|Convenor||Professor Bart Meuleman (University of Leuven )|
The paper aims to examine whether the probability of producing item-nonresponse was higher in a concurrent mixed-mode (CATI and online) than in a single-mode (CATI) treatment. Therefore data from an experimental health survey is used. The focus is of this analysis is on especially sensitive items covering income, drug consumption and sexual activity. Furthermore dropouts throughout the survey and their connection to item sensitivity and survey mode are analyzed.
We use matched survey and administrative data to study interviewer and interview related
determinants of misreporting on welfare receipt in interviews. In our data, 12.2 % of German welfare
recipients underreport benefit receipt. We find that underreporting is more likely in formal and
standardized interviews compared to those with a more conversational character. Further, low
interviewer education and matched interviewer-respondent characteristics with respect to
immigration and education are associated with higher reporting quality.
This paper introduces the concept of prejudicial nonresponse in experiments and quasi-experiments. Prejudicial nonresponse is a situation in which the non-response to a survey or similar data collection method is not ignorable, and where moreover the response mechanism for the treatment group differs from the response mechanism for the control group. The evidence suggests that when the parameter of interest is an odds ratio on a binary outcome variable, analyses can draw internally valid conclusions even with non-ignorable nonresponse, unless the response mechanism is prejudicial. The usefulness of prejudicial nonresponse as a tool-of-thought is discussed.
This paper uses the newest ESS data from all six rounds including the countries participating in two rounds at minimum. The ESS quality has been improving during six rounds from the point of view of the happiness study especially since it requires good control variables. Hence it is possible to estimate the age happiness better than earlier. The main focus is on estimating whether the relationship between age and happiness is U-shaped that has been suggested by recent psychological economic theories. Our cross-sectional results do not completely support this theory.
Although data on attitudes towards domestic violence (wife beating) collected from a household-based survey with complex sample design (e.g., Multiple Indicator Cluster Survey or Demographic Health Survey) has features of a repeated measures data, in practice, this feature commonly has been neglected for data analysis. In this paper, multilevel modeling is proposed as an alternative, so that the potential dependence among the responses of same respondents (e.g., women express their opinion regarding several scenarios where domestic violence is likely to be justified) is taken into account along with the complex sample design feature of the survey.
Near-peer teaching and learning is a common practice in the medical profession. However, measurements of the impact of near-peer teaching and learning on participating students are scarce. The current study reports on the development and validation of a scale for perception of effectiveness of near-peer learning and learning in a medical program.
The scale consists of 21 items found consists of three factors explaining 47% of the variance: Benefits of near-peer learning; Benefits of near-peer teaching; Promotes future near-peer teaching and learning. CFA indices support the construct validity of the scale.
Randomness, probability e representativeness: playing with the concepts in sample surveys
Both communities sociologists and statisticians rely on scientific methodologies for producing reliable and verifiable knowledge of the objective reality. However, often common sense and scientific words as well, have lost their sharpness and thus in a sense their scientific trait for we have misused them, consumed with inappropriate over intended meaning even if unawares. This appears to be the case also for crucial terms such as randomness, representativeness and probability which are often improperly used in social sciences with effects both perverse and limiting.
In an effort to improve the employment data collected by universities, we rely on an external source of data to ensure the validity of survey data. Using a multi-mode alumni survey (web and phone) and employment data from LinkedIn, a social networking site centered around career connections, we can measure the validity of how recent graduates respond to questions about their employment status with data provided by the social networking website. The goal of this paper is to test the validity of a new methodology for surveyors interested in collecting occupational and employment data.
This poster presents the Survey Quality Predictor (SQP) 2.0, an online program to predict the quality of survey questions, available for free at sqp.upf.edu. SQP 2.0 is also a database of questions with quality estimates and predictions. Using this program, the users can obtain a prediction of the question quality in many languages and for more than 20 countries. Thus, SQP 2.0 is a very powerful tool both at the stage of questionnaire design and at the stage of data analysis.
Analysing the 5th wave of the World Values Survey, items on the perception of ecological burdens and suitable measures for environmental protection will be compared among industrialised countried, emerging economies and devolping countries. Within societes different socio-economic groups will be examined concerning possible relations between environmentalism, perception on environmental problems and the individual ability and attitudes how to deal with those.
We have conducted several experiments in Japan aiming to explore mail survey implementation methods that could improve response rate. The examined factors include handwritten request letter, professional designed color leaflet, translucent envelope through which people can see an incentive without opening the envelope, a small number of questionnaire pages with two-column format, and sending a questionnaire so that it will be delivered on January 1st when most Japanese are paying attention to their mailbox expecting New Year's cards. All the factors examined seemed to have little effect on the improvement of response rate.
SANT DataLink has developed a system for managing linked data using graph databases and graph (network) algorithms.
This approach allows great flexibility in the management and analysis of linked data in terms of: research project composition; the timeliness, quality and richness of data provided to researchers; the ability to add and remove data sets without having to 'relink' data; the improvement of data quality.
The system stores data in a graph database and traverses the graph using graph algorithms to extract clusters of related records.
The lessons learnt during the development of this significant project will be presented.
The paper presents an innovative analytic technique for the analysis of survey data based on network analysis. It uses data from two surveys on sustainable consumption, one conducted in 2005 with a representative sample (1500 questionnaires) of the Italian population, and the other collected monthly from July 2010 to April 2011 from a large sample of UK Tesco consumers (26366 questionnaires). In both examples I will analyse the structure of connections between variables at different levels of correlations, and then I will focus on some interesting local neighbourhoods that suggest elements for interpretative frameworks.
Within this paper I attempt:
(1) To compare the communication strategies of the interviewed persons and their answers during the interviews and in their usual working environment;
(2) To compare the manner of speaking about the issues and the ideas expressed by the same persons in several interviews, conducted with an interval of about two years;
(3) To see if the change of a status of an interviewer is influencing the character and the content of the answers and the behavior the interviewed during the interview;
(4) To develop approaches enabling t interpreting the interviews under such conditions.
Imputing missing data in longitudinal panel studies presents special problems when there are many waves and many variables in the analysis. Missing data can stem from missed waves and on items within waves. In this study we compare several approaches to handling missing data by replicating a multilevel analysis which included 35 variables and 8 waves in a dataset with 2022 participants. We used multiple imputation approaches in the wide and long formats, and with a two-fold conditional expectation method which only informs the missing values with variables in adjacent waves. The results found little gain from the imputations
A form processing application (as HP Teleform) allows transforming scanned forms in an electronic format. The aim of the contribution is to present the whole development of such a solution in a practical project where no previous knowledge in this matter was available in the team. We discuss the advantages and the limits of the approach, the changes in the way the questionnaire is designed and the data collected. Its costs is also be addressed (licenses, hardware, workforce). Moreover, we present the duration of each step of the encoding phase. Finally, the quality of the process is assessed and discussed.
In 2013 an online registration tool for time-use research, named MOTUS (Modular Online Time Use Survey), was developed and tested. However, nowadays people increasingly use smartphones and tablets to consult the Internet and to share information. Therefore, the Research Group TOR decided to also develop an application for smartphones and tablets. Such an application offers great opportunities for (online) time-use research. It allows people to register real-time behaviour.
We investigated the Social Hygiene Monitoring datasets to explore: i) social-medical understanding of refusal of risky behaviour (smoking) and ii) modeling regarding "past smoking vs present". Data sources were Surveys at Vocational Colleges (N=1554) and at High Schools (N=1500) in Russian city. In logistic regression models dependent variable was a dichotomous «Experienced Non-Smoker», the significant predictors were: i) Knowledge about Smoking (Own-Experience and Self-Observation vs Adults&Mass-media); ii) age at first smoking; iv) Gender; v) school performance. This reveals Self-reflexivity as a Personality feature that prevents teenager from continuing to smoke.
Internet access is easy to measure, while intensity of Internet use remains somewhat elusive. The current work presents an index to measure the intensity of Internet use on the basis of two large Canadian surveys (CIUS 2011, 2013 [Canadian Internet Use Survey]). Based on a ranked and additive number of 22 online platforms people access, the proposed Internet Intensity Index weights the time spent online in 2010 (replicated in 2012). The intensity index assists our understanding which social groups most intensely surfs the Internet. The ambiguity of what this index captures is discussed regarding online tracking technologies and data sovereignty.
Everybody is Doing Beauty (refers to the German word Schönheitshandeln) – women use make-up daily and men shave. They have one thing in common: Doing Beauty means portraying yourself and securing one's identity.
However, are there any differences between heterosexual and homo-/bisexual men and women respectively?
The focus is the analysis of these differences: Does Doing Beauty vary between groups of different sexual orientation? The analysed Beauty Actions (refers to the German word Schönheitshandlung) cover a broad range: From day-to-day actions to more durable ones. Furthermore the satisfaction with the own body is analysed.
We present the results of a quasi-experiment on promised versus prepaid incentives during the recruitment process for the GESIS Panel, a self-administered mixed-mode panel of the general population. First analyses show a substantial increase of survey participation from 77% (promised) to 90% (prepaid). The increase of participation rate is more pronounced in the group of offline respondents (66% promised, 88% prepaid) in comparison to online respondents (84% promised, 93% prepaid). We also find that mailed prepaid incentives increase panelists´ motivation to keep their contact addresses up to date.
Using data from the German Socio-Economic Panel (SOEP) I evaluate the hypothesis that individuals connected to a distance communication network exhibit a higher propensity to trust others than individuals outside the network. Using a broad set of statistical tools, I find evidence of a causal effect of network connectivity on trust, mediated by beliefs about others’ trustworthiness.
This paper is motivated by some common misconceptions among students on random sampling in survey research. In order to identify if these misconceptions are common among researchers, the author conducted a content analysis of 38 journal articles that have been published in peer reviewed journals from July 2013 to June 2014. Two misconceptions were identified among young researchers and in published articles i.e. meaning of random sampling and the use of inferential statistics with data from a non-random sample. The number of published article with these misconceptions is small but it presents a challenge to the teaching survey.
The ability to use knowledge and skills to effectively manage financial resources is described by the concept of financial literacy. Although, it is substantial for many research questions, financial literacy is seldom measured directly in general surveys.
We developed an eight item Likert scale which ran in 2014 on the 8th wave of the German panel study `Labour Market and Social Security' (PASS).
To verify that the same construct is measured across different groups, we perform tests of measurement invariance. We focus on differences between person with low and high income as well as person with low and high education.
Statistics Canada, like many statistical organizations throughout the world, has observed a downward trend in response rates. Statistics Canada is currently investigating many options to improve data collection strategies and processes. In the meantime, recent surveys and field tests have provided us with a better understanding of factors that currently impact response rates. This paper provides an overview of the Statistics Canada’s plans to improve response rates and describes the results and lessons learned from these recent surveys and tests.
Four methods using telephone paradata (days and hour of call, number of calls, refusals, non-contact, appointments, beginning of questionnaire) are compared in the weighting process of a random survey, all based on homogenous response groups to correct for total non-response: logit modeling, clustering (based on the total calls or by call days), harmonic qualitative analysis. The respondents sample (n=8638) is considered the total population and three non-response mechanisms are simulated within (50% rate): MCAR, MAR, NMAR. Empirical biases and variances for 9 variables are computed based on 1000 replications for each method.
This paper reports on a household survey focused on electric vehicle (EV) uptake in Western Australia.
Besides information on household's characteristics and attitudes towards environment, renewable energy, new technologies, eight choice experiments with four alternatives helped us gauge the valuation of various EV features.
An initial sample showed high interest in EV and environmentally friendly technologies, and over-represented households with higher education and socio-economics status, therefore an additional online panel sample was used to compensate for the bias. Results from the choice models indicate significant differences between these samples, and between best only and best-worst data.
In this paper we seek to contribute to the understanding on how best to achieve a uniformity of approach in cross-country cognitive interviewing. In the past few years we have seen the growth of cognitive interview research from small scale 30 to 40 respondent studies to studies that include over 100 respondents. These interviews are now been conducted in short periods of time (3- 4 months). In this paper we outline the strategies that we have designed to improve uniformity in data collection and maintain data quality across geographical locations and languages.
Over the past decades, labor market participation has become an inherent part of mothers’ life course in Germany. In order to support work-life balance, family-friendly policies have become a central issue of the human resources policy in establishments. By focusing on long-term trends over time, the study identifies organizational and context factors, which induce establishments to initiate work-life balance measures for their employees.
We use the “IAB Establishment Panel”, conducted by the Institute for Employment Research (IAB) in Germany. Based upon longitudinal and multilevel analyses, we investigate the organizational change of implementing measures over time.
Sharing data has become the norm; however this requires an infrastructure to manage it. As there are different platforms across which data are shared, interoperability is critical to ensure continued collaboration and data sharing across disciplines and institutions. Statistics Canada has been providing access to Canadian research microdata for 18 years through the Data Liberation Initiative (DLI) and Research Data Centre (RDC) Program. This session will explore the importance of metadata interoperability, with emphasis on the use of the Data Documentation Initiative (DDI) standard to create survey metadata for its master and public use microdata in Microdata Access Division.
The paper deals with the question of work-family conflict focusing on men, assuming that they face the difficulties of harmonizing work and private life differently depending on their employment status, age, marital status and other characteristics. A cross-national analysis is used on the data of European Social Survey’s 5th round to see which attitude, work and family factors influence the perception of European men on work-to-family and family-to-work conflicts. The participating countries are sorted in four groups according to a typology based on welfare and gender dimensions.
This paper explores time use allocation using a sample of children and adolescents from the Child Development Supplement of the Panel Study of Income Dynamics. Tobit and OLS estimation results showed differences in time use patterns according to age, gender, race, and maternal education and employment. Given a constrained endowment of 24 hours each day, a higher participation in one activity reduces the chance for engaging in others. Where does the trade-off take place? The results from an economic share equation approach reveal similar general findings.
Assessment and diagnosis of adult attention deficit/hyperactivity disorder (ADHD) is complex, involving multiple tasks and several sources of information. This study aims to study the psychometric properties of the Catalan adaptation of the Conners Adult ADHD Rating Scales—Long Version (CAARS-L), in its self-report and observer forms. The sample comprised 606 participants. Dimensional structure, age and gender differences, reliability of the test and differential item functioning were analysed. The results show that the Catalan versions of the CAARS-L present adequate validity and reliability, and they can therefore be used for diagnostic purposes and cross-cultural comparisons.
We designed a feasibility study to see whether a CAPI/CASI survey of self-reported criminal activity among children and young adults from 10 to 22 years of age can be redesigned as a web survey. As we know very little of the response propensities in web of this group, we designed an experiment, in which we varied with incentives (promised iPad-minis versus no incentive) and additional materials (a large envelop versus a standard envelop, and a sheet of example questions with the advance letter versus no examples). Paradata will allow extensive monitoring of children's behaviour.
Entrepreneurial orientation involves key entrepreneurial processes related to how new ventures are undertaken. The concept entails psychological orientations of the owner that relate to the owner’s daily tasks and fit with the environmental requirements. Despite the importance of measuring and promoting entrepreneurial orientation, there is a scarcity of measures for evaluating this construct in the educational field. Thus, the present study aimed to develop a scale for assessing entrepreneurial orientation in the academic environment. After developing a preliminary set of scale items, an expert consultation, cognitive interviews and a pilot study were carried out to examine the content validity.
The aim of our communication is to investigate the time necessary to join a person in a CATI survey. We propose a time to event model with a starting time at the first attempt to join a person by a Pareto distribution function with two parameters which are related to the “reachabilty” of a person. The model takes also into account the daily density of attempts of call. The model is estimated on the paradata about a sample of persons elected to be interviewed for a pilot survey of the Swiss household panel. Results give some insights about the duration
The School Climate Surveys are a suite of survey instruments being developed for schools, school districts, and states in the United States. In this poster we present findings from the cognitive interviews conducted with parents and students in English and Spanish and focus on findings for the items that address questions of bullying and cyberbullying. We show examples of the tested items and the final version of these questions.
We focus on variations in panel attrition patterns through short periods of time. The project’s underlying database is the AID:A survey of the German Youth Institute whose second wave of interviewing was conducted in 2013. This particular wave also targeted children aged 1-4. First wave interviews took place between 1-4 years before the second interviews in 2013, thus enabling us to compare attrition patterns in total over a 5 year period.
In the city of Guadalajara, in Jalisco México, every year, from 2011 to 2014, a nonprofit organization has conducted a survey of quality of life. Through this survey they analyzed the subjective well-being through direct questions of satisfaction with life. It has also developed an index through the analysis of different dimensions and determinants of quality of life. The results reflect different levels of analysis, mediation of perception. On the one hand, a data tracking that allows observing affectations in specific areas, and the other, citizen’s feelings, expression of satisfaction with life as a whole.
Social integration theory offers an explanation for nonresponse. Individuals who are more integrated are more likely to respond to a survey request because they feel compelled to adhere to their group norms or perceive that their participation may result in benefits for their group. Integration is a latent construct that cannot be directly measured. We conduct a latent class analysis using data available for all sampled individuals to create a social integration index. We then place the index in a regression model to predict response and test the theory.
Recent studies indicate that cognitive processes underlying creative problem-solving tasks are influenced by affective dispositions and motivational preferences. Given the scarcity of measures that jointly examine cognition, emotions, and motivation, the aim of the present study was to develop the "Emotion/motivation-related Divergent/Convergent styles Scale" (EDICOS). Moreover, we aimed to provide different sources of validity evidence and to test its reliability in a large sample of higher education students.
Overall, the obtained results justify the future use of EDICOS for the assessment of individual differences in creativity and the identification of particular cognitive/affective/motivational associations.
This contribution presents systematic biases in the process of generating health data by using a step-by-step explanation of the DISEASE FILTER, a heuristic instrument that we designed in order to better understand and evaluate health data. The systematic bias in health data generally varies by data type and the operationalization of health outcomes. Self-reported subjective health and disease assessments, for instance, underlie a different selectivity than do data based on medical examinations or health care statistics. The contribution give some suggestions on how to deal with biases in health data and how to avoid them.
The paper analyzes the level of integration of the Third Country Nationals (TCN) in Luxembourg with respect to membership and volunteering in expressive and instrumental associations. The paper uses the EU-SILC/PSELL data from 2011. The TCNs are compared to natives and immigrants from the EU-countries. Propensity score matching is employed to assure comparability of the analyzed groups with respect to observed covariates and to estimate correctly group differences in the level of associative behaviour. The results of propensity score matching are compared to the outcomes of the regression analysis.
The paper explores whether the Greek economic crisis has triggered the interest of the global academic community and whether the country has acted as a ‘case study’ of the etiologies and impacts of an unprecedented crisis. Based on techniques of systematic review methodology and collecting data from academic databases the analysis demonstrates the exponential increase in publishing scientific papers associated with the Greek economic crisis in different scientific fields and from scholars worldwide. The focus specifically on the social sciences field indicates similar increasing trends; however a further exploration unveils that an empirical approach is still in a fledgling state.
The decision to participate in a survey may be associated with particular interests and leisure activities. Abraham, Helms and Presser (2009) demonstrated in a study on American data that people who volunteer are more likely to participate in surveys. The apparent consequence of such a pattern among potential respondents is that estimates of volunteering could be biased. An analysis of panel data from a Danish population survey on volunteering show that people who volunteer themselves are more likely to remain in the panel.
In psychological and social surveys, people often do not respond truthfully when asked personal or sensitive questions. Warner (1965) developed a data collection procedure, the randomized response technique (RRT), which allows researchers to obtain sensitive information while guaranteeing privacy to respondents
Warner’s study generated a rapidly-expanding body of alternative techniques for eliciting suitable RR schemes in order to estimate a population proportion. However, most statistical commercial software does not incorporate estimation procedures for RRT. In this paper we present a R package that implement some important randomized response techniques for estimating totals and means of qualitative and quantitative
A study on the data of Andalusian students from the PISA survey is carried out, so that the students’ competencies on provinces and shires with similar characteristics can be evaluated. As the Andalusian sample size is small, calibration, parametric and nonparametric methods are used to obtain efficient-enough indirect estimates.
Putnam suggests that social connectedness is essential in fostering generalized trust. Though the notion of social connectedness may refer to very different types of social ties, the vast majority of the literature has tested this claim by focusing on the role of associational membership. In this paper, I probe the solidity of Putnam's argument in the US context by looking at a variety of social ties (e.g. friendship and neighborhood relations). In addition, I test two main mechanisms that might explain why interactions with people we know would create trust in people we do not know.
The time required to conceive is an increasingly frequent concern among couples of reproductive age. No data on the prevalence of infertility in southern Spain have been published by population studies. The objective of this study is to use different approaches to determine the prevalence of infertility among women of childbearing age in a population of southern Spain. For doing this study we have conducted a face-to-face cross-sectional population-based survey of 443 women aged between 30 and 49 years residing in Huelva, southern Spain. Calibration estimators were used to extrapolate the infertility prevalence to Spain.
EU-SILC has become the major source of statistics and indicators on income, social inclusion and living conditions in Europe. Ensuring the accuracy of the measurements and their evolution over time is becoming paramount as well as the continuous improvement of the whole quality of the instrument. Eurostat has implemented an approach to measure standard error of SILC based estimates and of net change. At the same time the quality reporting process has been streamlined. The paper aims at describing the outcomes of these two actions aiming at improving the quality of the EU-SILC instrument.
This contribution deals with a methodological problem in the analysis of social change, and changing value orientations on the basis of replicative surveys. It is analysed, how changes in the Interviewer age composition of the surveys affect the time-series for post-materialist value orientations. The data based on the cumulative German General Social Survey (GGSS). These consist of independent cross-sectional surveys of the adult household population of Germany. 18 surveys conducted between 1980 and 2012 in West Germany, the average age of interviewer varied between 39 and 61 years.
Sentiment analysis is a methodology with increasing diffusion. Literature review reports several negative cases in which web listening supported the improvement of damaged corporates’ positioning.
Furthermore web listenting is not limitless: multiple time periods, multiple metrics, multiple venues and multiple venues with different metrics. Scheweidel et al. (2015) propose to model the relationship between what people post and where they post in an effort to extract the underlying brand sentiment across multiple social media venues.In this paper the Schweidl model was modified using a cluster sampling of enterprises in the luxury sectors and a Latent a Latent Semantic Analysis.
The process of the demographic modernization became widely-spread in the second half of the XX century in majority of the developed countries and has led to a number of changes, in particular, liberalization of family and marriage relations and, as a result, growth of number of illegitimate births.
This paper focuses on the illegitimate birth rate in Russia. The study of non-marital fertility phenomenon is of paramount significance for understanding and forecasting fertility trends as the main factor of natural reproduction of the population.
Political behaviour tends to be measured mainly at the national level. This paper analyses and compares the reported levels of discussion about “local politics” and “national politics” across Europe (n = 30216; 31 countries). A series of multilevel logit models are fitted to data from Eurobarometer 73.4 (2010), using covariates at the national and individual level. At the country level, a higher GDP and a more decentralized institutional structure are related with a higher likelihood of discussing local politics. At the individual level, the varying performance of indicators depending on the territorial level highlights the relevance of this topic.
The postponed marriages and childbirth delays could be explained the rise of job and income instability. The growth of flexible market relations increased uncertainty and job insecurity in many European countries during the last decades. This paper aims to disclose and explain the impact of job instability on fertility intentions taking into account the employment legislation of the country. The empirical analysis is grounded on the European Social Survey, 2010. Polarized dual labour market resulted from rigid employment legislation decreases fertility intentions while liberal labour markets contribute to childbirth planning
Surveys in developing countries pose a number of problems not present or not as important elsewhere. Among these are missing infrastructure and missing or poor base data. Besides these more or less expected problems an additional one are expectations based on western mental models that are not aligned with local reality
Based on field experience in developed and developing countries all over the world we describe specific challenges and solutions to the problems outlined above. To ease future work we outline lessons learned and requirements for better software support and examine currently available solutions as well as missing links.
According to well established researchers in the field of data fusion, ‘Statistical Matching’ should aim to preserve data distributions and multivariate relations between never jointly observed variables, however, it is not destined to reconstruct entire individual records. Although this approach of Statistical Matching is promising with regard to some particular applications, it, nevertheless, has its shortcomings, since on a micro level, data will not be approximated to their true but unknown values. The approach suggested here is to alleviate this deficit by auxiliary data,e.g. expert knowledge as external source, applicable in the case of fusing most similar cases.
This communication addresses problems arose and solutions adopted in sample design and sample selection for the first Portuguese National Health Examination Survey, a cross-sectional study that aims to obtain epidemiological data for the Portuguese population at regional and national level.
The sample was based on two-stage stratified cluster design, with selection of geographical areas in the first stage and individuals in the second stage.
Results suggest that sampling frame validation and use of multiple data sources in the survey planning stage are extremely important to minimize survey errors and selection bias.
Time is the basic resource most equally shared across populations. Time use research reveals the functioning of societies at the social equivalent of the atomic level. The Multinational Time Use Study, developed by the Centre for Time Use Research, archives 70 surveys spanning 1961-2013 from 22 countries. This poster details data restoration, adjusting data collected by new technologies for backwards comparability, identification of harmonisable demographic variables explaining differences in time use, and plans to increase accessibility of the data. MTUS informs surveys CTUR collects and advice CTUR provides. This poster explains how MTUS archiving contributes to time use research.
Statistics Korea conducts the Social Survey to provide information on subjective opinions related to people’s quality of life. It includes 5 topics in every year and covers a total of 10 topics. It is not possible to analysis of cross-tabulation across all topics. In this study, we construct a new dataset through the linking of records from the 2012-2013 Social Surveys. We describe the process of record linkage, assessment of the linkage method and the analytical results of the linked data. Two different linkage methods are considered: the deterministic linkage method and the probabilistic method.