!work in progress! TEAM MEMBERS:
Ekaterina Kartseva, Ehrenreich Leonie
BEFORE THE RESEARCH
Prepare for the research
Those can include:
⦿ Dominance of single-country studies
⦿ Lack of expansive cross-country investigations (Bene, 2021, p. 1711)
⦿ Informal structures
⦿ Body language
⦿ Ethos
⦿ Cross-check on the data obtained in interview
⦿ Process:
Structured → Unstructured / In-field setting complete participant → Non-participating OBSERVATION
Formal Observation:
⦿ Sense of setting
⦿ Rich description
⦿ Norms and values in cultures
(Simons, 2009, p.55)
Develop a codebook, […] where the cross-country reliability test show satisfying coefficients (as conducted in Tonnesen et al., 2023, p. 11)
⦿ Open questions, providing a frame (act as reminder of focus)
⦿ Purposive sampling (most often)
⦿ Define dependent variables (numbers of “angry” and “love” reactions from the automatedly collected meta, as done inTonnesen et al., 2023, p. 10).
LITERATURE REVIEW
Theoretical basis
⦿ 4 types of literature reviews:
1) Integrate what others have done / said
2) Criticising previous works
3) Buildong bridges between topics
4) Identify central issues in a field (Cooper, 2010, as cited in Cresswell, ch 2)
⦿ When using exhaust data: have a ground truth and thick description (can help the researcher understand the social phenomena at play) (Bjerre-Nielsen & Glavind, 2022, p. 3)
⦿ Distinguish between primary data (collected for the specific research question) and secondary data (collected for another purpose and subsequently made available to others to reuse and analyze) (Kitchin, 2014, as cited in Bjerre-Nielsen & Glavind, 2022, p. 2)
RESEARCH DESIGN
Practical advice
⦿ Open questions, providing a frame, act as reminder as reminder of focus (Simons, p.31)
Quantitative | Qualitative |
|
|
1) Experiments/quasi-experiments 2) Surveys 3) Causal modelling 4) Cost–benefit analysis
| Three qualitative methods often used to facilitate in-depth analysis and understanding:
|
MEASURES: Small-scale surveys, patterns of examination results, questionnaires, descriptive statistics, content analysis
| MEASURES: Critical incidents, open letters, discourse analysis, narratives, video analysis, photographs, log entries, artefacts, case studies, open-ended interviews, observations, document reviews |
|
|
(Simons, 2009, p.34f.) (Simons, 2009, p.55)
Mixed Method | Action oriented: |
|
|
1) Surveys
2) Interviews
3) Observations
4) Document analysis
5) Panel review | Qualitative methods /Stakeholder participation / Interpretative analysis
|
|
|
|
|
Choose the right mix:
– Ethnographic data + quantitative data and data science techniques – useful inputs that complement other parts of the qualitative analysis. In particular, they can be helpful in the early stages of fieldwork, as a guide to finding patterns that deserve further in-field exploration (Bjerre-Nielsen & Glavind, 2022, p. 4)
– Ethnographic data + models of specific outcomes (machine learning tools)
EXAMPLE: Bjerre-Nielsen et al. (2021) show predictions made from task-specific information (e.g. using high-school grades when predicting college grades) can outperform predictions made using ned grained individual big data. It is conceivable that ethnographers can supply such task-specific data and thereby enhance prediction. (Bjerre-Nielsen & Glavind, 2022, p. 4)
⦿ Consider using grounding frameworks:
⦿ Political / Cultural ideologies:
Authoritarianism / Libertanism / Communitarism / Individualism / Combination of the any of the above
⦿ Communicative approaches:
Transactional approach to the publuc / Dialogical approach to the public / Open systems approach to data work
⦿ Governing rationales:
Democratic values / Market capitalism / Privacy protection / Efficency / Pro-innovation
⦿ Governing actors:
Governmental actors / Non-governmental actors / Corporate actors
(table 1, Liu and Wang, 2022, p. 3) ⦿
⦿ What should be considered + subject to what sense of governance and regulation + who has the right to govern data work, and for whose interest (Liu and Wang, 2022, p. 4)
⦿ When applying quantitative tools to data from ethnographic fieldwork – ethnographic data needs to be structured to some degree and preferably digitized, depending on the format of the data and the application (Bjerre-Nielsen & Glavind, 2022, p. 4)
Toggle Cont
RESEARCH
Practical advice
⦿ Ensure homogeneity during the data sampling
HOW: setting limitations
EXAMPLE: “To ensure the relative homogeneity of the cross-country sample, the investigation is limited to Facebook, which is the dominant social media platform in each member state” (Newman et al., 2020; Teperik et al., 2018, as cited in Bene, 2021, p. 1712)
⦿ Consider country-speficic aspects
Specific domain of social life (identify pertinent behaviors / attitudes of the citizens to be regulated, as argued by H.-J. Andreß et al, p.17)
NB! The nation is seen as a context in which citizens are embedded. But, most of the time, nations are entities that are remote from the lives of their citizens, and they differ in many ways.
⦿ Consider how you are going to perform the reliability tests and tackle the language barrier
EXAMPLE: “Due to the national elections taking place at different times and the coding materials being in different languages, cross-country reliability tests Tonnesen et al., 2023, p. 11) could not perform these tests.
4 approaches (by H.-J. Andreß et al.):
1) Analyses of Aggregate Data (beware of the composition bias)
compare national aggregates (means, proportions, correlations, regression coefficients) across countries, mostly along a descriptive approach (p. 7)
2) Two-Step Analyses
Other studies go one step further, and use country estimates of means, proportions, correlations, or regression coefficients as dependent variables in regression models with country characteristics as explanatory variables (p.8)
3) Analyses of Disaggregated Data
Disaggregates the contextual information to the lower level and treats these macro- (and meso-)data as if they were microdata. (p.8)
4) Mixed Effects Analyses
Takes the hierarchical nature of the data into account, and estimates individual and contextual effects simultaneously (p.8)
⦿ Draft a discussion section, where you discuss the implications of the results
⦿ Individual-level results should be generalizable within countries. However, individual-level effects in multilevel models are not only identified by variation within countries, but also by between-country variation (see Bell et al. 2018)
⦿ To examine change, the analysis must be broadened by introducing a longitudinal design (Andreß et al., 2019, p.12)
⦿ 2 Weaknesses of CNCR:
1) It cannot determine the direction of causality inherent in the correlations. And as every country is observe
2) CNCR does not control for time-constant unobserved heterogeneity – may cause bias
(Andreß et al., 2019, p.13)
Other issues: data availability (Andreß et al., 2019, p.48) / the problem of having only a small number of higher-level units; and issues of omitted variable bias (Andreß et al., 2019, p.89) / self-selection or convenience sampling in the international surveys (Ebbinghaus 2005) / small number of included countries in the international surveys (Andreß et al., 2019, p.94) / planning fieldwork in a foreign country or experiencing a language barrier (Andreß et al., 2019, p.199
⦿ Certain cultures / countries are overrepresented in the social science research
CASE STUDY EXAMPLE
Qualitative design: ethnography & critical analysis
Part 1: two months, in-country fieldwork:
in-person interviews
participant observation
document analysis
students’ journals
Part 2:
two months of data collection
document analysis
cooperative journaling, phone interviews and Facebook conversations to address findings from in the field