Hubbry Logo
Sampling frameSampling frameMain
Open search
Sampling frame
Community hub
Sampling frame
logo
7 pages, 0 posts
0 subscribers
Be the first to start a discussion here.
Be the first to start a discussion here.
Contribute something
Sampling frame
Sampling frame
Sampling frame
View on Wikipedia
from Wikipedia

In statistics, a sampling frame is the source material or device from which a sample is drawn.[1] It is a list of all those within a population who can be sampled, and may include individuals, households or institutions.[1]

Importance of the sampling frame is stressed by Jessen[2] and Salant and Dillman.[3]

In many practical situations the frame is a matter of choice to the survey planner, and sometimes a critical one. [...] Some very worthwhile investigations are not undertaken at all because of the lack of an apparent frame; others, because of faulty frames, have ended in a disaster or in cloud of doubt.

— Raymond James Jessen

A slightly more general concept of sampling frame includes area sampling frames, whose elements have a geographic nature. Area sampling frames can be useful for example in agricultural statistics when a suitable and updated agricultural census is not available. In environmental surveys, area sampling frames may be the only option.

Obtaining and organizing a sampling frame

[edit]

In the most straightforward cases, such as when dealing with a batch of material from a production run, or using a census, it is possible to identify and measure every single item in the population and to include any one of them in our sample; this is known as direct element sampling.[1] However, in many other cases this is not possible; either because it is cost-prohibitive (reaching every citizen of a country) or impossible (reaching all humans alive).

Having established the frame, there are a number of ways for organizing it to improve efficiency and effectiveness. It's at this stage that the researcher should decide whether the sample is in fact to be the whole population and would therefore be a census.

This list should also facilitate access to the selected sampling units. A frame may also provide additional 'auxiliary information' about its elements; when this information is related to variables or groups of interest, it may be used to improve survey design. While not necessary for simple sampling, a sampling frame used for more advanced sample techniques, such as stratified sampling, may contain additional information (such as demographic information).[1] For instance, an electoral register might include name and sex; this information can be used to ensure that a sample taken from that frame covers all demographic categories of interest. (Sometimes the auxiliary information is less explicit; for instance, a telephone number may provide some information about location.

Sampling frame qualities

[edit]

An ideal sampling frame will have the following qualities:[1]

  • all units have a logical, numerical identifier
  • all units can be found – their contact information, map location or other relevant information is present
  • the frame is organized in a logical, systematic fashion
  • the frame has additional information about the units that allow the use of more advanced sampling frames
  • every element of the population of interest is present in the frame
  • every element of the population is present only once in the frame
  • no elements from outside the population of interest are present in the frame
  • the data is 'up-to-date'[4]

Types of sampling frames

[edit]

The most straightforward type of frame is a list of elements of the population (preferably the entire population) with appropriate contact information. For example, in an opinion poll, possible sampling frames include an electoral register or a telephone directory. Other sampling frames can include employment records, school class lists, patient files in a hospital, organizations listed in a thematic database, and so on.[1][5] On a more practical levels, sampling frames have the form of computer files.[1]

Not all frames explicitly list population elements; some list only 'clusters'. For example, a street map can be used as a frame for a door-to-door survey; although it doesn't show individual houses, we can select streets from the map and then select houses on those streets. This offers some advantages: such a frame would include people who have recently moved and are not yet on the list frames discussed above, and it may be easier to use because it doesn't require storing data for every unit in the population, only for a smaller number of clusters.

Sampling frames problems

[edit]

The sampling frame must be representative of the population and this is a question outside the scope of statistical theory demanding the judgment of experts in the particular subject matter being studied. All the above frames omit some people who will vote at the next election and contain some people who will not; some frames will contain multiple records for the same person. People not in the frame have no prospect of being sampled.

Because a cluster-based frame contains less information about the population, it may place constraints on the sample design, possibly requiring the use of less efficient sampling methods and/or making it harder to interpret the resulting data.

Statistical theory tells us about the uncertainties in extrapolating from a sample to the frame. It should be expected that sample frames, will always contain some mistakes.[5] In some cases, this may lead to sampling bias.[1] Such bias should be minimized, and identified, although avoiding it completely in a real world is nearly impossible.[1] One should also not assume that sources which claim to be unbiased and representative are such.[1]

In defining the frame, practical, economic, ethical, and technical issues need to be addressed. The need to obtain timely results may prevent extending the frame far into the future. The difficulties can be extreme when the population and frame are disjoint. This is a particular problem in forecasting where inferences about the future are made from historical data. In fact, in 1703, when Jacob Bernoulli proposed to Gottfried Leibniz the possibility of using historical mortality data to predict the probability of early death of a living man, Gottfried Leibniz recognized the problem in replying:[6]

Nature has established patterns originating in the return of events but only for the most part. New illnesses flood the human race, so that no matter how many experiments you have done on corpses, you have not thereby imposed a limit on the nature of events so that in the future they could not vary.

— Gottfried Leibniz

Leslie Kish posited four basic problems of sampling frames:[7]

  1. Missing elements: Some members of the population are not included in the frame.
  2. Foreign elements: The non-members of the population are included in the frame.
  3. Duplicate entries: A member of the population is surveyed more than once.
  4. Groups or clusters: The frame lists clusters instead of individuals.

Problems like those listed can be identified by the use of pre-survey tests and pilot studies.

References

[edit]
Revisions and contributorsEdit on WikipediaRead on Wikipedia

Sampling frame

View on Grokipedia
from Grokipedia
A sampling frame is a structured list, database, or of all units within a defined from which a sample is drawn in statistical surveys and , ensuring that the sample can represent the target as accurately as possible. This frame acts as the foundational source material for probability sampling methods, where each unit has a known probability of selection, and it ideally includes every accessible element of the without omissions or duplicates. In practice, the sampling frame is often derived from existing records such as census data, voter registries, or administrative databases, though it may represent only a subset of the full if complete is infeasible. The importance of a well-constructed sampling frame lies in its in minimizing nonsampling errors, particularly coverage errors that can survey results by excluding or overrepresenting certain subgroups. For instance, undercoverage occurs when key segments of the —such as rural households or recent immigrants—are absent from the frame, leading to skewed estimates that fail to reflect true characteristics. Overcoverage, conversely, involves duplicate or ineligible units, which can inflate costs and complicate analysis without improving accuracy. High-quality frames are thus essential for producing reliable inferences, especially in large-scale applications like national health surveys or agricultural censuses, where frame construction involves integrating multiple data sources to achieve comprehensiveness. Sampling frames can vary in type depending on the study context, including list frames (e.g., telephone directories), area frames (e.g., geographic maps for household selection), or multi-stage frames that combine elements for complex populations. Challenges in frame development often arise from dynamic populations, outdated records, or resource constraints, prompting ongoing methodological advancements to enhance frame accuracy and adaptability in modern data environments.

Fundamentals

Definition and Scope

A sampling frame is defined as the concrete list, database, or representation of all units within the target from which a probability sample is drawn for survey or purposes. This frame serves as the practical foundation for selecting sample elements, ensuring that each unit has a known, nonzero probability of inclusion in the study. The scope of a sampling frame distinguishes it from the broader theoretical population, which encompasses all conceptual elements of interest, by focusing on the operational or accessible units that can actually be sampled. In practice, the frame may not perfectly align with the theoretical population due to exclusions, such as unlisted individuals, but it provides the workable roster for probability-based selection. Common examples include voter registries, which list eligible voters for political polling, and telephone directories, which enumerate households for consumer surveys. The concept of the sampling frame emerged in the 1940s through the pioneering efforts of statisticians at the U.S. Bureau, particularly Morris Hansen, who advanced probability sampling techniques during wartime and postwar survey designs. Hansen, along with collaborators Hurwitz and Madow, formalized the term in their influential 1953 treatise on sample survey methods, establishing it as a core element of modern survey theory.

Relation to Population and Sample

The target population encompasses all conceptual units eligible for inclusion in a study, defined by specific characteristics relevant to the objectives, whereas the sampling frame constitutes a practical, operational list or database of these units (or proxies for them) from which the actual sample is selected. This frame often represents only a of the target population due to logistical constraints, potentially introducing frame error—the systematic discrepancy between the two, which can manifest as undercoverage (omission of some target units from the frame) or overcoverage (inclusion of ineligible units). Such errors compromise the representativeness of the sample and the validity of inferences drawn about the broader population. In the sampling , the frame serves as the foundational mechanism for probability-based selection, ensuring that every unit within it has a known, non-zero probability of inclusion, which allows for unbiased and generalizability to the target population. This known inclusion probability is calculated based on the frame's structure and the sampling , facilitating the use of statistical to quantify sampling variability and construct intervals. Absent a well-defined frame, selection relies on non-probability methods, where inclusion chances are unknown or unequal, limiting the to make probabilistic inferences and increasing reliance on subjective judgments. For example, in the National Hospital Discharge Survey (through ), the target consists of all inpatient discharges from non-federal short-stay hospitals , while the sampling frame is the master facility inventory of such hospitals; a probability sample of approximately 500 hospitals is selected, from which a sample of discharge (around 300,000 annually) is drawn for to represent national trends.

Construction

Sources for Obtaining Frames

Sampling frames are typically constructed from primary sources that provide direct, authoritative listings of population elements. Administrative records, such as tax rolls maintained by revenue agencies, serve as a key by offering comprehensive lists of households or individuals based on fiscal obligations. Similarly, school enrollment records from education departments function as frames for studies targeting students or families, capturing current demographic details like age and location. Registries, including business licenses issued by government agencies, enable sampling of commercial entities by providing up-to-date operational data. In areas lacking robust records, field enumerations involve on-site mapping and listing of households, particularly in remote or rural regions, to create bespoke frames through direct observation and verification. Secondary sources supplement primary data by offering accessible, pre-compiled datasets for frame development. Purchased databases, such as commercial mailing lists from vendors like or , provide enhanced frames with appended variables including income estimates and contact information, often derived from aggregated administrative and consumer records. Public datasets, exemplified by the 2020 U.S. Decennial , deliver broad population frames through geocoded address files and demographic summaries, enabling researchers to sample from verified housing units nationwide. Ensuring the currency of sampling frames is essential to minimize discrepancies between the frame and the target population, particularly in dynamic contexts like demographic studies where births, deaths, and migration alter compositions. Population registers, updated routinely with vital events, help maintain frame accuracy by incorporating these changes, as seen in systems used by national statistical offices. Failure to update frames can introduce undercoverage bias; for instance, outdated voter registration lists in election polling may exclude recent movers or deceased individuals, skewing results toward stable urban demographics.

Methods for Organizing Frames

Organizing a sampling frame begins with structuring the data to facilitate efficient access and selection during the sampling process. One fundamental approach involves assigning unique to each unit in the frame, such as numerical IDs or codes that ensure distinctiveness and prevent overlap, which is essential for accurate unit tracking in surveys. For instance, in agricultural master sampling frames, units like holdings are given unique codes combining administrative levels to maintain clarity across regions. Another key structuring method is stratification, where the frame is divided into subgroups based on relevant variables like geographic location or demographic characteristics, allowing for targeted sampling within clusters to improve representativeness. plays a critical role in this organization, converting frames into electronic formats compatible with database systems such as SQL, which enable querying, sorting, and integration for large-scale operations. Maintenance of the sampling frame requires ongoing processes to preserve its accuracy and over time. Periodic updates are typically achieved through linkage to external sources, such as administrative records or vital statistics registries, which allow for the addition, deletion, or modification of units to reflect real-world changes like births, , or migrations. For example, the U.S. Census Bureau's Master Address File is continuously updated using Postal Service files and federal agency to incorporate new housing units and group quarters. Handling duplicates is a vital aspect of , often employing deduplication algorithms that compare fields like names, addresses, and to flag and resolve overlaps systematically. In the World Trade Center Health Registry, such an algorithm reduced the frame by over 20,000 records by matching locator and demographic , minimizing overcoverage. Various tools and best practices support these organization and maintenance efforts, particularly in specialized contexts. Geographic Information Systems (GIS) are widely used for spatial frames, enabling the layering of points, lines, and polygons to structure area-based data with precise georeferencing for environmental or land-use surveys. Software like SAS facilitates frame management through procedures such as PROC SURVEYSELECT, which treats input datasets as frames for selecting samples while handling stratification and allocation. A practical example is organizing frames for agricultural surveys, where holdings are stratified by farm size—such as fully enumerating large holdings while sampling smaller ones—to optimize and ensure coverage of diverse production scales.

Characteristics

Essential Qualities

A sampling frame's in unbiased probability sampling hinges on three core attributes: completeness, accuracy, and non-duplication. These qualities that the frame serves as a reliable representation of the target population, minimizing coverage errors that could distort survey estimates. Completeness requires that the frame encompasses all units of the target population, providing each with a non-zero probability of selection. In practice, under-coverage—such as omitting nomadic households or new housing units in an outdated frame—can lead to biased estimates by systematically excluding certain subgroups. Accuracy refers to the correct and up-to-date representation of units in the frame, free from errors in identification or attributes like addresses or eligibility status. Inaccurate frames, such as those with misspelled names or invalid contact information, can result in failed sample selections or misclassification of units, thereby compromising the precision of survey results. Non-duplication ensures that each unit appears exactly once in the frame, preventing over-representation. The duplication rate highlights this issue; even low rates can cause over-sampling of certain units, inflating their influence on estimates and introducing positive . To mitigate this, often employ unique or post-processing to eliminate repeats, as seen in multi-list frames where overlaps must be resolved through weighting adjustments. A high-quality sampling frame for urban employment surveys, such as the U.S. Bureau of Labor Statistics' Quarterly Census of Employment and Wages, covers more than 95% of U.S. jobs, exemplifying strong completeness while maintaining accuracy and non-duplication through rigorous list maintenance.

Criteria for Evaluation

Evaluating the quality of a sampling frame involves standardized techniques to ensure it accurately represents the target population and supports reliable sampling. One primary evaluation technique is auditing subsets of the frame through random checks against external sources, such as census data or administrative records, to verify completeness and accuracy. This process identifies discrepancies like duplicates or omissions before full implementation. Another key technique is computing frame coverage error, which quantifies undercoverage or overcoverage relative to the target population. Important metrics for assessment include the uniformity of inclusion probabilities, where each unit in the frame should have a known and ideally equal probability of selection to minimize in probability-based sampling. Additionally, the cost-effectiveness compares the expenses of building and maintaining the frame against the improvements in sampling efficiency, such as reduced variance or higher response rates, to determine practical viability. Diagnostic tools, such as total survey error frameworks developed by Leslie Kish in the 1960s, provide frameworks for decomposing errors into components like coverage, nonresponse, and measurement biases, enabling targeted improvements. For instance, in evaluating a sampling frame, analysts may assess nonresponse by comparing respondent characteristics, such as age or from zip-code-level data, against known population benchmarks to detect systematic exclusions of certain groups. These criteria build on essential qualities like completeness and accuracy by offering quantifiable ways to measure and enhance frame performance prior to sampling.

Classifications

List-Based Frames

List-based sampling frames consist of explicit, enumerated lists of all units within a finite target population, providing a complete roster from which samples can be drawn. These frames typically include identifying information such as names, contact details, or identifiers for each unit, making them ideal for populations that can be comprehensively cataloged. Examples include maintained by businesses for , rosters at educational institutions for surveys on academic performance, and registries in healthcare settings for clinical studies. Such frames are particularly suitable for finite populations where every member can be identified and listed without omission or duplication. A primary advantage of list-based frames is their compatibility with simple random sampling, where each unit has an equal probability of selection, ensuring unbiased representation when the list is complete and up-to-date. This approach facilitates the use of generators or tables to select samples efficiently. Additionally, these frames enable straightforward stratification by allowing researchers to divide the list into subgroups based on characteristics like age, , or , thereby improving sample precision and representativeness across diverse segments. In applications, list-based frames are commonly employed in processes, such as selecting batches of manufactured products from a production roster to inspect for defects. They are also integral to clinical trials, where patient lists from hospital databases allow for randomized assignment to treatment groups while ensuring ethical and representative selection. A notable historical example is the 1936 Literary Digest poll, which used lists compiled from telephone directories, automobile registrations, and voter rolls to survey 10 million potential respondents; however, the frame's toward wealthier individuals led to a grossly inaccurate prediction of the U.S. outcome.

Area-Based and Multi-Frame Types

Area-based sampling frames divide geographic space into discrete segments to represent populations that are difficult to enumerate explicitly, such as households or agricultural units spread across large areas. These frames typically use maps, , or geographic information systems (GIS) to delineate primary sampling units (PSUs), such as city blocks, enumeration districts, or land parcels, from which secondary units like dwellings or farms are selected with known probabilities. This approach contrasts with list-based frames by relying on spatial coverage rather than pre-existing rosters, enabling comprehensive sampling in dynamic environments. In agricultural contexts, area-based frames have been pivotal, as exemplified by the U.S. Department of Agriculture's (USDA) National Agricultural Statistics Service (NASS) crop frames, which segment land into tracts typically ranging from 0.1 to 1 (approximately 64 to 640 acres), depending on the and , to estimate crop acreage and yields nationwide. These frames incorporate data and field enumerations to classify , ensuring unbiased estimates for non-point-frame populations like small farms or remote fields. National area probability sampling, a foundational application of this method, emerged in the 1940s through U.S. initiatives, including the Bureau's innovations in probability-based area selection for population and economic surveys. Multi-frame sampling types integrate multiple overlapping to enhance coverage for populations elusive to single-frame approaches, such as combining list-based sources like directories with area . A prominent example is dual-frame surveys, which merge and to address shifts in communication patterns, with samples drawn independently from each . Overlaps between are adjusted using inclusion probabilities, where the probability of selection for units in multiple is accounted for in estimation procedures to avoid double-counting and ensure unbiased totals. In contemporary applications, multi-frame designs extend to web-based surveys by combining email lists, platforms, and other digital sources to capture diverse online populations, improving representativeness in hard-to-reach groups like young adults or remote workers. These hybrid frames leverage algorithmic selection and probability adjustments to integrate disparate data sources, as seen in recent statistical agency implementations for broad societal surveys.

Challenges

Common Errors and Biases

One of the most prevalent errors in sampling frames is undercoverage, which occurs when certain members of the target population are systematically excluded from the frame, leading to non-representative samples and biased estimates. For instance, address-based frames often fail to capture transient populations such as frequent movers, resulting in underestimation of prevalence rates for issues like or health disparities among marginalized groups. This exclusion particularly distorts subpopulation analyses, as underrepresented groups contribute disproportionately to overall bias in survey inferences. Overcoverage represents another common issue, where the sampling frame includes units that do not belong to the target population, such as ineligible or outdated entries, which inflates the sample size unnecessarily and reduces . An example is frames that include closed facilities or converted group quarters, leading to wasted resources on non-viable contacts and potential dilution of valid responses. While overcoverage may not always introduce severe if ineligible units are screened out, it complicates fieldwork and can indirectly affect representativeness by straining survey operations. Beyond coverage issues, sampling frames can suffer from clustering, where units within the frame are not independent but grouped in ways that violate assumptions of simple random sampling, thereby increasing variance and introducing dependence bias. Additionally, temporal misalignment arises when the frame becomes outdated relative to the sampling period, capturing a population state that no longer aligns with current conditions and skewing results toward historical rather than contemporary realities. A historical illustration is the 1948 U.S. polls, where quota-based sampling led to biased selection by overrepresenting urban Republicans, contributing to erroneous predictions of a Dewey victory.

Strategies for Mitigation

To address undercoverage in sampling frames, post-stratification weighting calibrates sample estimates to known population benchmarks, effectively adjusting for discrepancies caused by incomplete frames. This method models inclusion probabilities for units in the frame and iteratively minimizes an objective function to align weighted sample totals with external controls, from omissions without requiring frame reconstruction. Frame augmentation complements this by incorporating supplemental lists from alternative sources, such as administrative records or field enumerations, to expand coverage of underrepresented subpopulations. For instance, in address-based sampling, vendors append data from the USPS No-Stat File or commercial databases to capture unlisted residences, improving rural coverage by 4% while minimizing overcoverage through targeted matching. Design strategies further mitigate frame limitations by altering the sampling process itself. Multi-stage sampling divides the population into hierarchical clusters, such as geographic areas, allowing random selection of clusters before subsampling individuals within them, which eliminates the need for a comprehensive frame of the entire population. This approach is particularly useful for large-scale studies where frame construction is infeasible, as it reduces logistical demands while maintaining probabilistic representation. Adaptive or responsive designs enable dynamic updates to the frame during data collection, using propensity models based on paradata (e.g., contact history) to prioritize high-response units or switch modes, thereby addressing emerging undercoverage in real-time without full redesign. For example, in multi-phase surveys, initial phases cap efforts on low-propensity cases, then reallocate resources to supplement the frame via incentives or mode shifts, controlling costs while boosting response rates. Recent advancements as of 2025 address evolving challenges, such as using multiple overlapping frames and mixed-mode designs to improve coverage in digital and mobile populations, while navigating privacy regulations like GDPR that restrict data integration for frame construction. Best practices emphasize proactive validation to detect frame errors early. Pilot testing involves administering the survey to a small, nonrandom convenience sample (typically 50-100 cases) that mirrors the target population, revealing issues like accessibility gaps or selection biases in the frame before full implementation. This process simulates production conditions, including interviewer training and mode of administration, to identify and correct frame deficiencies, thereby minimizing nonsampling errors. Post-stratification weighting has been applied in U.S. election surveys to adjust for coverage discrepancies by aligning samples with census benchmarks on demographics, helping to reduce bias in estimates.

References

  1. https://www.nlm.nih.gov/oet/ed/stats/02-100.html
  2. https://www.umsl.edu/~lindquists/sample.html
  3. https://pmc.ncbi.nlm.nih.gov/articles/PMC6662929/
  4. https://unstats.un.org/unsd/demographic/meetings/egm/sampling_1203/docs/no_3.pdf
  5. https://www.westga.edu/academics/research/vrc/assets/docs/SamplingFrames-Errors_EssayChumney_09282012.pdf
  6. https://www.nass.usda.gov/Education_and_Outreach/Understanding_Statistics/Statistical_Aspects_of_Surveys/whats_in_a_frame.pdf
  7. https://pmc.ncbi.nlm.nih.gov/articles/PMC4817645/
  8. http://people.umass.edu/~biep540w/pdf/sampling.pdf
  9. https://www.surveypractice.org/article/2827-scientific-surveys-based-on-incomplete-sampling-frames-and-high-rates-of-nonresponse
  10. https://www.pewresearch.org/methods/2018/10/09/comparing-survey-sampling-strategies-random-digit-dial-vs-voter-files/
  11. https://marketing-bulletin.massey.ac.nz/V3/MB_V3_A5_Esslemont.pdf
  12. https://www150.statcan.gc.ca/n1/pub/12-001-x/2017002/article/54888/02-eng.htm
  13. https://ccsg.isr.umich.edu/glossary/coverage-error/
  14. https://pmc.ncbi.nlm.nih.gov/articles/PMC5325924/
  15. https://www.cdc.gov/nchs/nhds/nhds_sample_design.htm
  16. https://nap.nationalacademies.org/read/27150/chapter/11
  17. https://www.samhsa.gov/data/sites/default/files/standards_stat_surveys.pdf
  18. https://aapor.org/wp-content/uploads/2022/11/AAPOR_Report_1_7_16_CLEAN-COPY-FINAL-2.pdf
  19. https://openknowledge.fao.org/server/api/core/bitstreams/00e74c7c-5e5d-434e-87fd-43d300069617/content
  20. https://unstats.un.org/capacity-development/handbook/chapters/Chapter12_Handbook_2025A_202505.pdf
  21. https://nap.nationalacademies.org/read/13387/chapter/6
  22. https://www.rti.org/sites/default/files/resources/murphy_jsm2004_paper.pdf
  23. https://www.fao.org/fileadmin/templates/ess/documents/world_census_of_agriculture/chapter10_r7.pdf
  24. https://unstats.un.org/unsd/demographic/sources/surveys/handbook23june05.pdf
  25. https://www.nass.usda.gov/Education_and_Outreach/Reports%2C_Presentations_and_Conferences/allreports/Statistical_Aspects_of_a_Census.pdf
  26. http://www.asasrms.org/Proceedings/papers/1988_056.pdf
  27. https://www.bls.gov/cew/overview.htm
  28. https://www150.statcan.gc.ca/n1/pub/62f0026m/2010005/section3-eng.htm
  29. https://ccsg.isr.umich.edu/chapters/sample-design/
  30. https://academic.oup.com/poq/article/74/5/849/1817502
  31. https://academic.oup.com/aje/article/178/8/1337/83207
  32. https://statisticsbyjim.com/basics/sampling-frame/
  33. https://uca.edu/psychology/files/2013/08/Ch7-Sampling-Techniques.pdf
  34. https://www.omniconvert.com/blog/statistical-sampling/
  35. https://gwern.net/doc/statistics/bias/2017-lohr.pdf
  36. https://www.worldbank.org/en/programs/lsms/sampling
  37. https://projecteuclid.org/journals/statistical-science/volume-7/issue-3/US-Government-Contributions-to-Probability-Sampling-and-Statistical-Analysis/10.1214/ss/1177011230.pdf
  38. https://www.jstor.org/stable/2279251
  39. https://www150.statcan.gc.ca/n1/en/pub/12-001-x/2011002/article/11608-eng.pdf
  40. https://research.google/pubs/combining-landline-and-mobile-phone-samples-a-dual-frame-approach/
  41. https://www.tandfonline.com/doi/abs/10.1198/016214506000000195
  42. https://isi-iass.org/home/wp-content/uploads/Survey_Statistician_2025_January_N91_03.pdf
  43. https://www150.statcan.gc.ca/n1/en/pub/12-001-x/2021002/article/00008-eng.pdf
  44. https://study.sagepub.com/sites/default/files/Fricker.pdf
  45. https://pure.rug.nl/ws/files/93751314/smy016_1_.pdf
  46. https://link.springer.com/article/10.1007/s11160-025-09921-2
  47. https://link.springer.com/article/10.1007/s42379-022-00102-3
  48. https://arxiv.org/pdf/2107.14166
  49. https://www.nass.usda.gov/Education_and_Outreach/Reports%2C_Presentations_and_Conferences/reports/cal_paper_rev3.pdf
  50. https://nces.ed.gov/fcsm/pdf/J1_Tourangeau_2015FCSM.pdf
  51. https://journals.sagepub.com/doi/abs/10.1177/0282423X251324098
  52. https://methods.sagepub.com/ency/edvol/encyclopedia-of-survey-research-methods/chpt/pilot-test
  53. https://aapor.org/wp-content/uploads/2022/11/Report-of-the-Task-Force-on-Transitions-from-Telephone-Surveys-FULL-REPORT-FINAL.pdf
Add your contribution
Related Hubs
Contribute something
User Avatar
No comments yet.