Download full text
(910.5Kb)
Citation Suggestion
Please use the following Persistent Identifier (PID) to cite this document:
https://nbn-resolving.org/urn:nbn:de:0168-ssoar-427053
Exports for your reference manager
Finding Respondents in the Forest: A Comparison of Logistic Regression and Random Forest Models for Response Propensity Weighting and Stratification
[journal article]
Abstract Survey response rates for modern surveys using many different modes are trending downward leaving the potential for nonresponse biases
in estimates derived from using only the respondents. The reasons for nonresponse may be complex functions of known auxiliary variables or
unknown latent variables... view more
Survey response rates for modern surveys using many different modes are trending downward leaving the potential for nonresponse biases
in estimates derived from using only the respondents. The reasons for nonresponse may be complex functions of known auxiliary variables or
unknown latent variables not measured by practitioners. The degree to which the propensity to respond is associated with survey outcomes
casts light on the overall potential for nonresponse biases for estimates of means and totals. The most common method for nonresponse
adjustments to compensate for the potential bias in estimates has been logistic and probit regression models. However, for more complex
nonresponse mechanisms that may be nonlinear or involve many interaction effects, these methods may fail to converge and thus fail to
generate nonresponse adjustments for the sampling weights. In this paper we compare these traditional techniques to a relatively new data
mining technique- random forests – under a simple and complex nonresponse propensity population model using both direct and propensity
stratification nonresponse adjustments. Random forests appear to offer marginal improvements for the complex response model over logistic
regression in direct propensity adjustment, but have some surprising results for propensity stratification across both response models.... view less
Keywords
random sample; weighting; sampling error; regression analysis; model construction; regression; measurement; social stratification; response behavior; survey research; estimation
Classification
Methods and Techniques of Data Collection and Data Analysis, Statistical Methods, Computer Methods
Document language
English
Publication Year
2015
Page/Pages
p. 1-17
Journal
Survey Methods: Insights from the Field (2015)
Issue topic
Weighting: Practical Issues and ‘How to’ Approach
DOI
https://doi.org/10.13094/SMIF-2015-00003
ISSN
2296-4754
Status
Published Version; peer reviewed
Licence
Creative Commons - Attribution-Noncommercial-No Derivative Works