Integration of Nonprobability and Probability Samples via Survey Weights

Nandram, Balgobin; Choi, Jin Woo; Liu, Yang

doi:10.5539/ijsp.v10n6p5

Cited by 6 publications

(6 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Alternative matching measures such as Mahalanobis distance have also been considered. 40 This is very closely related to the approach in Stuart et al 2011.…”

Section: Brief Review Of Non-probability Inferencementioning

confidence: 79%

Improving transportability of randomized controlled trial inference using robust prediction methods

Elliott,

Carroll,

Grieve

et al. 2023

Stat Methods Med Res

View full text Add to dashboard Cite

Randomized trials have been the gold standard for assessing causal effects since their introduction by Fisher in the 1920s, since they can eliminate both observed and unobserved confounding. Estimates of causal effects at the population level from randomized controlled trials can still be biased if there are both effect modification and systematic differences between the trial sample and the ultimate population of inference with respect to these modifiers. Recent advances in the survey statistics literature to improve inference in nonprobability samples by using information from probability samples can provide an avenue for improving population causal inference in randomized controlled trials when relevant probability samples of the patient population are available. We review some recent work in “transporting” causal effect estimates from trials to populations, focusing on the setting where there is a “benchmark” or population-representative sample along with the RCT sample. We then propose estimators using either inverse probability weighting (IPWT) or prediction that can accommodate unequal probability of selection in the “benchmark” or population, and use Bayesian additive regression trees for both inverse probability of treatment weighting and prediction estimation that do not require specification of functional form or interaction. We also consider how the assumption of ignorability may be assessed from observed data and propose a sensitivity analysis under the failure of this assumption. We compare our proposed approach with existing methods in simulation and apply these alternative approaches to a study of pulmonary artery catheterization in critically ill patients. We also suggest next steps for future work.

show abstract

“…Alternative matching measures such as Mahalanobis distance have also been considered. 40 This is very closely related to the approach in Stuart et al 2011.…”

Section: Brief Review Of Non-probability Inferencementioning

confidence: 79%

Improving transportability of randomized controlled trial inference using robust prediction methods

Elliott,

Carroll,

Grieve

et al. 2023

Stat Methods Med Res

View full text Add to dashboard Cite

show abstract

“…For the future, we want to compare the proposed methodology with other techniques that are appearing to combine probability and nonprobability samples as Kim and Tam ( 2021 ) or Nandram et al. ( 2020 ). Extensions to small domain estimation (Rao & Molina, 2015 ) and variance estimation under nonparametric PSA will also be future research topics.…”

Section: Discussionmentioning

confidence: 99%

Enhancing estimation methods for integrating probability and nonprobability survey samples with machine‐learning techniques. An application to a Survey on the impact of the COVID‐19 pandemic in Spain

et al. 2022

View full text Add to dashboard Cite

Web surveys have replaced Face‐to‐Face and computer assisted telephone interviewing (CATI) as the main mode of data collection in most countries. This trend was reinforced as a consequence of COVID‐19 pandemic‐related restrictions. However, this mode still faces significant limitations in obtaining probability‐based samples of the general population. For this reason, most web surveys rely on nonprobability survey designs. Whereas probability‐based designs continue to be the gold standard in survey sampling, nonprobability web surveys may still prove useful in some situations. For instance, when small subpopulations are the group under study and probability sampling is unlikely to meet sample size requirements, complementing a small probability sample with a larger nonprobability one may improve the efficiency of the estimates. Nonprobability samples may also be designed as a mean for compensating for known biases in probability‐based web survey samples by purposely targeting respondent profiles that tend to be underrepresented in these surveys. This is the case in the Survey on the impact of the COVID‐19 pandemic in Spain (ESPACOV) that motivates this paper. In this paper, we propose a methodology for combining probability and nonprobability web‐based survey samples with the help of machine‐learning techniques. We then assess the efficiency of the resulting estimates by comparing them with other strategies that have been used before. Our simulation study and the application of the proposed estimation method to the second wave of the ESPACOV Survey allow us to conclude that this is the best option for reducing the biases observed in our data.

show abstract

“…The problem of combining a number of p-values, from the studies on the same subject, is one of data integration, which is currently a hot topic, see, for example, Nandram et al (2021) for model-based methods using both non-Bayesian and Bayesian approaches.…”

Section: Bayesian Model For Combining P-valuesmentioning

confidence: 99%

Combining Correlated P-values From Primary Data Analyses

Choi¹,

Nandram²,

Choi³

2022

IJSP

View full text Add to dashboard Cite

Research results on the same subject, extracted from scientific papers or clinical trials, are combined to determine a consensus. We are primarily concerned with combining p-values from experiments that may be correlated. We have two methods, a non-Bayesian method and a Bayesian method. We use a model to combine these results and assume the combined results follow a certain distribution, for example, chi-square or normal. The distribution requires independent and identically distributed (iid) random variables. When the data are correlated or non-iid, we cannot assume such distribution. In order to do so, the combined results from the model need to be adjusted, and the adjustment is done “indirectly” through two test statistics. Specifically, one test statistic (TS** ) is obtained for the non-iid data and the other is the test statistic (TS) is obtained for iid data. We use the ratio between the two test statistics to adjust the model test statistic (TS**) for its non-iid violation. The adjusted TS** is named as “effective test statistics” (ETS), which is then used for statistical inferences with the assumed distribution. As it is difficult to estimate the correlation, to provide a more coherent method for combining p-values, we also introduce a novel Bayesian method for both iid data and non-iid data. The examples are used to illustrate the non-Bayesian method and additional examples are given to illustrate the Bayesian method.

show abstract

Integration of Nonprobability and Probability Samples via Survey Weights

Cited by 6 publications

References 16 publications

Improving transportability of randomized controlled trial inference using robust prediction methods

Improving transportability of randomized controlled trial inference using robust prediction methods

Enhancing estimation methods for integrating probability and nonprobability survey samples with machine‐learning techniques. An application to a Survey on the impact of the COVID‐19 pandemic in Spain

Combining Correlated P-values From Primary Data Analyses

Contact Info

Product

Resources

About