# Exports Guide (ibe-pp)

This note explains the key fields you’ll see in CSV/Excel exports from the demo sessions and how to interpret them in class.

## Common fields
- session_code, participant_code: Internal identifiers for linking records.
- app_name, round_number, page_name: Where a row came from.
- payoff: Participant payoff in the app (points or currency units).

## Classic/Policy Games
- Prisoner’s Dilemma: `cooperate` (True/False), `payoff`.
- Trust (simple/framed): `sent_amount`, `sent_back_amount`, `payoff`.
- Dictator (norms): `kept` (amount P1 kept), `payoff`.
- Public Goods (base/defaults): `contribution`, group totals (if included), `payoff`.

Interpretation tips
- Trust: reciprocity shows in `sent_back_amount` relative to `sent_amount`×multiplier.
- Public Goods: highlight total vs. private trade-off driving `payoff`.
- Dictator: distribution between P1 and P2 indicates fairness preferences; note framing.

## Survey Biases (policy_survey_biases)
- Overconfidence (90% CI): `oc1_low`, `oc1_high`, `oc2_low`, `oc2_high`, `oc3_low`, `oc3_high`.
  - In class: discuss calibration by checking how often intervals would contain the true value (avoid revealing “correct” answers live).
- Overplacement: `self_rank_policy_knowledge` (1–7).
  - In class: look at distribution vs. mid-point 4.
- Confirmation: `agree_statement_before`, `agree_statement_after` (1–7).
  - In class: look at whether responses moved after the snippet; both change and no-change are informative.
- Availability: `recall_event` (free text), `priority_energy`, `priority_public_health` (1–7).
  - In class: discuss how salient events might shift priorities.

## Exporting with bots (optional)
- Use Make targets to run bots and then export from the admin. For quick bot runs:
  - `make otree-test-classic`
  - `make otree-test-policy`
  - `make otree-test-anchoring`
  - `make otree-test-survey`

## Notes
- Avoid showing “correct” answers for overconfidence during class; focus on calibration concepts.
- Frame findings as illustrations; behavior varies by group, context, and incentives.