Walk at all (≥1 day/week) vs. none (n=484) | Number of days/week walking dog (1–7) (n=300) | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | p | B | 95% CI | p | |
| Intercept (with centered variables): | 1.85 | 1.25, 2.72 | -- | 3.47 | 3.15, 3.99 | -- |
| Demographic Characteristics (Model 1) | ||||||
| Adolescent Age (years) | 1.15 | 0.79, 1.67 | .473 | −0.07 | −0.23, 0.09 | .390 |
| Adolescent Gender (Male) | 0.92 | 0.81, 1.06 | .239 | 0.02 | −0.41, 0.46 | .923 |
| Adolescent White Non-Hispanic | 1.01 | 0.66, 1.56 | .963 | −0.24 | −0.76, 0.28 | .362 |
| Parent married/living with a partner | 0.80 | 0.45, 1.40 | .423 | −0.46 | −1.08, 0.16 | .148 |
| Parent with college degree | 1.27 | 0.82, 1.97 | .518 | 0.69 | 0.16, 1.22 | .012 |
| House type (single family) | 0.77 | 0.36, 1.69 | .518 | −1.44 | −2.27, −0.60 | .001 |
| Psychosocial Characteristics (Model 2)a | ||||||
| Confidence in ability to do PA (self-efficacy) [1–5] | 1.19 | 0.93, 1.52 | .171 | 0.11 | −0.16, 0.38 | .430 |
| Decisional balance: (pros of PA - cons of PA) [−1–3] | 0.99 | 0.73, 1.34 | .939 | −0.08 | −0.41, 0.25 | .640 |
| Enjoyment of PA [1–5] | 0.17 | 0.92, 1.60 | .166 | 0.06 | −0.28, 0.40 | .737 |
| Home Environment (Model 3)a | ||||||
| Portable electronics ownership index [0–4] | 1.34 | 1.07, 1.67 | .011 | 0.04 | −0.23, 0.34 | .712 |
| Electronic items/things in bedroom index [0–6] | 1.01 | 0.90, 1.14 | .861 | 0.004 | −0.13, 0.14 | .957 |
| Activity rules index [0–14] | 0.99 | 0.92, 1.05 | .663 | −0.03 | −0.11, 0.04 | .411 |
| Perceived neighborhood environment (Models 4–5)a | ||||||
| Model 4: Parents NEWS | ||||||
| Aesthetics [1–4] | 1.31 | 0.95, 1.80 | .095 | 0.26 | −0.11, 0.63 | .169 |
| Traffic Safety [1–4] | 1.20 | 0.83, 1.73 | .326 | 0.08 | −0.32, 0.49 | .683 |
| Pedestrian Safety [1–4] | 1.05 | 0.77, 1.42 | .766 | −0.03 | −0.09, 0.47 | .888 |
| Low Crime Risk [1–4] | 0.95 | 0.75, 1.21 | .690 | 0.19 | −0.09, 0.47 | .186 |
| Low Stranger Dangers [1–4] | 0.91 | 0.67, 1.23 | .518 | −0.10 | −0.45, 0.24 | .559 |
| Model 5: Adolescent NEWS | ||||||
| Traffic Safety [1–4] | 0.91 | 0.65, 1.27 | .581 | −0.32 | −0.71, 0.07 | .107 |
| Pedestrian Safety [1–4] | 1.16 | 0.79, 1.70 | .447 | 0.13 | −0.31, 0.57 | .550 |
| Low Crime Risk [1–4] | 1.07 | 0.84, 1.37 | .588 | 0.005 | −0.27, 0.28 | .973 |
| Low Stranger Dangers [1–4] | 1.08 | 0.79, 1.49 | .629 | −0.11 | −0.47, 0.26 | .568 |
| Objective neighborhood environment (Models 6–7)a | ||||||
| Model 6: Built environment characteristics | ||||||
| Number parks [parks/sq km] | 1.03 | 0.90, 1.18 | .659 | −0.06 | −0.20, 0.09 | .426 |
| Residential density [housing units/parcel] | 0.96 | 0.91, 1.01 | .109 | 0.01 | −0.05, 0.07 | .788 |
| Street Connectivity [intersections/sq km] | 1.01 | 1.0, 1.02 | .221 | 0.01 | 0, 0.02 | .038 |
| Retail floor area ratio [building:parcel sq ft] | 2.13 | 0.54, 8.42 | .279 | −0.24 | −1.68, 1.20 | .745 |
| Mixed use [0=single 1=mixed] | 4.20 | 1.33, 13.23 | .014 | −1.29 | −2.41, −0.17 | .024 |
| Model 7: Walkability | ||||||
| Walkability index | 1.11 | 1.03, 1.20 | .009 | −0.01 | −0.10, 0.07 | .794 |
Models 2–7 controlled for adolescent age, gender, race/ethnicity, parent marital status, parent education and house type B denotes unstandardized regression coefficient.
In the final adjusted multi-level model, the significant main effect indicated that for every one-unit increase in neighborhood walkability, there was a 12% increase in the odds of walking the dog at all (P=0.006, Table 4). Residential density and mixed use were not included in this model because they were components of the walkability index. Five of 30 tested interactions were significant at P<0.1. Perceived aesthetics had significant interactions with adolescent age (P=0.007, Figure 1A) and adolescent race/ethnicity (P=0.090, Figure 1B). Portable electronics interacted significantly with adolescent age (P=0.062, Figure 1C), parental education (P<0.001, Figure 1D), and housing type (P=0.029, Figure 1E) in explaining the odds of walking the dog at all.
Table 4.
Final cross-level ecological model of correlates of dog walking (any vs. none) among adolescents who own a dog (N=484) in the Seattle and Baltimore regions. Significant (p<.10) interactions are shown in Figure 1: A–E.
| Walk at all (≥1 day/week) vs. none (n=484) | |||
|---|---|---|---|
| OR | 95% CI | p | |
| Intercept (with centered variables): | 1.89 | 1.25, 2.87 | -- |
| Final Ecological Model | |||
| Adolescent Age | 0.89 | 0.77, 1.03 | .112 |
| Adolescent Gender (Male) | 1.03 | 0.69, 1.54 | .892 |
| Adolescent White Non-Hispanic | 1.03 | 0.64, 1.63 | .916 |
| Parent married/living with a partner | 0.88 | 0.48, 1.60 | .668 |
| Parent with college degree | 1.25 | 0.76, 2.06 | .385 |
| House type (single family) | 0.72 | 0.31, 1.68 | .916 |
| Portable electronics ownership index | 1.24 | 0.84, 1.82 | .278 |
| Perceived aesthetics (Parents NEWS) | 1.20 | 0.83, 1.73 | .326 |
| Walkability Index | 1.12 | 1.03, 1.21 | .006 |
| Age*Perceived Aesthetics | 0.74 | 0.59, 0.92 | .007 |
| Race*Perceived Aesthetics | 1.88 | 0.91, 3.87 | .090 |
| Age*Portable electronics ownership | 0.85 | 0.72, 1.01 | .062 |
| Parent with college degree*Portable electronics ownership | 0.29 | 0.15, 0.55 | <.001 |
| House type*Portable electronics ownership | 2.64 | 1.10, 6.31 | .029 |
Figure 1.
Interactions with the odds of adolescents walking the dog at least one day a week (versus no days a week) in the Seattle and Baltimore regions
Correlates of dog walking frequency among adolescents who walked their dogs
Adolescent dog walkers reported walking their dogs 2.96 days/week (SD=1.97). Having a parent with a college degree was associated with more dog walking (P=0.012), and living in a single-family home was associated with less dog walking (P=0.001) (Table 3). Having a two-parent/guardian household, mixed use neighborhoods, and adolescent-perceived traffic safety were negatively associated, and objective street connectivity was positively associated with days of walking the dog at P<0.15, which met the criterion for inclusion in the final cross-level model (Table 5).
Table 5.
Final cross-level ecological model of correlates of dog walking frequency among adolescents who reported walking the dog at least 1 day/week (N=300) in the Seattle and Baltimore regions. Significant (p<.10) interactions are shown in Figure 2.
| Number of days/week walking dog (1–7) (n=300) |
||||
|---|---|---|---|---|
| β | B | 95% CI | p | |
| Intercept (with centered variables): | 2.94 | 3.48 | 3.07, 3.90 | -- |
| Final Ecological Model | ||||
| Adolescent Age (years) | −0.13 | −0.09 | −0.25, 0.07 | .273 |
| Adolescent Gender (Male) | 0.02 | 0.03 | −0.39, 0.46 | .879 |
| Adolescent White Non-Hispanic | −0.11 | −0.25 | −0.76, 0.25 | .325 |
| Parent married/living with a partner (marital status) | −0.15 | −0.42 | −1.03, 1.26 | .177 |
| Parent with college degree | 0.32 | 0.74 | 0.22, 1.26 | .006 |
| Housing type (single family) | −0.36 | −1.30 | −2.13, −0.48 | .002 |
| Traffic Safety (adolescent NEWS) | −0.22 | −0.36 | −0.71, 0 | .048 |
| Street Connectivity [intersections/sq km] | 0.32 | 0.02 | 0.01, 0.04 | .006 |
| Mixed Use [0=single 1=mixed] | −0.33 | −1.45 | −2.44, −0.46 | .004 |
| Marital status*street connectivity | −0.24 | −0.03 | −0.06, 0 | .083 |
B denotes unstandardized regression coefficient
β denotes standardized Beta
In the final adjusted multi-level model, living in a household where a parent had a college degree was associated with 0.74 more days/week of dog walking (P=0.006), and 10 more intersections per sq km (street connectivity) was associated with 2 more days/week of dog walking (P=0.006) (Table 5). Living in a single-family home was associated with 1.3 fewer days/week of dog walking (P=0.002), each unit increase in perceived traffic safety was associated with 0.36 fewer days/week of dog walking (P=0.048), and mixed use was associated with 1.45 fewer days/week of dog walking (compared to single use) (P=0.004). Only 1 of the 18 tested interactions was significant at P<0.1; marital status and street connectivity (P=0.083) (Fig 2).
Figure 2.
Interaction between parent marital status and neighborhood street connectivity with increased number of days of dog walking
DISCUSSION
Adolescents who walked their dogs at all obtained 4 to 5 minutes more MVPA per day than dog owners who did not walk their dogs. Those who walked their dogs 5 days a week had almost 8 minutes more of total MVPA/week than those not walking their dog. These findings are consistent with previous findings that dog walking contributes to more total physical activity among adults and adolescents.5–9,13,14 The relatively modest impact of dog walking on total MVPA minutes may be due to short dog walks or most dog walking being below the moderate-intensity threshold that would be captured in the accelerometer scores. Several demographic and environmental variables were related to dog walking. Youth living in more walkable neighborhoods and those with more portable electronics were more likely to walk their dogs. Supporting adolescents to walk their dogs regularly among households with dogs is a potential approach for interventions to increase total physical activity, but such interventions remain to be evaluated with adolescents.
There were no associations between dog walking and adolescent weight status. Previous dog walking studies found mixed associations between BMI and dog walking. Coleman et al. reported a significant inverse association among adults.13 Timperio et al. found the association between dog walking and adolescent weight varied by the type of dog owned, length of ownership and manner of interaction with the dog.7 Further studies of dog walking and BMI are needed to answer questions related to adolescent dog-walking patterns.
Correlates of any dog walking
The current study identified correlates at multiple levels of walking the dog at all. In the final model of any dog walking, neighborhood walkability emerged as the only significant main effect. Adolescents who lived in objectively more walkable neighborhoods were 12% more likely to walk their dogs in this study, consistent with a similar study of adults.13 A likely explanation is that people in single family homes with backyards do not perceive the need to walk their dogs for biological relief.
Ecological models predict cross-level interactions,11 and 5 such significant interactions were found in the final model for any dog walking by adolescents. A commonality among the interactions was that demographic factors moderated relations of perceived aesthetics and personal portable electronic ownership to dog walking. Aesthetics were positively related to dog walking in both younger and White non-Hispanic adolescents. It is possible that better aesthetics may lead parents to feel more comfortable letting their younger adolescents outside. In higher-minority neighborhoods, it is possible that aesthetics have a less influential role because other barriers to walking may be more salient (e.g. traffic or crime safety). Portable electronics were related to more dog walking among younger participants and those living in single-family and less-educated households. It is possible that portable electronics improve perceptions of safety or make dog walking more enjoyable, but this warrants further research.
Correlates of dog walking frequency
Several main effects were found for correlates of the frequency of adolescent dog walking. Adolescents who lived in single-family homes, had better perceived traffic safety, and lived in neighborhoods with objectively more mixed use walked their dogs less frequently. The present study examined housing type, with the hypothesis that dog owners in single-family homes would have less need to walk their dog because dogs could get both exercise and relieve themselves in private yards. This hypothesis was supported, as adolescents who lived in single-family homes walked their dogs about one day less per week than those who lived in multi-family homes or apartments. Cutt et al. found that an important reason dog owners overcame barriers to walk their dog was by recognizing the need to take their dog outside;10 having a yard may reduce this motivation. It was unexpected that better safety from traffic and more land use mix were associated with less dog walking, particularly because overall walkability was associated with the likelihood of walking the dog at all. These findings seem somewhat contradictory and could be due to confounding, but the authors could not identify a promising explanation.
Among adolescent dog walkers, those living in households with a college-educated parent and in neighborhoods with objectively higher street connectivity reported walking their dogs more frequently. Perhaps highly educated parents better understand the benefits of physical activity and are more likely to encourage their adolescents to walk dogs more. Higher street connectivity is an indicator of walkable neighborhoods, which was an important correlate of any dog walking. Because there was only one significant interaction related to frequency of dog walking, there is not strong evidence of moderation.
Strengths and limitations
Present findings support the utility of ecological models because significant correlates were found at multiple levels of influence. It is noteworthy that though no psychosocial correlates were significant in the final models, both reported and objectively measured environmental variables were significantly related to both dog walking outcomes. The cross-sectional design of the present study limits interpretation about causal pathways. The sample was intended to capture an even distribution of high and low walkable neighborhoods but not intended to be representative of the regions studied. Dog walking was self-reported by adolescents and potentially susceptible to social desirability bias. Though the short one-week recall of dog walking may reduce recall bias, this short time frame may limit representativeness of the dog walking measures. Frequency of dog walking was assessed, but it would be useful to obtain reports of usual duration of dog walking and if they walked with other people. Future studies could assess duration using improved measures such as the combination of GPS and accelerometry.
Conclusion
Because dog walkers had 7–8% more daily MVPA than non-dog-walkers, dog walking is a potential way to increase physical activity among adolescents with dogs. Present findings suggest intervention approaches that could be evaluated, particularly interventions that target multiple levels of the ecological model including walkability. Enhancing aesthetics of the neighborhood, such as more street trees, repainted buildings, and landscaping, might facilitate dog walking. Owning portable electronics could potentially increase dog walking because the adolescent can use these devices for entertainment or security. Less dog walking among those living in low-walkable neighborhoods and less frequent dog walking among those living in single-family homes seems to be an unanticipated consequence of current zoning laws. Present findings provide additional rationale for adoption of zoning laws that favor walkable neighborhood designs to limit barriers to walking, including dog walking.
Highlights.
Dog walking contributed 7–8% to daily MVPA among dog walking adolescents
Neighborhood walkability was a significant correlate related adolescents’ dog walking
Correlates from different levels of ecological models were found to be related to dog walking
Dog walking could be a promising intervention to increase PA among dog owners
Acknowledgments
This study was partially supported by NIH grant HL083454.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Contributor Information
Jessa K. Engelberg, Email: jkengelb@ucsd.edu, Family Medicine and Public Health, University of California, San Diego, 3900 5th Ave Suite 310, San Diego, CA 92103 USA, p: 619-260-2619; f: 619-260-1510.
Jordan A. Carlson, Email: jacarlson@cmh.edu, Center for Children’s Healthy Lifestyles & Nutrition, Children’s Mercy Hospital, 610 E. 22nd St., Kansas City, MO 64108, p: 816-234-9240; f: 816-302-9977.
Terry L. Conway, Email: tlconway@ucsd.edu, Family Medicine and Public Health, University of California, San Diego, 3900 5th Ave Suite 310, San Diego, CA 92103 USA, p: 619-260-5542; f: 619-260-1510.
Kelli L. Cain, Email: kcain@ucsd.edu, Family Medicine and Public Health, University of California, San Diego, 3900 5th Ave Suite 310, San Diego, CA 92103 USA, p: 619-260-5542; f: 619-260-1510.
Brian E. Saelens, Email: bsaelens@u.washington.edu, Department of Pediatrics, University of Washington, Seattle Children’s Research Institute, 2001 Eighth Avenue, Suite 400, Seattle, WA 98121 USA, p: 206-884-7800; f: 206-884-7801.
Karen Glanz, Email: kglanz@upenn.edu, Perelman School of Medicine and School of Nursing, University of Pennsylvania, 801 Blockley Hall, 423 Guardian Drive, Philadelphia, PA 19104, P: 215-898-0613; f: 215-573-5315.
Lawrence D. Frank, Email: ldfrank@urbandesign4health.com, School of Community and Regional Planning, University of British Columbia, Vancouver BC, #433-6333 Memorial Road, Vancouver, BC V6T 1Z2 CANADA.
James F. Sallis, Email: jsallis@ucsd.edu, Family Medicine and Public Health, University of California, San Diego, 3900 5th Ave Suite 310, San Diego, CA 92103 USA, p: 619-260-5535; f: 619-260-1510.
WORKS REFERENCED
- 1.Troiano RP, Berrigan D, Dodd KW, Mâsse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181. doi: 10.1249/mss.0b013e31815a51b3. [DOI] [PubMed] [Google Scholar]
- 2.Brener ND, Eaton DK, Kann LK, McManus TS, Lee SM, Scanlon KS, O’Toole TP. Behaviors Related to Physical Activity and Nutrition Among US High School Students. J Adolesc Health. 2013;53(4):539–546. doi: 10.1016/j.jadohealth.2013.05.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Carlson JA, Saelens BE, Kerr J, Schipperijn J, Conway TL, Frank LD, Chapman JE, Glanz K, Cain K, Sallis JF. Association between neighborhood walkability and GPS-measured walking, bicycling and vehicle time in adolescents. Health Place. 2015;32:1–7. doi: 10.1016/j.healthplace.2014.12.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.American Pet Products Manufacturers Association. [Accessed 26 March 2014];American Pet Products Manufacturers Association 2011–2012 National Pet Owners Survey. 2012 http://www.humanesociety.org/issues/pet_overpopulation/facts/pet_ownership_statistics.html#.UzWODa1dXw8.
- 5.Christian HE, Westgarth C, Bauman A, Richards EA, Rhodes RE, Evenson KR, Thorpe RJ., Jr Dog ownership and physical activity: a review of the evidence. J Phys Act Health. 2013;10(5) doi: 10.1123/jpah.10.5.750. [DOI] [PubMed] [Google Scholar]
- 6.Salmon J, Binh Chu AT, Veitch J. Dog ownership, dog walking, and children’s and parents’ physical activity. Res Q Exercise Sport. 2010;813:264–271. doi: 10.1080/02701367.2010.10599674. [DOI] [PubMed] [Google Scholar]
- 7.Timperio A, Salmon J, Chu B, Andrianopoulos N. Is dog ownership or dog walking associated with weight status in children and their parents? Health Promot J Aust. 2008;191:60. doi: 10.1071/he08060. [DOI] [PubMed] [Google Scholar]
- 8.Christian HE, Westgarth C, Bauman A, Richards E, Rhodes RE, Evenson KR, Thorpe RJ., Jr Dog Ownership and Physical Activity: A Review of the Evidence. J Phys Act Health. 2013 doi: 10.1123/jpah.10.5.750. [DOI] [PubMed] [Google Scholar]
- 9.Soares J, Epping J, Owens C, Brown D, Lankford TJ, Simoes E, Caspersen C. Odds of Getting Adequate Physical Activity by Dog Walking. Journal of physical activity & health. 2014 doi: 10.1123/jpah.2013-0229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cutt HE, Giles-Corti B, Wood LJ, Knuiman MW, Burke V. Barriers and motivators for owners walking their dog: results from qualitative research. Health Promot J Austr. 2008;19(2):118–124. doi: 10.1071/he08118. [DOI] [PubMed] [Google Scholar]
- 11.Sallis JF, Owen N. Ecological models of health behavior. In: Glanz K, Rimer BK, Viswanath K, editors. Health Behavior: Theory, Research, and Practice. 5. San Francisco: Jossey-Bass/Wiley; 2015. in press. [Google Scholar]
- 12.Toohey AM, McCormack GR, Doyle-Baker PK, Adams CL, Rock MJ. Dog-walking and sense of community in neighborhoods: Implications for promoting regular physical activity in adults 50 years and older. Health Place. 2013;22:75–81. doi: 10.1016/j.healthplace.2013.03.007. [DOI] [PubMed] [Google Scholar]
- 13.Coleman KJ, Rosenberg DE, Conway TL, Sallis JF, Saelens BE, Frank LD, Cain K. Physical activity, weight status, and neighborhood characteristics of dog walkers. Am J Prev Med. 2008;47(3):309–312. doi: 10.1016/j.ypmed.2008.05.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Christian H, Trapp G, Lauritsen C, Wright K, Giles-Corti B. Understanding the relationship between dog ownership and children’s physical activity and sedentary behaviour. Pediatr Obes. 2012 doi: 10.1111/j.2047-6310.2012.00113.x. [DOI] [PubMed] [Google Scholar]
- 15.Carlson JA, Sallis JF, Kerr J, Conway TL, Cain K, Frank LD, Saelens BE. Built environment characteristics and parent active transportation are associated with active travel to school in youth age 12–15. Br J Sports Med. 2014 doi: 10.1136/bjsports-2013-093101. bjsports-2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Frank LD, Sallis JF, Saelens BE, Cain K, Conway TL, Hess PM. The development of a walkability index: application to the Neighborhood Quality of Life Study. Br J Sports Med. 2010;44:924–933. doi: 10.1136/bjsm.2009.058701. [DOI] [PubMed] [Google Scholar]
- 17.Norman GJ, Sallis JF, Gaskins R. Comparability and reliability of paper- and 833 computer-based measures of psychosocial constructs for adolescent physical activity and sedentary behaviors. Res Q Exerc Sport. 2005;76:315–23. doi: 10.1080/02701367.2005.10599302. [DOI] [PubMed] [Google Scholar]
- 18.Rosenberg D, Sallis JF, Kerr J, Maher J, Norman GJ, Durant N, Harris SK, Saelens BE. Brief scales to assess physical activity and sedentary equipment in the home. Int J Behav Nutr Phy. 2010;7:10. doi: 10.1186/1479-5868-7-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rosenberg D, Ding D, Sallis JF, Kerr J, Norman GJ, Durant N, Saelens BE. Neighborhood Environment Walkability Scale for Youth (NEWS-Y): reliability and relationship with physical activity. Prev Med. 2009;49(2):213–218. doi: 10.1016/j.ypmed.2009.07.011. [DOI] [PubMed] [Google Scholar]
- 20.Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, Flegal KM, Guo SS, Wei R, Johnson CL. CDC growth charts: United States. Advance data. 2000;(314):1–27. [PubMed] [Google Scholar]
- 21.Freedson PS, Melanson E, Sirard J. Calibration of the Computer Science and Applications, Inc. (CSA) accelerometer. Med Sci Sports Exerc. 1998;29(suppl):45. doi: 10.1097/00005768-199805000-00021. [DOI] [PubMed] [Google Scholar]
- 22.Trost SG, Loprinzi PD, Moore R, Pfeiffer KA. Comparison of accelerometer cut points for predicting activity intensity in youth. Med Sci Sports Exerc. 2011;43(7):1360–1368. doi: 10.1249/MSS.0b013e318206476e. [DOI] [PubMed] [Google Scholar]
- 23.Richards EA, Ogata N, Ting J. Dogs, Physical Activity, and Walking (Dogs PAW) Acceptability and Feasibility of a Pilot Physical Activity Intervention. Health Promot Pract. 2014 doi: 10.1177/1524839914553300. 1524839914553300. [DOI] [PubMed] [Google Scholar]