Search

We discuss ways that survey researchers can learn on the job and ways that managers, supervisors, and organization leadership can support and encourage on-the-job training (OJT). We discuss training that is needed beyond classroom survey methodology training and explore some OJT scenarios.

There is a wonderful challenge facing the field of survey methodology. With high demand for survey-based information and increasing complexity in the socio-economic environment within which we collect these data, survey researchers and survey methodologists are faced with ensuring our workforce is well trained to meet the need for high quality, timely, and cost-effective survey data. This article presents the impetus for this special issue of Survey Practice on training needs in survey research methods and provides an overview of our field of study, the job environment, and some of the key challenges to successful training in survey research.

This paper presents the outline of technical skills that a modern survey methodologist needs to be effective in research, study design and survey analyses positions. We provide lists of skills at three tiers: skills required for all survey methodologists, including those who concentrate on social science related research (outlined in another paper in this issue1 skills expected of the master-level survey statisticians; and skills at advanced level, typically associated with doctorate-level survey statisticians. We also identify the training gaps, (i.e., the skills from our lists that we consistently find to be the most difficult ones to develop in academic graduate program setting). We believe that these lists of proficiencies will be helpful to several groups of survey researchers: 1) current students who might have opportunities to find the courses leading to these proficiencies, or even to select the programs to enroll in; 2) early or mid-career statisticians who want to explore their opportunities in the survey profession; 3) and university professors and administrators who want to identify the opportunities to develop new courses and programs that can train highly employable graduates.

This final article of the special issue on survey training outlines the future of survey methodology education. While traditional academic programs will certainly play a major role in training survey specialists, we also see a growing role for a variety of other forms of continuing professional education. Flexible formats such as short courses, webinars, and other online programs and courses (e.g., massive open online courses or MOOCs) will make it easier for survey researchers to update their skills throughout their careers. Survey researchers of the future face the challenges of staying up-to-date with survey methodology findings and applying them to the increasingly-complex survey environment. This will likely involve additional interdisciplinary connections, particularly in new fields like big data and data science. In order to harness their potential in survey data collection, the survey researcher of the future must use a range of training options to keep his or her knowledge base fresh.

In this talk, I am demonstrating how a multiple imputation procedure can be built by a user from scratch. The motivating example comes from a public opinion survey in which the sampled respondents were providing their responses on the web or by phone. As is known in survey methodology literature, presence of an interviewer on the phone produces higher reports of socially desirable behaviors, such as number of friends or political engagement, or lower reports of undesirable behaviors, such as illicit drug use. Treating these less accurate responses as partially missing data, I develop a non-standard multiple imputation model that is driven by a concept of utility from choice and decision literature in economics. My implementation is aligned to supply the data to Stata -mi- suite, in the sense that I create the imputations, and -mi- combines them using Rubin's rules. As a side track, the workflow of the mode effect detection also featured multiple testing corrections. They required extensive -post- operations and exchanging the lists of variables between the do-files of the project. My work on this is also demonstrated.

In this talk, I introduce the main ideas of structural equation models (SEMs) with latent variables and Stata tools that can be used for such models. The two approaches most often used in applied work are numeric integration of the latent variables and covariance structure modeling. The first approach is implemented in Stata via gllamm, which was developed by Sophia Rabe-Hesketh. The second approach is currently implemented in confa for confirmatory factor analysis models. Also, introduction of the generalized method of moments (GMM) estimation and testing framework in Stata 11 make it possible to estimate SEMs by using moderately complex parameter and matrix manipulations. I provide working examples with some popular datasets (Holzinger–Swineford factor analysis model and Bollen’s industrialization and political democracy model).

This talk will introduce the main ideas of structural equation models (SEMs) with latent variables and Stata tools that can be used for such models. The two approaches most often used in the applied work are numeric integration of the latent variables and covariance structure modeling. The first approach is implemented in Stata via -gllamm- (developed by Sophia Rabe-Hesketh). The second approach is currently implemented in -confa- for confirmatory factor analysis models. Also, introduction of the generalized method of moments (GMM) estimation and testing framework in version 11 of Stata made it possible to estimate SEMs by using moderately complex parameter and matrix manipulations. Working examples will be provided with some popular data sets (Holzinger-Swineford factor analysis model and Bollen's industrialization and political democracy model).

In this presentation, I will review the bootstrap for complex surveys with designs featuring stratification, clustering, and unequal probability weights. I will present the Stata module bsweights, which creates the bootstrap weights for designs specified through and supported by svy. I will also provide simple demonstrations highlighting the use of the procedure and its syntax. I will discuss various tuning parameters and their impact on the performance of the procedure, and I will give arguments for the bootstrap by the method of weights in nonsurvey settings.

I will present a set of routines to conduct a one-factor confirmatory factor analysis in Stata. The use of Mata in programming will be highlighted. Corrections for non-normality, as common in the structural equation modeling literature, will be demonstrated. Indications for further development into multifactor models and, eventually, structural equation models, will be given.