Add About section to methodology report addressing prompt hacking

- Reference Kosch & Feger (2025) "Prompt-Hacking: The New p-Hacking?"
- Explain how CatLLM uses standardized prompts for reproducibility
- Addresses concern that prompt variability undermines replicability

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

app.py

@@ -83,11 +83,22 @@ def generate_methodology_report_pdf(categories, model, column_name, num_rows, mo
 
     story = []
 
-    # === PAGE 1: Title, Category Mapping ===
+    # === PAGE 1: Title, About, Category Mapping ===
     story.append(Paragraph("CatLLM Methodology Report", title_style))
     story.append(Paragraph(f"Generated: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}", normal_style))
     story.append(Spacer(1, 15))
 
+    # About CatLLM - addressing prompt hacking
+    story.append(Paragraph("About This Report", heading_style))
+    about_text = """This methodology report documents the classification process for reproducibility and transparency. \
+CatLLM addresses an issue identified by researchers in "Prompt-Hacking: The New p-Hacking?" (Kosch & Feger, 2025): \
+researchers could keep modifying prompts to obtain outputs that support desired conclusions, and this variability \
+in pseudo-natural language poses a challenge for reproducibility since each prompt, even if only slightly altered, \
+can yield different outputs, making it impossible to replicate findings reliably. CatLLM restricts the prompt to a \
+standard template that is impartial to the researcher's hypothesis or inclinations, ensuring consistent and reproducible results."""
+    story.append(Paragraph(about_text, normal_style))
+    story.append(Spacer(1, 15))
+
     # Category mapping
     story.append(Paragraph("Category Mapping", heading_style))
     story.append(Paragraph("Each category column contains binary values: 1 = present, 0 = not present", normal_style))