What Is Time-Location Sampling?
Time-location sampling (TLS) is a probability-based method for reaching populations that don't have a fixed residential address or aren't reachable through standard household sampling frames. Instead of sampling people from where they live, TLS samples them from where they go. Researchers first map out all the venues and time periods where the target population congregates, bars, clinics, parks, community centers, transit hubs, events, then randomly select a subset of venue-time combinations and intercept everyone (or a random sample of those) who shows up during the selected windows. Because the sampling units are venue-day-time blocks rather than individuals, TLS can produce valid probability samples of populations that would be invisible to traditional methods: homeless individuals, migrant workers, nightlife-connected communities, or anyone who moves through public spaces in predictable patterns.
Why Time-Location Sampling Matters
Populations defined by behavior rather than geography can't be captured by address-based or household sampling frames. TLS provides a structured, probability-based alternative that yields calculable inclusion probabilities, something convenience sampling and snowball methods can't offer. Public health agencies adopted TLS specifically because it produces defensible prevalence estimates for mobile and hidden populations, making it essential for HIV surveillance, harm reduction research, and migrant health studies.
How Time-Location Sampling Works
TLS requires extensive formative research before a single interview happens. The method has four distinct phases.
Phase 1: Venue Universe Construction
Before sampling, you need a comprehensive list of every venue-time block where your target population gathers. This requires ethnographic fieldwork, key informant interviews, community mapping, and direct observation. For a study of men who have sex with men, the venue universe might include specific bars, clubs, bathhouses, parks, and community organizations, each broken into time blocks (e.g., Saturday 10pm-2am vs. Tuesday 6pm-9pm).
The completeness of this list directly determines the coverage of your sample. If major gathering points are missing from the venue universe, everyone who frequents those locations is excluded, introducing coverage bias.
Phase 2: Random Selection of Venue-Time Events
From the complete venue universe, randomly select a manageable number of venue-day-time blocks for each sampling period (typically monthly). Selection can be done with equal probability or with probability proportional to estimated attendance, busier venues get selected more often, which is more efficient.
Each selected venue-time block becomes a sampling event where a field team will intercept participants.
Phase 3: Field Enumeration and Recruitment
At each selected event, the field team counts everyone who enters or is present during the sampling window (enumeration) and then approaches individuals for eligibility screening and participation. You can attempt to recruit everyone present (take-all approach) or systematically sample every nth person. The enumeration count is critical, it provides the denominator needed to calculate inclusion probabilities and response rates.
Consent, screening, and data collection happen on-site. Interviews need to be short enough that people will participate in a venue setting, typically 15-30 minutes.
Phase 4: Estimation and Weighting
Each participant's inclusion probability depends on how many venue-time blocks they attend (self-reported attendance patterns), the selection probability of each venue-time block, and the within-event sampling fraction. Participants who frequent many venues have higher inclusion probabilities and receive lower weights. The Horvitz-Thompson estimator produces unbiased population estimates when these probabilities are calculated correctly.
The math gets complicated when participants attend multiple selected events during a study period. Multiplicity adjustments prevent double-counting by weighting down people who had multiple opportunities to be sampled.
When to Use Time-Location Sampling
- Public health surveillance for mobile or hidden populations: people experiencing homelessness, substance users, sex workers, migrant communities
- Studies where the target population congregates at identifiable venues but doesn't appear on any list-based frame
- Populations defined by behavior (nightlife participation, event attendance, service utilization) rather than demographics
- Urban research where foot traffic patterns provide natural sampling opportunities at transit hubs, markets, or public spaces
- When you need probability-based estimates but household sampling would miss the target group entirely
Common Mistakes to Avoid
- Incomplete venue mapping that misses major gathering points, creating coverage gaps. Invest in thorough formative research, the venue universe is the foundation of the entire design.
- Skipping enumeration at sampling events. Without a count of everyone present, you can't calculate response rates or inclusion probabilities, and the sample reduces to a convenience intercept.
- Ignoring multiplicity adjustments for people who attend multiple venues. Failing to account for higher inclusion probabilities among frequent venue-goers biases your estimates toward the most socially active members of the population.
How Quali-Fi Supports Time-Location Sampling
Quali-Fi's mobile survey tools support on-site data collection with offline capability, GPS verification, and timestamp logging that maps directly to your venue-time sampling matrix. The platform's real-time field dashboards track completion rates by event, making it easy to monitor which venue-time blocks have been covered and flag events that need follow-up.
Frequently Asked Questions
How is TLS different from venue-based sampling?
TLS is a specific probability-based framework within the broader category of venue-based sampling. Generic venue-based sampling can be convenience-based (intercepting whoever is available at a location). TLS adds the formal structure: random selection of venue-time blocks, systematic enumeration, and probability-weighted estimation. The structure is what makes TLS a probability method.
How many venue-time events do I need to sample?
Most TLS studies sample 15-30 venue-time events per month over 3-12 months, targeting 200-500 total participants. The exact number depends on population density at venues, your precision requirements, and the diversity of the venue universe. More events from different venues reduces clustering effects and improves coverage.
Can TLS work in rural settings?
It's harder but possible. TLS works best when the population concentrates at identifiable locations. In rural areas, gathering points may be fewer and more dispersed, weekly markets, religious services, water points. If you can map them, you can sample them, but the logistics are more challenging and per-interview costs increase.
Related Topics
- Venue-Based Sampling
- Area Probability Sampling
- Respondent-Driven Sampling
- Network Sampling
- Self-Selection Bias in Sampling
Reach mobile populations where they gather. Start a free trial with Quali-Fi and use GPS-verified mobile surveys, real-time field dashboards, and offline data capture for venue-based research.