Issues of Nonresponse and Imputation in the Survey of Income and Program Participation

Kalton, Graham; Kasprzyk, Daniel; Santos, Robert

doi:10.1016/b978-0-12-426280-5.50032-0

Cited by 39 publications

(23 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For respondents who were employees in both waves, and had not changed their jobs, 32 per cent registered a change in their 3-digit SOC code, and almost 18 per cent changed their SOC major group (in similar work on the US Survey of Income and Program Participation, Kalton and McMillen (1986) found 40 per cent change at the three digit level). 15 A sample of the relevant questionnaires was examined and while there were various explanations for the discrepancies, the bulk of cases were caused by either different codings of what was essentially the same job description, or different descriptions of the same job.…”

Section: Measurement Error In Job Description: Some Evidencementioning

confidence: 93%

Unified BHPS Work-Life Histories: Combining Multiple Sources into a User-Friendly Format

Halpin

1998

Bulletin of Sociological Methodology/Bulletin de Méthodologie S

View full text Add to dashboard Cite

The British Household Panel Study collects extensive labour market history information from its respondents, both during the panel period and retrospectively from labour market entry. That this information is of necessity stored in multiple locations, and of varying levels of detail, has made use somewhat inconvenient. This paper describes an exercise to bring the labour market information together in a more convenient format. It also considers some of the problems of retrospective and panel longitudinal data, and discusses issues of recall error and measurement error.The data files described will be made available through the Data Archive. SummaryThe British Household Panel Study collects extensive information on respondents' labour market status, (i) at the time of interview at each wave of the panel, (ii) through the period between 1 September a year before and the interview date, and (iii) retrospectively from first leaving full-time education. Because the retrospective information was collected in two tranches (one focusing on employment status, the other on occupational information) there are four different types of labour-market history information, located (at Wave 5) in twelve different files in the BHPS database. This complexity is a necessary aspect of longitudinal information, but it has inhibited use of the work-life history data. In order to facilitate such use, a set of 'reconciled' files has been created, constituting single continuous records each containing all the information of a particular type in a single location. This paper describes the creation of the files, examines their output and discusses some aspects of measurement error and recall bias relevant to the exercise.The first part of the exercise is to take the 'current status' information and combine it with the inter-wave history, for each wave, and then to combine the five waves thus creating a continuous record from September 1990 to the September 1995 (and later). The second stage is to take the life-time employment status history collected at Wave 2, and the life-time occupational history collected at Wave 3, and to combine each of them with analogous information drawn from the combined panel file, thus creating employment and occupational histories that stretch from labour-market entry to the latest wave. The third stage is to combine these two extended life-time histories into a single record which contains both employment-status information (with good information about non-employed spells) and occupational information (that is, details about the job held during each employed spell).The paper describes the methods used, in terms of an initial specification and its detailed implementation, and goes on to consider the output produced. By including the retrospective data we have information stretching back many decades, though from the point of view of breatdh of detail and quality of recall the panelderived data (covering 1990-95) are much better. When we compare data from different sources, we find systematic difference...

show abstract

Section: Measurement Error In Job Description: Some Evidencementioning

confidence: 93%

Unified BHPS Work-Life Histories: Combining Multiple Sources into a User-Friendly Format

Halpin

1998

Bulletin of Sociological Methodology/Bulletin de Méthodologie S

View full text Add to dashboard Cite

show abstract

“…Rubin [1] addressed the missing data concept and suggested various imputation methods which make the data structurally complete at the beginning of the analysis. Some important works based on imputation method were carried out by [2], [3], [4] and [5]. Later on utilizing the information on an auxiliary variable and using the missing completely at random (MCAR) response mechanism, [6], [7], [8], [9], [10], [11], [12], [13] and [14] among others have suggested several interesting imputation methods with success.…”

Section: Introductionmentioning

confidence: 99%

On the use of imputation methods for missing data in estimation of population mean under two-phase sampling design

Singh¹,

Suman²,

Kadılar³

2018

HJMS

View full text Add to dashboard Cite

Non-response is an unavoidable feature in sample surveys and it needs to be carefully handled to avoid the biased estimates of population characteristics/parameters. Imputation is one of the latest fascinating methods which is attracting the attention of survey practitioners to deal with the problems of non-response because it makes the survey data complete before the beginning of generating the survey estimates. The present work proposes some new imputation methods to compensate the missing data in two-phase sampling when non-response observed in samples of both the phases. The proposed imputation methods result in chain type estimators of population mean of study variable and the resultant estimators have shown the ecacious performances in terms of producing the more precise estimates. Properties of the proposed estimators are examined with the help of empirical and simulation studies. Results are critically analyzed and suitable recommendations are put forward to the survey practitioners.

show abstract

“…It is typically used when it is needed to substitute missing item values with certain fabricated values in a survey or census. To deal with the missing item values effectively [13], [14], [16] and [9] suggested imputation methods that make an incomplete data set structurally complete and its analysis simple. Imputation may also be carried out with the aid of an auxiliary variable if it is available.…”

Section: Introductionmentioning

confidence: 99%

Some imputation methods for missing data in sample surveys

Singh¹,

Maurya²,

Khetan³

et al. 2015

HJMS

View full text Add to dashboard Cite

The present work suggests some imputation methods to deal with the problems of non-response in sample surveys. The imputation methods presented in this work lead to the precise estimation strategies of population mean. Empirical studies are carried out with the help of data borrowed from natural populations to show the superiorities of the suggested imputation methods over usual mean, ratio and regression methods of imputation in terms of the mean square error criterions. Suitable recommendations have been put forward for the survey practitioners.

show abstract

Issues of Nonresponse and Imputation in the Survey of Income and Program Participation

Cited by 39 publications

References 1 publication

Unified BHPS Work-Life Histories: Combining Multiple Sources into a User-Friendly Format

Unified BHPS Work-Life Histories: Combining Multiple Sources into a User-Friendly Format

On the use of imputation methods for missing data in estimation of population mean under two-phase sampling design

Some imputation methods for missing data in sample surveys

Contact Info

Product

Resources

About