Table 1.

Key terms and operational definitions

Term	Operational Definition
Cohen’s Kappa coefficient	The agreement between primary and secondary retail food outlet data sources that takes into account the agreement occurring by chance
Concordance	The proportion of the retail food outlets both observed during primary data collection and listed by the secondary retail food outlet data sources among all the outlets ascertained via primary and secondary retail food outlet data sources
GPS-assisted	GPS or other forms of remote sensing technologies capture precise locational data (i.e., the latitude and longitude of a retail food outlet)
Ground-truthed	Primary data on retail food outlet type and location, gathered by trained observers not guided in the field by a list and/or map of retail food outlets identified through secondary data sources A systematic canvass of the targeted study area is conducted, with or without the use of GPS or other remote sensing technologies.
Intra- or inter-rater reliability	Evidence of intra-rater reliability included comparisons of retail food outlet data entered by the same rater. Evidence of inter-rater reliability included comparisons of raters’ decisions about whether to identify a retail food outlet as a convenience store or fast-food restaurant or both, as well as how to distinguish a small grocery store from a supermarket from a convenience store.
Omnidirectional Observations	Uses omnidirectional imagery (i.e., sources that simultaneously collect images in multiple directions from a single location producing a panoramic view such as Google Street View) to visually tour a targeted study area, not guided by a list of predetermined retail food outlets in the study area from primary or secondary data sources
On-Site Verification	Primary data on retail food outlet type and location, gathered by trained observers guided in the field by a list and/or map of food outlets identified through secondary data sources that could occur with or without a systematic canvass of the targeted study area and with or without the use of GPS or other remote sensing technologies
Percentage Agreement	The percentage of the primary retail food outlet data that matched the secondary retail food outlet data
Positive Predictive Value	The proportion of the retail food outlets listed by the secondary retail food outlet data sources that were observed during primary data collection
Primary Retail Food Data	Data collected through direct field observations by the team conducting the research to characterize the local retail food environment Primary data are considered the gold standard to characterize retail food environments given that secondary retail food outlet data sources have been found to under- and over-estimate food access, when compared to primary data.
Retail Food Outlet	Retail or commercial outlet in the business of selling food to the public; does not include household availability or institutional food service such as child care centers, schools, hospitals, correctional facilities, or municipal
Secondary Retail Food Data	Data collected by someone else; for example, government sources, such as local food inspection registries; commercial sources, such as InfoUSA and Dun and Bradstreet; online directories, such as Yellow Pages; and omnidirectional sources, such as Google Street View and Google Earth These sources have been shown to under- and over-count the number of retail food outlets in comparison to primary data.
Sensitivity	The ratio of the number of retail food outlets ascertained via primary data that matched retail food outlets ascertained via secondary data source(s), to the number of retail food outlets ascertained via primary data that matched retail outlets ascertained via secondary data source(s) plus the number of retail food outlets ascertained via primary data that did not match retail food outlets ascertained via secondary data source(s)
Specificity	The proportion of negatives (i.e., nonretail food outlets) that are correctly identified as not being retail food outlets
Systematic Canvass	Thorough and detailed primary data examination of a defined geographic setting using defined geographic parameters Evidence of a systematic canvass includes a detailed description or discussion of study maps marking areas to include and exclude during primary data collection and were not limited to the areas where secondary data sources indicated the presence of a retail food outlet. Ground-truthed studies by definition include systematic canvasses, while onsite verification studies could occur with or without a systematic canvass.
Targeted Observational Field Data	Primary data gathered by trained observers that targets a specific study area such as a study participant’s residential block or selected street block segments These observations do not systematically canvass beyond the targeted field areas. These observations may or may not use GPS or other remote sensing technologies. These studies do not include a list of predetermined resources in the study area to target the field observations, but the observational area is limited or guided by a participant’s residential address or based on study selection criteria such as high-walkability block segments in New York City.
Validity	This review focused on criterion-related validity, defined as the accuracy with which secondary data sources identified the type and location of retail food outlets, using primary data to represent the gold standard.
Virtual Verification	Uses omnidirectional imagery such as Google Street View or Google Earth to visually verify the existence of a retail food outlet identified through primary or secondary data sources