Refine
Has Fulltext
- no (12)
Document Type
- Article (12) (remove)
Is part of the Bibliography
- yes (12)
Keywords
- Bildungsexpansion (1)
- Bourdieu (1)
- COVID-19 (1)
- Causal Inference (1)
- Custom Online Panel (1)
- Descriptive Inference (1)
- Directed Acyclic Graph (DAG) (1)
- Distinction (1)
- Distinktion (1)
- Educational Expansion (1)
Institute
- Fachgruppe Soziologie (12) (remove)
Since COVID-19 became a pandemic, many studies are being conducted to get a better understanding of the disease itself and its spread. One crucial indicator is the prevalence of SARS-CoV-2 infections. Since this measure is an important foundation for political decisions, its estimate must be reliable and unbiased. This paper presents reasons for biases in prevalence estimates due to unit nonresponse in typical studies. Since it is difficult to avoid bias in situations with mostly unknown nonresponse mechanisms, we propose the maximum amount of bias as one measure to assess the uncertainty due to nonresponse. An interactive web application is presented that calculates the limits of such a conservative unit nonresponse confidence interval (CUNCI).
Im Vergleich zu Umfragen an Wahrscheinlichkeitsstichproben bieten Umfragen an Access-Panels, die auf Nicht-Wahrscheinlichkeitsstichproben basieren, unbestreitbare wirtschaftliche Vorteile. Diese Vorteile gehen jedoch mit unvermeidbaren Qualitätseinbußen einher, die auch dann bestehen bleiben, wenn Erstere sehr niedrige Responseraten haben. Daher müssen die wirtschaftlichen Vorteile und die methodischen Einschränkungen gegeneinander abgewogen werden. Es wird argumentiert, dass diese Abwägung anhand normativer Festlegungen erfolgen muss. Unter Anwendung der hier vorgeschlagenen Maßstäbe kommt der Beitrag zu dem Schluss, dass die Qualitätsansprüche an über Massenmedien verbreitete Meinungsumfragen höher sein sollten als für rein (sozial)wissenschaftliche Zwecke.
Pulp Science?
(2023)
A review of all research papers published in the European Sociological Review in 2016 and 2017 (N = 118) shows that only a minority of papers clearly define the parameter of interest and provide sufficient reasoning for the selected control variables of the statistical analysis. Thus, the vast majority of papers does not reach minimal standards for the selection of control variables. Consequently, a majority of papers interpret biased coefficients, or statistics without proper sociological meaning. We postulate that authors and reviewers should be more careful about control variable selection. We propose graphical causal models in the form of directed acyclic graphs as an example for a parsimonious and powerful means to that end.
The long term relationship between medicaid expansion and adult life-threatening chronic conditions
(2023)
We test whether the expansions of children's Medicaid eligibility in the 1980s–1990s resulted in long-term health benefits in terms of severe chronic conditions. Still relatively rare in the field, we use prospective individual-level panel data from the Panel Study of Income Dynamics (PSID) along with the higher quality income measures from the Cross-National Equivalent File (adjusting for taxes, transfers and household size). We observe severe chronic conditions (high blood pressure/heart disease, cancer, diabetes, or lung disease) at ages 30–56 (average age 43.1) for 4670 respondents who were also prospectively observed during childhood (i.e., at ages 0–17). Our analysis exploits within-region temporal variation in childhood Medicaid eligibility and adjusts for state- and individual-level controls. We uniquely concentrate attention on adjusting for childhood income. A standard deviation greater childhood Medicaid eligibility significantly reduces the probability of severe chronic conditions in adulthood by 0.05 to 0.12 (16%–37.5% reduction from mean 0.32). Across the range of observed childhood Medicaid eligibility, the probability is approximately cut in half. Greater childhood Medicaid eligibility also substantially reduces childhood income disparities in severe chronic conditions. At higher levels of childhood Medicaid eligibility, we find no significant childhood income disparities in adult severe chronic conditions.
This paper compares the usability of data stemming from probability sampling with data stemming from nonprobability sampling. It develops six research scenarios that differ in their research goals and assumptions about the data generating process. It is shown that inferences from data stemming from nonprobability sampling implies demanding assumptions on the homogeneity of the units being studied. Researchers who are not willing to pose these assumptions are generally better off using data from probability sampling, regardless of the amount of nonresponse. However, even in cases when data from probability sampling is clearly advertised, data stemming from nonprobability sampling may contribute to the cumulative scientific endeavour of pinpointing a plausible interval for the parameter of interest.
Was sollten Mitarbeiter in einem empirisch ausgerichteten Forschungsprojekt können, und welche Kernkompetenzen sollte die Ausbildung an den Universitäten daher vermitteln? Die Antworten auf diese Fragen hängen – wie sollte es anders sein – von der inhaltlichen Fragestellung und methodischen Ausrichtung des jeweiligen Forschungsprojektes ab. Natürlich sollten Projektmitarbeiter über Vorkenntnisse zum Forschungsthema verfügen. Natürlich sollten Kenntnisse des projektspezifischen (statistischen) Methodenarsenals vorliegen.
Emerging evidence has highlighted the important role of local contexts for integration trajectories of asylum seekers and refugees. Germany's policy of randomly allocating asylum seekers across Germany may advantage some and disadvantage others in terms of opportunities for equal participation in society. This study explores the question whether asylum seekers that have been allocated to rural areas experience disadvantages in terms of language acquisition compared to those allocated to urban areas. We derive testable assumptions using a Directed Acyclic Graph (DAG) which are then tested using large-N survey data (IAB-BAMF-SOEP refugee survey). We find that living in a rural area has no negative total effect on language skills. Further the findings suggest that the "null effect" is the result of two processes which offset each other: while asylum seekers in rural areas have slightly lower access for formal, federally organized language courses, they have more regular exposure to German speakers.
This article draws on the experience from an ongoing research project employing respondent-driven sampling (RDS) to survey (illicit) 24-hour home care workers. We highlight issues around the preparatory work and the fielding of the survey to provide researchers with useful insights on how to implement RDS when surveying populations for which the method has not yet been used. We conclude the article with ethical considerations that occur when employing RDS.
In times of educational expansion, privileged families are looking for new strategies of distinction. Referring to Pierre Bourdieu’s theory of distinction, we argue that choosing Latin at school – a language that is no longer spoken and therefore has no direct value – is one of the strategies of privileged families to set themselves apart from less privileged families. Based on two surveys we conducted at German schools, the paper analyzes the relationship between parents’ educational background and the probability that their child will learn Latin. Results indicate that historically academic families have the strongest tendency towards learning Latin, followed by new academic families, and leaving behind the non-academic families. We distinguish between four causal mechanisms that might help to explain these associations: cultural distinction, selecting a socially exclusive learning environment, beliefs in a secondary instrumental function of learning Latin, and spatial proximity between the location of humanist Gymnasiums and the residential areas of privileged families. The hypotheses are formalized by means of Directed Acyclic Graphs (DAG). Findings show that the decision to learn Latin is predominately an unintended consequence of the selection of a socially exclusive learning environment. In addition, there is evidence that especially children from historically academic families learn Latin as a strategy of cultural distinction.