Table 3 Methods used to account for neighbourhood self-selection in the articles considered in the systematic review

Alves, Silva (2013) [15]

Cross-sectional

Dietary behaviour and physical activity

Disadvantage

Model adjustment

Sociodemographic characteristics

3 variables: i) Age; ii) Education; iii) Marital status

No

Boarnet, Joh (2011) [16]

Cross-sectional

Physical activity

Residential density, Block size, Intersections, Commercial destinations

Restricted population (only consider a small study area, arguing “If residential location choice mostly determines the study area where persons live, but not where along the corridor residents live, then travel behaviour differences within the corridors will be due to direct effects of differences in the built environment and business concentration, and not residential preferences.”)

N/A

N/A

No

Boone-Heinonen, Gordon-Larsen (2011) [17]

Longitudinal

Physical activity

Disadvantage

Model adjustment and Fixed effects regression (Considered both mixed and fixed effects regression)

i) Sociodemographic characteristics

5 variables: i) Education; ii) Income; iii) Race; iv) Marital status; v) Children

Yes

Boone-Heinonen, Diez Roux (2011) [18]

Longitudinal

Dietary behaviour

Fast food chain restaurants, Supermarkets, Smaller grocery stores

Fixed effects regression

N/A

N/A

No

Brown, Pantin (2013) [19]

Cross-sectional

Physical activity

Walkability

Restricted population (only considered recent Cuban immigrants who overwhelmingly reported that they did not select their neighbourhood based on built environment characteristics)

N/A

N/A

No

Brown, Lombard (2014) [20]

Cross-sectional

Physical activity

Walkability, Distance to urban development boundary, Distance to central business district

Restricted population (only considered recent Cuban immigrants who overwhelmingly reported that they did not select their neighbourhood based on built environment characteristics)

N/A

N/A

No

Cerin, Frank (2011) [21]

Cross-sectional

Physical activity

Land use mix

Restricted population (limited to middle and high-income residents who could self-select for reasons other than affordability)

N/A

N/A

No

Christian, Knuiman (2013) [22]

Longitudinal

Physical activity

Neighbourhood type (hybrid, liveable, conventional)

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

21 items [not provided; referenced another article]. Factor analysis identified 5 factors: i) streets are pedestrian and cycling friendly; ii) access to services, jobs or place of study; iii) access to school; iv) close to parks and recreational facilities; v) safe, diverse and easy living community.

No

Foster, Hooper (2016) [23]

Longitudinal

Physical activity

Crime

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

1 item: i) Importance of safety from crime

No

Frank, Saelens (2007) [24]

Cross-sectional

Physical activity

Walkability, Household income

Model adjustment

i) Importance of characteristics for living in or moving to neighbourhood/new house, ii) Neighbourhood preference

10 items in reasons for moving:

i) Low crime, ii) Affordability, iii) Closeness to job, iv) Near shops and services, v) Near major roads and interstates, vi) Ease of walking, vii) Low transportation costs, viii) Near outdoor recreation, ix) Quality of schools, x) Near to public transit.

Principal components analysis identified 1 factor with low transportation costs, near to public transit and ease of walking having highest loads. The average score of these three items was split into quartiles and used as the self-selection variable. 7 trade-offs used to assess preferences: i) walkability vs. commercial-residential land use separation, ii) commute distance vs. residential density, iii) urban vitality vs. low-density and single-use neighbourhoods, iv) commute distance vs. living on quieter cul-de-sac street, v) availability of alternatives to the car vs. home size, vi) accommodation of automobile vs. accommodation of pedestrians and cyclists, vii) availability of alternatives to the car vs. neighbourhood privacy. Principal components analysis identified 1 factor. This was normalised and split into quartiles.

No

Frank, Kershaw (2014) [25]

Cross-sectional

Physical activity

Walkability

Model adjustment

Neighbourhood preference

7 trade-offs used to assess preferences in walkable vs. auto-orientated neighbourhoods: i) Closeness to shops and services; ii) Level of activity and mix of housing; iii) Home size and travel options; iv) Lot size and commute distance; v) Street design and travel options; vi) Public recreation and lot size; vii) Access to and size of food stores. Trade-offs were evaluated using 11-point scales for each of three questions: 1) “Your neighbourhood preference is … “, 2) “Indicate if your current neighbourhood is more like “A” or “B” … “, 3) “Regarding [the described attributes], the neighbourhood you’d hope to find would be [more like “A” or “B”] than your current neighbourhood”. Principal component analysis was used to extract a single neighbourhood preference component. This was split into quartiles.

No

Giles-Corti, Bull (2013) [26]

Longitudinal

Physical activity

Transport-related walking destinations (post offices, bus stops, delicatessens, supermarkets, train stations, shopping centres or CD or DVD stores), Recreation-related walking destinations (beach, park or sports field)

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

21 items [not provided; referenced another article]. Factor analysis identified 5 factors: i) streets are pedestrian and cycling friendly; ii) access to services, jobs or place of study; iii) access to school; iv) close to parks and recreational facilities; v) safe, diverse, easy living community. These five were included as separate categorical variables (Not important or not important at all/Somewhat important/Important). In addition, a self-selection scale (not important or somewhat important) used in previous studies was considered.

Yes

Hajna, Ross (2016) [27]

Longitudinal

Physical activity

Walkability

Model adjustment

Residential self-selection

Residential self-selection: 11 items from the Neighbourhood Quality of Life Study questionnaire (reference but no details provided)

Yes

Handy, Cao (2008) [28]

Cross-sectional

Physical activity

Institutional destinations (bank, church, library, post office), Maintenance destinations (grocery store, pharmacy), Eating out destinations (bakery, pizza, ice cream, takeaway), leisure destinations (health club, bookstore, bar, theatre, video rental)

Model adjustment

Importance of characteristics when looking to move to neighbourhood/new house

34 items [not all provided]: i) Easy access to a regional shopping mall, ii) Easy access to downtown, iii) Other amenities such as a pool or community centre available nearby, iv) Shopping areas within walking distance, v) Easy access to the freeway, vi) Good public transit service (bus or rail), vii) Good bicycle routes beyond neighbourhood, viii) Sidewalks throughout neighbourhood, ix) Parks and open spaces nearby, x) Quiet neighbourhood, xi) Low crime rate within neighbourhood, xii) Low level of car traffic on neighbourhood streets, xiii) Safe neighbourhood for walking, xiv) Safe neighbourhood for kids to play outdoors, xv) Good street lighting, xvi) Diverse neighbourhoods in terms of ethnicity, race, and age, xvii) Lots of people out and about within my neighbourhood, xviii) Lots of interaction among neighbours, xix) Economic level of neighbours similar to my level, xx) Attractive appearance of neighbourhood, xxi) High level of upkeep in neighbourhood, xxii) Variety in housing styles, xxiii) Big street trees, xxiv) Large backyards, xxv) Large front yards, xxvi) Lots of off-street parking (garages or driveways). Principal components analysis identified six factors: accessibility, physical activity options, attractiveness, outdoor spaciousness, safety, and socialising.

No

Jack and McCormack (2014) [29]

Cross-sectional

Physical activity

Walkability

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

19 items [not provided]. Principal components analysis identified four factors: access to places that support physical activity, access to local services, sense of community, ease of driving. These were transformed into z-scores.

No

Kaczynski and Mowen (2011) [30]

Cross-sectional

Physical activity

Park

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

1 item: i) Importance of closeness to open space

Yes

Knuiman, Christian (2014) [31]

Longitudinal

Physical activity

Street connectivity, Residential density, Land use mix, Service destinations (dry cleaners, post offices, pharmacies, video stores), Convenience destinations (delis, general stores, supermarkets, green grocers, seafood shops, gas stations, other food shops, shopping centres), Public open space destinations (parks, sports fields, beaches), Railway station

Fixed effects regression

N/A

N/A

Yes

Lee, Zegras (2013) [32]

Cross-sectional

Physical activity

Net density, Land use mix, Open space, Trail length, Intersections, Hilliness, Retail destinations, Transport destinations, Traffic volume, Traffic crashes

Model adjustment

Self-selection

Unclear. Used structural equation modelling to enable the inclusion of latent characteristics to control for self-selection.

Yes

MacDonald, Stokes (2010) [33]

Cross-sectional and Longitudinal

Physical activity

Light rail transit introduction [longitudinal], residential density [cross-sectional], park [cross-sectional], Food (grocery, convenience, restaurants) and alcohol destinations [cross-sectional]

Propensity score and Quasi-experiment

i) Sociodemographic variables, ii) Plans to use light rail transit

Sociodemographic variables/baseline characteristics included 7 items: i) gender, ii) race, iii) age, iv) employed, v) miles to work, vi) education level, vii) rent

No

McCormack, Shiell (2012) [34]

Cross-sectional

Physical activity

Street connectivity, Land-use mix, Residential density

Propensity score

i) Importance of characteristics for living in or moving to neighbourhood/new house; ii) Number of years in current neighbourhood; iii) Sociodemographic characteristics

19 items for characteristics: i) Affordability, ii) Proximity to parks, iii) Proximity to job/school, iv) Proximity to transit, v) Proximity to stores/services, vi) Ease of walking, vii) Sense of community, viii) Safety from crime, ix) Proximity to recreation facilities, x) Access to highways, xi) Attractive streets, xii) Proximity to family/friends, xiii) Views of scenery (e.g., mountains), xiv) Cleanliness of streets, xv) Proximity to downtown, xvi) Proximity to trails, xvii) Places to be physically active, xviii) Places to walk/cycle to, xix) Ease of driving. Sociodemographic characteristics: i) home ownership status, ii) gender, iii) age, iv) education, v) number of dependents < 18 years at home

No

McCormack, Friedenreich (2012) [35]

Cross-sectional

Physical activity

Walkability

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

21 items [not provided; referenced another article]. Factor analysis identified 5 factors: i) Pedestrian and cycling friendly streets; ii) Accessible services for daily living; iii) Accessible schools or places of study; iv) Accessible parks and recreation facilities; 5) Housing affordability and choice. Factors included as covariates in the models.

Yes

McCormack, McLaren (2017) [36]

Cross-sectional

Physical activity

Walkability, Business destinations, Bus stops, Parks, Recreational facilities, Sidewalk length, Residential density, Green space, Cycle paths

Propensity score

Importance of characteristics for living in or moving to neighbourhood/new house

13 items: i) Proximity to transit, ii) Proximity to recreational destinations, iii) Proximity to non-recreational destinations, iv) Proximity to work, v) Proximity to schools, vi) Proximity to downtown, vii) Access to highways and major roads, viii) Access to community associations, ix) Sense of community, x) Attractiveness, xi) Cleanliness of streets, xii) Housing type variety, xiii) quality of recreational facilities. Responses to each item were collapsed from “not at all”, “somewhat” and “very important” into “not important” and “important”. Propensity scores were created.

No

Nichani, Dirks (2016) [37]

Cross-sectional

Physical activity

Green space

Model adjustment

Neighbourhood preference

1 item: “Why do you live in this neighbourhood?: I like the local lifestyle.” No/Yes. Local lifestyle included access to community resources (e.g. green space, recreational facilities, public transport, shopping, education, healthcare, social and cultural facilities).

Yes

Norman, Carlson (2013) [38]

Cross-sectional

Physical activity

Walkability

Model adjustment

i) Importance of characteristics for living in or moving to neighbourhood/new house; ii) Neighbourhood preference

4 items in moving: i) Ease of walking; ii) Near public transit; iii) Near shops and services; iv) Near outdoor recreation. The average rating of these items, all measured on 5-point scales, was split at the median to categorise as low/high importance of walkability. 3 items in preference: i) residential density; ii) land use; iii) street connectivity. The average rating of these items, all measured on 11-point scales, was split at the median to categorise as low/high preference for a high-walkability neighbourhood.

No

Owen, Cerin (2007) [39]

Cross-sectional

Physical activity

Walkability

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

4 items: i) Closeness to job or school; ii) Closeness to public transportation; iii) Desire for nearby shops and services; iv) Ease of walking. The average rating of these items, all on 5-point scales, was used as neighbourhood self-selection measure.

Yes

Saelens, Sallis (2012) [40]

Longitudinal

Physical activity

Residential density, Land use mix, Intersection density, Retail destinations, Parks

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

3 items [chosen based on factor analysis of 11 residential selection items]: i) Closeness to public transportation; ii) Desire for nearby shops and services; iii) Ease of walking. The average rating of these items, all on 5-point scales, was used as residential selection variable.

No

Sallis, Saelens (2009) [41]

Cross-sectional

Physical activity

Walkability, Household income

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

3 items: i) Desire for nearby shops and services; ii) Ease of walking; iii) Closeness to recreational facilities. The average rating of these items, [scale not provided but reference to paper provided] was used as measure of walkability-related self-selection of neighbourhoods.

Yes

Van Dyck, Cardon (2011) [42]

Cross-sectional

Physical activity

Walkability

Model adjustment

Importance of characteristics for living in or moving to neighbourhood/new house

21 items: i) House price; ii) Importance of living in city centre; iii) Importance of living in a quiet neighbourhood; iv-vii) Social/emotional reasons (e.g. living close to family and friends); viii-xxi) Walkability related items (e.g. importance of closeness to shops, closeness to work/school, traffic safety, amount and quality of sidewalks/footpaths). All items were scored on a 5-point scale. A single variable was created but it is not clear how this was defined. Those scoring higher than the median were considered to have walkability as an important reason for neighbourhood selection.

Yes

Wells and Yang (2008) [43]

Longitudinal

Physical activity

Land use mix, Land use density, Street network pattern

Quasi-experiment (examine post-move - pre-move change in exposure on post-move outcome controlling for pre-move outcome)

N/A

N/A

No

West and Shores (2015) [44]

Longitudinal

Physical activity

Greenway

Quasi-experiment (pre-post design with control group)

N/A

N/A

No

Witten, Blakely (2012) [45]

Cross-sectional

Physical activity

Street connectivity, Dwelling density, Land use mix, Service and amenity destinations, Urbanicity

Model adjustment

i) Sociodemographic characteristics, and ii) Neighbourhood preference

8 sociodemographic variables: i) Age; ii) Sex; iii) Ethnicity; iv) Marital status; v) Household income; vi) Educational qualifications; vii) Occupation; viii) Household car access. 2 items: i) Prefer lower-density suburban neighbourhood suburban or urban environment located 10–15 min by car from common destinations or a higher-density urban neighbourhood with most destinations accessible on foot or by public transportation within 10–15 min; ii) Strength of preference for suburban or urban environment. Responses were combined as: strongly prefer walkable, moderately prefer walkable, neutral, moderately prefer less walkable, strongly prefer less walkable.

Yes