Gallup Most Important Problem: Open-Ended Answer Coding with the Policy Agendas Codebook
Deductive coding of open-ended answers to Gallup's "most important problem" question into the Policy Agendas Project major topics, with a fixed codebook, two independent coders per answer, Krippendorff's alpha on the admin dashboard, private memos, and adjudication of disagreements. The Policy Agendas data are yearly topic shares and carry no individual answers, so the ten sample answers are synthetic.
Try it live — no install
Boot the real Potato server in your browser (WebAssembly) and annotate with this exact config. Nothing leaves your machine.
Run live in your browserGet This Design
Clone or download from the repository
Quick start:
git clone https://github.com/davidjurgens/potato-showcase.git
cd potato-showcase/text/computational-social-science/gallup-most-important-problem-coding
potato start config.yamlAbout this dataset
The Policy Agendas Project's Most Important Problem dataset codes the responses reported in Gallup's Most Important Problem polls by policy topic. The current release on the Comparative Agendas Project site holds 1,575 observations spanning 1947 to 2023, aggregated to the yearly share of responses in each major topic.
According to the dataset's codebook, the project used the standard Policy Agendas major topics plus major topic 25 for answers of "don't know" or an "other" topic, and assigned each entry one and only one content code. The topic numbers and names in the config are the major topics of the Comparative Agendas Project Master Topics Codebook (version 1.1, August 2022), and each code's tooltip lists the subtopics that topic covers.
The design applies the same code frame to individual answers, coded by two people working independently. Once both coders finish the shared answers, Potato reports Krippendorff's alpha for the topic scheme on the admin dashboard, and answers the coders disagree on go to an adjudicator, who must write a note when overriding a coder. The codebook is fixed, so neither coder can add a code partway through, and memos are private until reliability has been computed. The Policy Agendas codebook describes no two-coder procedure, so that part comes from this design.
The ten synthetic answers sit on boundaries the master codebook settles. Inflation and the unemployment rate fall under Macroeconomics rather than Labor, gas prices under Energy, climate change under Environment, border enforcement agencies under Law and Crime while immigration itself has its own topic, drug trafficking under Law and Crime while drug abuse treatment falls under Health, and homelessness under Housing. Wlezien (2005) discusses what answers to the most important problem question measure.
Comparative Agendas Project datasets and codebooks are released under CC BY-NC-SA 4.0. The Policy Agendas data carry no individual answers, so every sample answer is synthetic.
- Observations
- 1,575
- Years covered
- 1947 to 2023
- Codes
- 21 major topics plus topic 25 (don't know / other)
- Codes per entry
- One
- Licence
- CC BY-NC-SA 4.0
Configuration File
This Potato config reproduces the annotation task. Save it as config.yaml and run potato start config.yaml to try it.
# Gallup "Most Important Problem": Coding Open-Ended Answers with the
# Policy Agendas Project Codebook, Two Coders
#
# Dataset and code frame:
# Bryan D. Jones, Frank R. Baumgartner, Sean M. Theriault, Derek A. Epp,
# Shruti Khandekar, and Daniel Little (2025). "Policy Agendas Project: Most
# Important Problem." Comparative Agendas Project. 1,575 observations, 1947
# to 2023. https://www.comparativeagendas.net/datasets_codebooks
# Codebook: "Gallup's Most Important Problem Data Codebook", Policy Agendas
# Project (2015).
# Topic system: Comparative Agendas Project Master Topics Codebook, version
# 1.1 (August 9, 2022). https://www.comparativeagendas.net/pages/master-codebook
# Background on the measure:
# Christopher Wlezien (2005). "On the salience of political issues: The
# problem with 'most important problem'." Electoral Studies 24(4): 555-579.
# https://doi.org/10.1016/j.electstud.2005.01.009
#
# What the Policy Agendas Project did (from its MIP codebook): it coded the
# responses reported in Gallup's Most Important Problem polls by policy
# content, using the standard Policy Agendas major topics "with the addition
# of a category for those answering 'don't know' or an 'other' topic: major
# topic 25", and "each entry was assigned one and only one content code".
# The published dataset is aggregated to yearly proportions per major topic.
#
# What this design adds: the same code frame applied to individual open-ended
# answers by two independent coders, so intercoder reliability can be reported
# before the codes are used. Krippendorff's alpha for each scheme appears on
# the admin dashboard once both coders have finished the shared items, and
# disagreements go to an adjudication queue. The two-coder procedure is this
# design's, not something the Policy Agendas codebook describes.
#
# Licence: Comparative Agendas Project datasets and codebooks are
# (c) Comparative Agendas Project, released under CC BY-NC-SA 4.0. The topic
# numbers and names below are the CAP major topics; each tooltip lists the
# subtopics that topic contains. Read the master codebook's subtopic
# descriptions before coding: they settle most boundary cases.
#
# Reproduction notes:
# - The Policy Agendas data are yearly topic shares and carry no individual
# answers, so every answer in sample-data.json is SYNTHETIC. The ten
# answers are chosen to land on documented topic boundaries:
# * inflation, prices, and cost of living are subtopic 101, and the
# unemployment rate is 103, both under 1 Macroeconomics, not 5 Labor;
# * oil and gas prices are subtopic 803, under 8 Energy;
# * climate change is subtopic 705, under 7 Environment;
# * border enforcement agencies are subtopic 1201 (12 Law and Crime),
# while immigration and refugees are 900 (9 Immigration);
# * illegal drug crime and trafficking are 1203 (12 Law and Crime),
# while drug and alcohol abuse treatment is 342 (3 Health);
# * homelessness is subtopic 1409, under 14 Housing.
# - One code per answer, as in the PAP data. When an answer names two
# problems, code the one named first and write a memo; this rule is this
# design's.
# - The codebook is fixed: coders cannot add or change codes during coding.
annotation_task_name: "Gallup Most Important Problem: Policy Agendas Topic Coding"
task_dir: "."
data_files:
- sample-data.json
item_properties:
id_key: "id"
text_key: "text"
output_annotation_dir: "annotation_output/"
export_annotation_format: "jsonl"
port: 8000
# A deductive code frame: the codebook page shows each code's definition,
# and codes change only through the config or the command line.
codebook_mode: fixed
# Coders keep memos on hard cases. Private memos keep the two coders from
# seeing each other's reasoning before reliability is computed.
annotation_ui:
memos: true
visibility: private
# Every answer is coded by two people.
num_annotators_per_item: 2
agreement_metrics:
enabled: true
min_overlap: 2
# Answers the two coders disagree on go to an adjudicator, who records why.
adjudication:
enabled: true
adjudicator_users:
- adjudicator
min_annotations: 2
show_annotator_names: false
require_notes_on_override: true
annotation_instructions: |
Each item is one answer to Gallup's question about the most important
problem facing the country. Assign it to exactly one Policy Agendas Project
major topic.
1. Code what the answer names, not what you think caused it. "Gas prices"
is Energy, even if the respondent is worried about inflation in general.
2. Hover over a topic to see the subtopics it covers. When two topics both
seem to fit, find the subtopic that names the issue and use its topic.
3. Use 25 for "don't know", for no answer, and for answers that fit no
major topic.
4. If an answer names more than one problem, code the first one named and
write a memo listing the others.
5. Code on your own. Do not discuss items with the other coder until
reliability has been computed; disagreements are settled in adjudication.
instance_display:
fields:
- key: text
type: text
label: "Answer to the most important problem question"
annotation_schemes:
- annotation_type: radio
name: major_topic
description: "Which Policy Agendas Project major topic does this answer name?"
codebook: true
labels:
- name: "1 Macroeconomics"
tooltip: "Interest rates, inflation, cost of living and prices (101); unemployment rate (103); monetary policy; national budget and debt; tax code; industrial policy; price control."
- name: "2 Civil Rights"
tooltip: "Minority, gender, age, and disability discrimination; voting rights; freedom of speech; right to privacy; anti-government activity."
- name: "3 Health"
tooltip: "Health care reform; health insurance; drug industry; medical facilities; providers; medical liability; disease prevention; mental health; long-term care; drug costs; tobacco, drug, and alcohol abuse (342)."
- name: "4 Agriculture"
tooltip: "Agricultural trade; subsidies to farmers; food inspection and safety; food marketing; animal and crop disease; fisheries."
- name: "5 Labor"
tooltip: "Worker safety; employment training; employee benefits and pensions; labor unions; minimum wage and fair labor standards; youth employment; migrant and seasonal workers."
- name: "6 Education"
tooltip: "Higher education; elementary and secondary schools; education of underprivileged students; vocational and special education; educational excellence."
- name: "7 Environment"
tooltip: "Drinking water; waste disposal; hazardous waste; air pollution and climate change (705); recycling; indoor hazards; species and forest protection; land and water conservation."
- name: "8 Energy"
tooltip: "Nuclear energy; electricity; natural gas and oil, including oil and gas prices and shortages (803); coal; alternative and renewable energy; energy conservation."
- name: "9 Immigration"
tooltip: "Immigration, refugees, and citizenship (900)."
- name: "10 Transportation"
tooltip: "Mass transit; highways; air travel; railroads; maritime transport; infrastructure."
- name: "12 Law and Crime"
tooltip: "Law enforcement agencies, including border agencies (1201); white collar crime; illegal drug crime and trafficking (1203); courts; prisons; juvenile crime; child abuse; family issues and domestic violence; criminal code; crime control; police."
- name: "13 Social Welfare"
tooltip: "Low-income assistance; elderly assistance and government pensions; disability assistance; volunteer associations; child care."
- name: "14 Housing"
tooltip: "Community and urban development; rural housing and development; low-income housing; veterans' and elderly housing; homelessness (1409)."
- name: "15 Domestic Commerce"
tooltip: "Banking; securities; consumer finance; insurance regulation; bankruptcy; corporate management; small business; patents and copyrights; disaster relief; tourism; consumer safety; sports regulation."
- name: "16 Defense"
tooltip: "Alliances; intelligence; military readiness; nuclear arms; military aid; personnel; procurement; installations; reserve forces; foreign military operations."
- name: "17 Technology"
tooltip: "Space; science transfer; telecommunications; broadcasting; weather forecasting; computers."
- name: "18 Foreign Trade"
tooltip: "Trade agreements; exports; private investment; competitiveness; tariffs and imports; exchange rates."
- name: "19 International Affairs"
tooltip: "Foreign aid; resource exploitation; developing countries; international finance; specific regions and countries; human rights; international organizations; terrorism; diplomats."
- name: "20 Government Operations"
tooltip: "Intergovernmental relations; bureaucracy; postal service; government employees and appointments; currency; procurement; tax administration; scandals; branch relations; political campaigns; census."
- name: "21 Public Lands"
tooltip: "National parks; indigenous affairs; public lands; water resources; dependencies and territories."
- name: "23 Culture"
tooltip: "General cultural policy."
- name: "25 Don't know / Other"
tooltip: "Don't know, no answer, or a problem outside every major topic. Major topic 25 in the Policy Agendas MIP data."
user_config:
allow_all_users: true
Sample Data
[
{
"id": "mip_001",
"text": "Inflation. Groceries cost twice what they did a few years ago.",
"synthetic": true
},
{
"id": "mip_002",
"text": "Unemployment. People can't find decent jobs around here.",
"synthetic": true
}
]
// ... and 8 more itemsTry it live — no install
Boot the real Potato server in your browser (WebAssembly) and annotate with this exact config. Nothing leaves your machine.
Run live in your browserGet This Design
Clone or download from the repository
Quick start:
git clone https://github.com/davidjurgens/potato-showcase.git
cd potato-showcase/text/computational-social-science/gallup-most-important-problem-coding
potato start config.yamlDataset & paper
Christopher Wlezien. 2005. On the salience of political issues: The problem with 'most important problem'. Electoral Studies 24(4): 555-579.
Citation (BibTeX)
@misc{jones2025pap_mip,
title = {Policy Agendas Project: Most Important Problem},
author = {Jones, Bryan D. and Baumgartner, Frank R. and Theriault, Sean M. and Epp, Derek A. and Khandekar, Shruti and Little, Daniel},
year = {2025},
howpublished = {Comparative Agendas Project},
url = {https://www.comparativeagendas.net/datasets_codebooks}
}
@article{wlezien2005salience,
title = {On the salience of political issues: The problem with `most important problem'},
author = {Wlezien, Christopher},
journal = {Electoral Studies},
volume = {24},
number = {4},
pages = {555--579},
year = {2005},
doi = {10.1016/j.electstud.2005.01.009}
}Details
Annotation Types
Domain
Use Cases
Tags
Found an issue or want to improve this design?
Open an Issue(opens in a new tab)Related Designs
FTC Non-Compete Rule Public Comments: Open Coding in QDA Mode
Open coding of public comments on the FTC's Non-Compete Clause Rule (docket FTC-2023-0007) in Potato's QDA Mode. The design demonstrates a codebook grown while reading, in-vivo codes, full-text search with search-and-claim, and shared memos raised as issues. Commenters keep the rights to their comments, so the ten sample comments are synthetic.
Political Discourse Analysis (AgoraSpeech)
Multi-task annotation of political speeches covering sentiment, polarization, populism, topic identification, and named entities. Based on AgoraSpeech (Sermpezis et al., 2025), featuring human-validated labels for comprehensive political discourse analysis.
AgentBoard Progress Scoring
Assess multi-turn LLM agent progress by identifying achieved milestones, scoring overall progress, categorizing the agent environment, and noting partial progress observations.