Independent evaluation & analytics consultancy

Defensible analysis that informs the decisions that matter.

Archimedes Evaluation Services provides mixed-method program evaluation, statistical modeling, predictive analytics, and fraud-detection methodology for government, nonprofit, and private-sector clients. Rigorous methods. Stakeholder-ready reports. Decisions you can defend.

Start a conversation See selected work
20+
Years applied analysis
$6.5B
Portfolio analyzed
35%+
Predictive accuracy lift
10k+
Learners evaluated
Erik Halvorson, Principal
Erik Halvorson, Principal
Founder & Lead Methodologist
Ph.D. Cand., Research Methods & Statistics M.S., Columbia USAF Veteran
About Archimedes

An agile evaluation practice built on rigorous method and plain-English findings.

Founded in 2021, Archimedes Evaluation Services helps organizations make better decisions through defensible mixed-method evaluation, advanced statistical analysis, predictive modeling, and fraud-detection research.

Our work sits at the intersection of quantitative depth and qualitative insight. We design custom psychometric instruments, run mixed-method evaluations with criteria-based dimensions and sub-dimensions, build predictive models with formal validation reporting, and apply text mining and sentiment analysis to data others overlook.

Clients have included Jackson Laboratory, Habitat for Humanity, Northwest Center, mindSpark Learning, A Little Help, the Colorado Department of Education, an international airline, and ACFE Research Institute-funded federal fraud research. Every engagement delivers technical rigor and findings that non-technical leadership can act on.

What we do

Six lines of work, all grounded in method.

Capabilities range from formal program evaluation to predictive fraud modeling. We scope every engagement to the decision the client needs to make.

Mixed-Method Program Evaluation

Formative, process, and outcome evaluations using surveys, interviews, focus groups, psychometric instruments, and criteria-based rubrics with dimensions and sub-dimensions.

Evaluability assessment Logic models Process eval

Statistical Analysis & Modeling

Inferential statistics, factor analysis, latent variable modeling, sensitivity analysis, cost-benefit analysis, break-even reasoning, and Monte Carlo-style simulation for decisions under uncertainty.

SEM Factor analysis Sensitivity

Predictive & Fraud Analytics

Machine learning models, NLP feature engineering, network analytics, and risk scoring frameworks built with formal validation reporting and explainability standards.

ML & NLP Anomaly detection Risk scoring

Psychometric Instrument Design

Custom survey instrument development, scale construction, reliability and validity testing, and pre/post analysis for outcome measurement in education, behavioral health, and program settings.

Scale development Pre/post Reliability

Technical Reporting & Visualization

Methodology documentation, technical findings briefs, executive summaries, and infographic-format visualizations that translate complex analysis into decision-useful guidance.

Tech reports Infographics Executive briefs

Oversight & Integrity Methodology

Risk-based oversight frameworks, fraud-detection methodology, defensible assumption documentation, and program-integrity analysis informed by 13 years of federal oversight experience.

Risk-based oversight Integrity reviews Fraud research
Selected work

A portfolio across nonprofit, government, and private-sector clients.

Engagements range from text mining at scale to multi-state cost-driver analysis to predictive fraud modeling. Each project is scoped to the decision the client needs to make.

Completed

Jackson Laboratory

Bar Harbor, Maine

Educational analytics for a nonprofit biomedical research organization. Text mining and sentiment analysis of unstructured open-response data across 10,000+ learners and 41 continuing-education modules. Pre/post question alignment supports module updates and journal-publication research.

Text mining Sentiment analysis Pre/post
Completed

Northwest Center, At-Risk Youth Programming

Chicago, Illinois

Year-two participatory evaluation of a new at-risk youth program. Psychometric instrument design, pre/post analysis, interview and focus-group design, and rubric-based evaluation using criteria-based dimensions and sub-dimensions to support iterative program refinement.

Participatory eval Psychometrics Rubric design
Completed

Habitat for Humanity — Repair Corps

Decatur, Georgia

Analytics, visualization, and evaluability assessment for the Repair Corps Veterans support program, in partnership with Habitat's U.S. Research, Outcome, and Measurement Group. Data review, descriptive and trend analysis, and infographic-format reporting.

Evaluability Trend analysis Infographic
Funded Research

SBIR Fraud Predictive Evaluation

ACFE Research Institute, Austin TX

ACFE Research Institute-funded study of the U.S. Small Business Innovation Research program using principal component analysis, qualitative inquiry, novel NLP-derived feature engineering, and predictive modeling. Greater than 35% improvement in predictive accuracy in validation.

PCA NLP features Predictive modeling
Completed

A Little Help

Denver, Colorado

Mixed-method evaluation of a nonprofit supporting older adults aging in place. Geographic mapping, statistical analyses, semi-structured interviews, synthesis methodology, and sentiment analysis. Two technical reports delivered, expanding the program's logic model and evaluability.

Mixed-method Geospatial Sentiment
Completed

mindSpark Learning

Lakewood, Colorado

Outside evaluator for a nonprofit delivering professional development to educators. Macro-level business-alignment review, custom psychometric design, sentiment analysis of unstructured survey data, and an exploratory evaluation using geographic mapping and thematic qualitative review.

Exploratory eval Psychometrics Thematic
Completed

International Airline — Hub Selection

New York, New York

Multi-factor demand-analytics framework integrating airline customer data with millions of datapoints from US Census, US and international economic indexes, travel data, and foreign-arrival forms. Contributed to the establishment of a new hub in Trenton, New Jersey.

Demand analytics Multi-factor model C-suite reporting
Pending Research

Afghanistan Refugee & Veteran Narrative Inquiry

University of Denver

Mixed-method study with DU's Graduate School of Professional Psychology and nonprofit Task Force Argo, pairing refugee and veteran-evacuation-staff narratives to inform future emergency-evacuation policy and limit negative outcomes in wartime scenarios.

Narrative inquiry Mixed-method Policy
Speaking & Recognition

Trusted by professional bodies in fraud, oversight, and evaluation.

Selected speaking, instruction, and federal-scholar appointments.

ACFE Global Conference Speaker

"Incorporating Analytics and Research Methods into Anti-Fraud Program Management." 33rd Global Fraud Conference, Nashville, TN (2022).

DOJ National Institute of Justice LEADS Scholar

Selected federal scholar program advancing research-to-practice in justice, integrity, and oversight work. 2023 cohort.

Association of Inspectors General Instructor

Faculty for AIG programs on analytics-driven oversight methodology and evaluation design.

American Accounting Association Panelist

Expert panel contributor on emerging research methods, AI-enabled analytics, and the future of fraud detection in accounting.

ACFE Research Institute Grantee

Funded principal investigator on predictive-fraud research applying novel NLP feature engineering to a federal-program dataset.

Federal Grant Fraud Working Group

"From Imagination to Implementation: Creating a Proactive Anti-Fraud Endeavor." Quarterly meeting presentation.

Have a decision that deserves better analysis?

We work with a small number of clients at a time and scope each engagement to fit. Reach out to start a conversation.

erik.halvorson@archimedeseval.com 303.901.3070