Speed Talk Session 13

Back to Schedule

Title of session: Innovation in statistical production processes

Chair:

Room: Lecture room 3

Time: 13:15 - 14:00

Date: 28 June


Presenting AuthorAbstract
Edin Šabanović
e-mail: edin.sabanovic@bhas.gov.ba
Title: <<< About the Quality of the 2013 Census of Population, Households and Dwellings in Bosnia and Herzegovina: Statistical Results versus Political Controversies >>>
Census of Population in Bosnia and Herzegovina was conducted in 2013. Statistical offices conducted census in line with international standards. However, almost all main issues of the Census were politicized. The Census was given many non-statistical characteristics. It was presented by various interest groups as politically important activity, as a population register or a source of data, which could be used for tax policy or property management. For all these reasons, census activities were a subject of political controversies, topics for evaluation of several, mostly statistically unprofessional interest groups or individuals, who had more media attention than statistical expertise and whose influence to the attitude of Bosnian citizens was non-ignorable. The Post-enumeration survey (PES) was conducted in November 2013. The objective of the PES is to provide indicators of the census. In this paper, we will show how the PES was designed and conducted and which statistical models and methods were used for linkage of Census and PES data and for the analysis of Census quality. The last part of the paper will discuss the acceptance of the Census results by different data users such as statistical institutions, journalists, general and expert public. This part of the paper aims to give a rough picture of the general statistical culture in Bosnia and Herzegovina through the prism of the Population Census. In conclusion the evaluation of the quality of 2013 Census of Population, Households and Dwellings in Bosnia and Herzegovina will be done as well as some proposals for future work.
Daniel Roash
e-mail: danielr@cbs.gov.il
Title: <<< Improving Monthly Estimates of Job Vacancies Survey With Statistical Model >>>
The monthly Job Vacancies Survey in Israel began in 2009 by the Central Bureau Of Statistics. The survey's goals are: To serve as a leading indicator for the cyclicality in the labour market (during recession firms will begin by reducing their job openings and only later proceed with dismissing employees). To aid in assessing the demand for labour and identifying work opportunities by industry and composition of employed persons. To supply a broad view of the labour market by comparing estimates of job vacancies and the profile of workers requested by employers (labour demand) against the estimates of job seekers and their profiles as derived from labor force survey data (labour supply). In order to reduce the response burden, it was determined that the firms that belongs to the take some strata (small and medium firms) will be divided randomly during the quarter. While the large firms from the take all strata will report monthly. The strata are determined by the number of employees in the firm while the survey's target is the number of vacancies in each firm. Sometimes those variables (employees, vacancies) do not have high correlation (there is a large variation of the vacancies between firms in the same strata). This sample design led to high variation of the monthly estimates with an error pattern and a high correlation to the estimates in period t-3. We applied a State Space model that reduces the sample error and smooth's the estimates. This process improved the quality of the survey's estimates and I will present this method which is strongly relevant to surveys with panel sample. The presentation will also analyze the unique characters of the vacancies survey and it implication to imputing missing items.
Sabine Schuster
e-mail: sabine.schuster@statistik.gv.at
Title: <<< Flash estimates and increasing data quality as a result of modern IT-applications and job satisfaction >>>
From 2017 onwards, Statistics Austria publishes flash estimates for road freight transport statistics. These flash estimates enable data users to get first key figures of freight transport of Austrian road freight vehicles already one month after the reference quarter (t+1) instead of five months (t+5). Road freight transport statistics is – with reference to Council Regulation (EC) No 70/2012 on statistical returns in respect of the carriage of goods by road (recast) – performed as a sample survey. Weekly, questionnaires are sent to the respondents, who have to report all journeys of the selected trucks and return the web- or paper-based form to Statistics Austria two weeks later. To check the plausibility of the filled-in forms a Java-application is used since 2014. Prior to that, the reprocessing was very complicated and included different steps and different people in charge. Now, each questionnaire is attributed to one staff member, who is registering its completeness, checking its plausibility and - if necessary - is getting in contact with the respondent to clarify certain parameters. Once finalised, each questionnaire is directly added to the authentic database. Following the two-factor theory of Herzberg about job satisfaction, it was noticed, that the motivation of the staff members improved due to intrinsic factors such as increased responsibility and decision autonomy. The work tasks have become more demanding and therefore interesting with a better insight and understanding of the whole process. Thereby the quality and availability of the produced data enhanced. It was analysed, that one month after the reference quarter already more than 70% of the questionnaires could be correctly finalised. Calculated flash estimates showed similar outcomes to the final key results with a slightly higher sampling error which is still in the frame of precision requirements.
Michael Blohm
e-mail: michael.blohm@gesis.org
Title: <<< Do declining response rates negatively affect sample composition? A longitudinal analysis using data from the German General Social Survey (ALLBUS) >>>
The German General Social Survey (ALLBUS) is a repeated cross-sectional multi–thematic survey conducted every other year since 1980. Like many other surveys ALLBUS has been facing the problem of declining response rates. Since 1994 ALLBUS response rates decreased steadily from 54% to 35% at present – despite exerting higher fieldwork efforts. Declining response rates give rise to the question whether this affects survey data quality. Selective participation behavior might threaten the representativeness of a survey in general or with regard to specific questions. In our presentation we use ALLBUS data to analyze the relationship between the response rate and indicators of nonresponse bias over time. The ALLBUS provides a good basis to analyze this relationship, since it has used an almost identical study design since 1994. This includes the definition of the target population (adults living in private households in Germany), the sample design (samples of named individuals), the net sample size (3.500 completed interviews), the interview duration (around 70 minutes), and – particularly important in this context – a consistent calculation of response rates. To investigate nonresponse bias we analyse a) survey participation in ALLBUS with sample frame information from the population registers and b) deviations of the net sample from external benchmark data (namely the German (micro-) census from official statistics). First results indicate that decreasing response rates are not systematically associated with deterioration in demographic sample composition. We will argue that these findings only hold in “controlled” sampling design settings (for instance, when samples of named individuals have been used).
Barbara Kowalczyk
e-mail: bkowal@sgh.waw.pl
Title: <<< Design and analysis of the improved Poisson and negative binomial item count techniques >>>
Reliable data on stigmatizing, socially unaccepted or illegal features are very hard to obtain in direct questioning. Many indirect methods of questioning have been developed to help in eliciting honest answers to sensitive questions and to eliminate the social desirability bias. Item count techniques (ICTs) pioneered by Miller (1984) constitute an example of indirect survey techniques designed to deal with sensitive features. These techniques have many practical advantages: they are easy to implement, can be used in telephone and internet surveys, and the way they protect privacy is easy to understand by respondents. Recently Tian et al. (2017) proposed new item count techniques called Poisson and negative binomial ICTs. The new methods give many opportunities for further theoretical and practical developments. But the methods require large sample sizes to obtain a reasonable precision. Efficiency is an important issue in indirect methods of questioning. Protection of respondents’ privacy is usually achieved at the expense of the efficiency of the estimation. In the present paper we propose new improved Poisson and negative binomial ICTs in which each of the two subsamples serve both as a control and a treatment group. This procedure allows to increase efficiency of the estimation as compared to the classic Poisson and negative binomial ICTs and maintain respondents’ privacy at the same level. In the paper we introduce methodology of the proposed improved method and accompanying statistical theory. We analyze and compare best linear unbiased and maximum likelihood estimators of the population proportion of the sensitive attribute. We also compare the improved technique with previously proposed classic Poisson and negative binomial ICTs. The improvement is obtained in terms of efficiency. Gain in efficiency is achieved without affecting the privacy of respondents. Theoretical results presented in the paper are illustrated by comprehensive simulation studies.

Back to Schedule

Font Resize
Contrast