literature_review_critical: Optional[bool] = Field(
    default=None,
    description="""
        Does the literature review include a **critical assessment** of **previous empirical studies** (studies based on data)?

        -----------------------------
        DECISION RULES
        -----------------------------

        STEP 1 – Identify if a literature review exists:
        - Look for an "Introduction", "Background", or "Literature Review" section.
        - If no such section exists OR no previous research is cited → return False.

        STEP 2 – Are empirical studies cited?
        - Empirical studies include surveys, experiments, interviews, observations, or other data-based research.
        - If only conceptual papers, frameworks, policy reports, or definitions are cited (with no reference to empirical data studies) → return False.

        STEP 3 – Look for explicit **evaluative commentary** linked to empirical studies:
        - Trigger phrases include:
            "however", "although", "few studies", "contradictory", "inconsistent", 
            "limited evidence", "mixed results", "scarce research", "gap in research", 
            "future research needed", "weakness of prior work", "methodological limitation".
        - The trigger phrase must clearly refer to an **empirical study or group of studies**, for example:
            - "Previous studies relied only on small samples"
            - "Findings have been inconsistent across contexts"
            - "Earlier research used only self-reported data"

        STEP 4 – Distinguish conceptual context vs empirical critique:
        - If the evaluative language is applied only to conceptual arguments, definitions, or general context (e.g., "There are many ways to define diversity", "sustainability is complex") → return False.
        - Only return True if there is at least one explicit judgment of **empirical research methods, findings, or gaps**.

        STEP 5 – Final decision:
        - Return True → if there is at least one explicit evaluative statement about empirical studies.
        - Return False → if prior work is only described or contextualized, or only conceptual/theoretical sources are reviewed, or no literature review is present.

        Do NOT infer critique based on tone or general importance statements. Only return True when evaluative language is explicitly present **and clearly linked to empirical research**.
    """,
    json_schema_extra={
        "examples": [
            (
                # Purely descriptive (False)
                """
                Active learning strategies have been widely used in engineering education. Prince (2004) defines active learning as instructional methods that engage students in the 
                learning process. Freeman et al. (2014) conducted a meta-analysis showing that active learning improves student performance in STEM fields. Studies by Froyd et al. (2013) 
                and Michael (2006) have also documented benefits such as increased engagement and deeper understanding. Several models have been proposed to implement active learning 
                effectively in large classes (Borrego et al., 2018).
                """,
                False
            ),
            (
                # Explicit critique linked to empirical research (True)
                """
                While active learning has been shown to improve outcomes in engineering education (Freeman et al., 2014), much of the evidence relies on quasi-experimental designs, which 
                limit causal claims. For example, Froyd et al. (2013) primarily report student self-reports, raising questions about measurement validity. Although Borrego et al. (2018) 
                propose scalable models, their generalizability across different institutional contexts remains untested. In contrast to these optimistic accounts, Prince (2004) cautions 
                that active learning requires significant instructor preparation and may lead to uneven student participation. Overall, the literature demonstrates benefits but also 
                highlights gaps in understanding how contextual factors influence effectiveness.
                """, 
                True
            ),
            (
                # Evaluative but only about concepts (False)
                """
                There is no single agreed definition of diversity in engineering education, and approaches vary widely between institutions. Diversity is inherently complex and multidimensional.
                """,
                False
            ),
            (
                # Trigger words but conceptual (False)
                """
                Although sustainability is often considered essential, there is little consensus on how it should be incorporated into curriculum models.
                """,
                False
            ),
            (
                # Explicit empirical gap (True)
                """
                Few studies have examined how peer instruction affects confidence in minority-serving institutions, and those that exist use small sample sizes (Smith et al., 2018; Lee, 2019). 
                Findings are inconsistent across contexts, suggesting that more diverse settings need to be investigated.
                """,
                True
            ),
        ],
        "disclaimer_message": "Literature review might be biased"
    }
)