The goal of the final project is for students to demonstrate their ability to:

• Formulate a meaningful economic research problem realted to consumer behavior.
• Collect data (secondary data or own survey data).
• Apply appropriate econometric models.
• Interpret results within the economic literature context.
• Critically discuss implications and limitations.

Students are encouraged to focus on a real-world policy-relevant issue, with potential contribution to the literature:

• For BA students: a “contribution” may consist of applying established economic models to a specific dataset, context, or policy setting, and demonstrating correct empirical reasoning. Novel theoretical or methodological contributions are not required.
• MA students are encouraged to articulate more clearly how their study extends, refines, or challenges existing empirical evidence, even if the contribution is incremental.

The remainder of the report should contain the following sections, or contain the following content.

1. Introduction

• Motivation and background of the research problem.
• Clearly state the research question and its relevance.
• Contextualize the research problem within the related economic literature.
• Brief overview of methodology and data.

The introduction should motivate the research question and outline the approach. It should not present detailed results or policy conclusions.

2. Literature Review

The literature review must serve a clear economic purpose: it should justify the model structure, variable choices, and expected relationships.

• BA: the literature review must be grounded primarily in microeconomic theory and applied microeconometric studies. Managerial, psychological, or purely descriptive frameworks are not sufficient unless explicitly linked to an economic model.
• MA: interdisciplinary literature may be used, but the economic mechanism must remain central and explicit.

The literature review provides the conceptual and empirical rationale for the model structure presented later in the methodology section.

• Discuss relevant theoretical and empirical studies related to your research problem. What economic relationships have others studied? What models and variables have been used in similar contexts?
• Identify the economic mechanisms or conceptual frameworks that justify the expected relationships.
• Expected coefficient signs should be derived from theory or prior evidence. These are benchmarks, not targets. Empirical results that differ from expectations must be discussed rather than adjusted or ignored.
• Use empirical literature to justify your choice of variables, models, and expected coefficient signs.
• Explain how your study differs from or contributes to existing work – by using different data, a new context, or updated methodology.

3. Data Description

Students must clearly state the scope and limitations of their data. Claims about populations, policy relevance, or external validity must not exceed what the data can support.

• Source and structure of the data.
• Summary statistics (by cross-section and time if applicable).
• Scatter plots, if necessary.
• Data cleaning steps, if any.

Additional information:

• BA: simple descriptive statistics and basic data checks are sufficient.
• MA: More careful discussion of data construction, missing values, and measurement issues is expected.

4. Methodology

The methodology section translates the conceptual framework and empirical justification from the literature review into a concrete, testable model and explains how it is estimated using the panel dataset. This section should demonstrate your understanding of the econometric requirements of your chosen model and ensure the reliability of your findings.

• Present the econometric model clearly using one or more equations. Define all variables.
• Justify the model specification. Explain the inclusion of each variable based on economic theory and prior empirical studies (link to your literature review). Justify any interaction terms or nonlinear specifications (logarithm or squared).
• Explain the estimation method and model choice.
• Present how you will use relevant diagnostic tests, as applicable:
• Multicollinearity check (e.g., variance inflation factors).
• Diagnostic tests after IV regression.
• Discuss how you address any identified issues (e.g., clustering standard errors, using robust estimators, transforming variables, or refining instruments).

The choice of model should reflect the research question and data structure. More complex models do not automatically imply higher quality.

• BA: standard models (OLS, log-linear models, limited dependent variable models) are sufficient if correctly specified and interpreted.
• MA: More advanced methods (IV, panel models, nonlinear estimators) may be appropriate when clearly motivated and correctly implemented.

Instrumental variable methods and advanced diagnostics should only be used when the source of endogeneity is clearly explained and instruments are theoretically justified. Incorrect or superficial use of advanced techniques will be penalized more heavily than their omission.

When endogeneity is suspected but cannot be credibly addressed, students are expected to acknowledge and discuss it explicitly rather than ignore it. This discussion should identify which variables are potentially endogenous, explain the economic reasons for endogeneity (e.g., omitted variables, reverse causality, measurement error), and outline what types of instruments or research designs could address the issue in principle, even if they are not available in the current dataset.

5. Empirical results

Present regression outputs in well-formatted and numbered tables/figures. Each table/figure must:

• Have a clear title and number (e.g., Table 2: FE estimates of wage function),
• Be referred to and explained in the main text (do not insert tables without interpretation or discussion).
• Be free from redundant footnotes such as “Author’s calculation” or “Compiled by the authors” – these are unnecessary.

Interpret the estimated coefficients, focusing on:

• Sign, magnitude, and statistical significance,
• Whether the results align with theoretical expectations or prior literature.

Explain and interpret results from diagnostic tests:

• Discuss evidence of autocorrelation, heteroskedasticity, endogeneity, or instrument weakness (if applicable),
• Explain how these issues affect your model’s reliability and what adjustments have been made.

Present robustness checks to strengthen credibility:

• Alternative model specifications,
• Subsample analysis,
• Use of lagged variables or transformations.

Presentation standards for tables and figures:

• Every table and figure must be titled, numbered, and explicitly discussed in the text. Avoid placing tables without explanation.
• Use clear, descriptive variable names in tables and figures. Avoid coded or symbolic variable labels (e.g., “hhinc” or “gtinh”). Instead, write:
  • Household income ($/month),
  • Gender (1 = Male),
  • Education (years of schooling),
  • Region (1 = urban)
• For dummy variables generated from categorical variables (e.g., region, occupation, education level), always state the omitted (base) category used in the regression.

6. Discussion

The discussion section should interpret the empirical findings in a broader economic context. This section is not a repetition of the results, but an analytical reflection on what the results mean, how they relate to existing knowledge, and why they matter. Key components include:

• Interpret your findings in relation to your research question and economic theory.
• Compare your results to existing studies. Are they consistent with previous evidence? Do they offer a new angle or contradict established findings?
• Reflect on the limitations of your analysis and how these might affect interpretation.

Empirical findings should be used to draw broader insights about real-world economic issues. Policy discussion should be grounded in the results and framed cautiously. Students are not required to propose new policies.

Policy implication may take one or more of the following forms:

• Identifying policy-relevant patterns: highlight economic patterns or relationships that deserve attention (e.g., persistent consumption inequality, heterogeneous responses across income groups, barriers to access).
• BA: explaining why the pattern matters for policy debates is sufficient.
• MA: students should link these patterns more explicitly to policy objectives or constraints.
• Evaluating existing policies or assumptions: discuss whether the results support or challenge the assumptions underlying current policies, or whether existing interventions appear less effective than intended.
• Informing policy trade-offs: use the findings to clarify trade-offs faced by policymakers, such as efficiency versus equity, short-run versus long-run effects, or universal versus targeted interventions.
• Conditional policy suggestions (optional): when appropriate, students may suggest policy initiatives, programs, or reforms conditional on the empirical evidence and stated assumptions.
• BA: explicit policy recommendations are optional and will not be rewarded unless clearly supported by the analysis.
• MA: policy suggestions should be carefully justified, conditional, and linked explicitly to the empirical findings.
• Contribution to public debate: explain how the results inform public discussions by highlighting distributional consequences or identifying groups that benefit or are disadvantaged.

Important notes:

• Avoid strong normative claims or prescriptive language (“should”, “must”) unless the analysis clearly supports such conclusions.
• Policy implications should follow logically from the results; proposals unrelated to the empirical analysis will be penalized.

7. Conclusion

The conclusion should summarize findings without introducing new results, models, or data.

• Recap main findings and implications.
• Briefly mention unanswered questions or limitations.

8. References

• Use a consistent and recognized citation format, such as APA or Chicago author-date style.
•  Include DOI links for all academic journal articles and book chapters, where available. This ensures your references are verifiable and professionally formatted. If a DOI is not available, a stable link to the articles is acceptable.
• All in-text citations must correspond to full references in the bibliography section.

APA Style:

Acemoglu, D., & Autor, D. (2011). Skills, tasks and technologies: Implications for employment and earnings. In O. Ashenfelter & D. Card (Eds.), Handbook of labor economics (Vol. 4, pp. 1043–1171). Elsevier. https://doi.org/10.1016/S0169-7218(11)02410-5
Angrist, J. D., & Krueger, A. B. (1991). Does compulsory school attendance affect schooling and earnings? Quarterly Journal of Economics, 106(4), 979–1014. https://doi.org/10.2307/2937954

Chicago (Author-Date) Style:

Acemoglu, Daron, and David Autor. 2011. “Skills, Tasks and Technologies: Implications for Employment and Earnings.” In Handbook of Labor Economics, Vol. 4, edited by Orley Ashenfelter and David Card, 1043–1171. Amsterdam: Elsevier. https://doi.org/10.1016/S0169-7218(11)02410-5
Angrist, Joshua D., and Alan B. Krueger. 1991. “Does Compulsory School Attendance Affect Schooling and Earnings?” Quarterly Journal of Economics 106 (4): 979–1014. https://doi.org/10.2307/2937954

Final project evaluation criteria (Total: 100 points)

Non-compensation is a global rule. For example, serious econometric errors (mis-specification, incorrect interpretation, invalid methods) will lead to score reductions across multiple criteria and cannot be compensated by strong writing or literature review alone.

1. Research question & motivation (10 points)

What is assessed

• Clarity and focus of the research question
• Economic relevance to consumer behavior
• Logical motivation grounded in real-world economic issues

Score anchors

• 8–10: Clear, focused research question with strong economic motivation and relevance
• 5–7: Reasonable question but vague, overly broad, or weakly motivated
• 0–4: Unclear or descriptive topic without a well-defined economic question

Warning: a topic description or data-driven curiosity without a clear economic question will not receive a high score.

2. Literature review & theoretical justification (20 points)

What is assessed

• Use of relevant economic theory and empirical literature
• Ability to justify model structure, variable choice, and expected relationships
• Integration of literature (not listing or summarizing papers)

Score anchors

• 16–20: Literature clearly motivates the empirical model; theory and evidence are well integrated
• 10–15: Relevant literature cited but linkage to model or theory is weak or incomplete
• 0–9: Largely descriptive, managerial, or disconnected from the empirical analysis

Warnings:

• A long literature review does not guarantee a high score if it does not inform the model.
• Non-economic frameworks (e.g., managerial or purely psychological models) must be explicitly linked to economic mechanisms.

3. Methodology: model specification & estimation (20 points)

What is assessed

• Appropriateness of the econometric model for the research question
• Correct specification and presentation of equations
• Justification of variables, functional forms, and estimation method
• Appropriate (not excessive) use of diagnostic tests

Score anchors

• 16–20: Model is well specified, theoretically justified, correctly estimated, and clearly explained
• 10–15: Model broadly appropriate but with gaps in justification or minor technical errors
• 0–9: Mis-specified model, misunderstanding of econometric concepts, or unjustified complexity

Warnings:

• Advanced methods are not rewarded per se.
• Incorrect use of IV, nonlinear models, or diagnostics will be penalized more heavily than omission.
• Cosmetic sophistication leads to score reduction.

4. Data description & workflow (10 points)

What is assessed

• Clarity of data sources and structure
• Summary statistics and basic exploratory analysis
• Transparency and replicability of data processing and code

Score anchors

• 8–10: Data clearly described; workflow is transparent and replicable
• 5–7: Basic description provided but missing clarity or organization
• 0–4: Poor data documentation or irreproducible analysis

Warning: claims about population relevance or policy scope must not exceed what the data can support.

5. Empirical results: interpretation & robustness (15 points)

What is assessed

• Correct interpretation of coefficients (sign, magnitude, statistical significance)
• Distinction between statistical and economic significance
• Use and interpretation of robustness checks where appropriate

Score anchors

• 12–15: Accurate, economically meaningful interpretation; robustness clearly discussed
• 7–11: Interpretation mostly correct but mechanical or incomplete
• 0–6: Misinterpretation of coefficients, tests, or significance

Warnings:

• Confusing “significant” with “important” is penalized.
• Tables without interpretation do not earn points.
• Robustness checks must be explained, not just reported.

6. Discussion & policy implications (15 points)

What is assessed

• Integration of findings with economic theory and literature
• Ability to reflect on limitations and external validity
• Quality of policy-relevant reasoning (not policy prescription)

Score anchors

• 12–15: Clear, well-reasoned discussion linking results to theory and policy relevance with appropriate caution
• 7–11: Discussion present but generic, descriptive, or weakly connected to results
• 0–6: Over-claiming, unsupported policy recommendations, or repetition of results

Warnings:

• Policy implications is not policy design.
• Strong normative language (“should”, “must”) without causal justification will be penalized.
• Proposals unrelated to the empirical findings receive no credit.

7. Conclusion (5 points)

What is assessed

• Concise summary of key findings
• Clear acknowledgment of limitations and future directions

Score anchors

• 4–5: Clear, accurate, and focused conclusion
• 2–3: Summary present but vague or repetitive
• 0–1: Incoherent or misleading conclusion

Warning: No new results, models, or claims should appear in the conclusion.

8. Structure, writing & presentation (5 points)

What is assessed

• Logical organization and flow
• Academic writing quality
• Presentation of tables, figures, and citations

Score anchors

• 4–5: Clear, professional, and well structured
• 2–3: Understandable but poorly organized or formatted
• 0–1: Disorganized, unclear, or careless presentation

Warning: poor presentation can reduce the credibility of otherwise correct analysis.