LMSW guideTopic deep dives
Research and Program Evaluation on the LMSW Exam
How research and program evaluation appear on the ASWB Masters exam, with a plain-language table of the concepts the outline names and worked examples of matching a purpose to a method.
Short answer
Research and evaluation is one competency, IIIC Practice Evaluation and Research, inside Intervention and Practice, the 32% content area. ASWB does not publish how many items each competency gets, so expect a modest share rather than a section of its own. The outline names evaluating progress and practice, agency program evaluation (needs assessment, formative and summative evaluation, cost-effectiveness, cost-benefit, and outcomes), research design, data collection and analysis, and reliability and validity. Items usually ask you to choose the right method or concept for a stated purpose.
How much research is on the Masters exam?
Intervention and Practice is 32% of the Masters outline, about 35 of the 110 scored questions by Passy's arithmetic. It has four competencies: practice concepts, intervention methods, practice evaluation and research (IIIC), and supervision and administration. ASWB does not say how many items each competency receives, so there is no official research count to quote. Plan for a modest share of that area. Research ethics, such as review boards and participant consent, is a listed statement on the 2026 Clinical outline but not on the 2026 Masters outline.
The IIIC statements cover four things:
- techniques for evaluating progress and the effectiveness of interventions and service plans
- methods and instruments for evaluating social work practice
- methods of evaluating agency programs, with needs assessment, formative and summative evaluation, cost-effectiveness, cost-benefit analysis, and outcomes assessment as examples
- basic and applied research design, data collection and analysis, and ways to assess reliability and validity, including objective and subjective, qualitative and quantitative data
What is on the LMSW exam? shows how IIIC sits next to the other nine competencies.
The concepts in plain language
| Concept | What it means | The cue in a stem |
|---|---|---|
| Reliability | The measure gives consistent results: over time (test-retest), between raters (inter-rater), or across items (internal consistency) | Two workers score the same client very differently on the same scale |
| Validity | The measure captures what it is meant to measure | A checklist sold as a full depression measure asks only about sleep and appetite and never about mood. Sleep and appetite can be valid items; leaving out the rest of the construct is the problem |
| Reliability vs validity | A measure can be consistent and still wrong, but an inconsistent measure cannot be valid | A scale that always reads five pounds heavy is reliable, not valid |
| Internal vs external validity | Internal: did the intervention cause the change? External: do the results generalize to other people and settings? | "Could something else explain the improvement?" vs "Would this work at another agency?" |
| Needs assessment | Identifies unmet needs and service gaps before planning a program | A county asks whether to open a new service |
| Formative evaluation | Done to improve a program, usually while it is being developed or run; overlaps with process evaluation | A director wants to fix problems halfway through the first year |
| Summative evaluation | Done to judge whether a program achieved its intended results, often at the end of a program or cycle; overlaps with outcome evaluation | A funder asks whether the program worked |
| Cost-effectiveness | Cost per unit of a non-dollar outcome | Cost per family housed for a year |
| Cost-benefit | Costs and benefits both expressed in dollars | Program cost compared with money saved on emergency care |
| Single-subject (single-system) design | One client or system measured repeatedly: baseline (A), then intervention (B), sometimes withdrawn and repeated (ABA, ABAB) | A worker charts a behavior daily before and during an intervention |
| Experimental vs quasi-experimental | Experimental designs randomly assign participants to groups; quasi-experimental designs test effects without random assignment, by comparing nonequivalent groups or one group at different points in time, as in a time series | "Clients were randomly assigned" vs "We compared two existing caseloads" |
| Qualitative vs quantitative | Words, themes, and meaning vs numbers and measurement | Focus group themes vs scores on a standardized scale |
| Independent vs dependent variable | What is changed or compared vs what is measured as the result | The intervention vs the symptom score |
| Probability vs nonprobability sampling | Random selection, so every member has a known chance, vs convenience, purposive, or snowball samples | "Every fifth client on the roster, from a randomly chosen starting point" vs "whoever came to group this week" |
Six worked mini-examples
The examples are invented for this page, not taken from ASWB.
- The drop-in center. A county board wants to know whether teenagers in two neighborhoods need an after-school drop-in center. The right first step is a needs assessment, using sources such as existing data, key informant interviews, a community forum, or a survey of families and youth.
- Halfway through year one. The director of a new job-readiness program wants to know whether sessions are running as designed and what to adjust. That is formative, or process, evaluation.
- The funder's question. Twelve months later the funder asks whether participants are more likely to be employed than before. That is summative, or outcome, evaluation.
- Two ways to count value. One analysis reports $4,800 per family stably housed. That is cost-effectiveness, because the outcome stays in housing terms. Another converts housing stability into dollars saved on shelter stays and emergency care and compares them with program costs. That is cost-benefit. These dollar figures are illustrative, not real program data.
- The classroom chart. A school social worker records a student's daily outbursts for two weeks, then starts a behavior plan and keeps recording. That is a single-subject AB design, and the first two weeks are the baseline.
- Two raters, two scores. Two case managers use the same functioning scale with the same family and get very different scores. The problem to fix first is inter-rater reliability, through training on the scale or clearer scoring rules.
Research ethics as background
The 2026 Masters outline has no separate research ethics statement, so treat this section as background for evaluation stems rather than a topic of its own. The NASW Code's standard 5.02 sets the ethics of evaluation and research. The parts worth knowing:
- Consent must be voluntary and informed, usually written, with no coercion, penalty, or undue inducement, and it must cover the nature, extent, duration, risks, and benefits of taking part (5.02(e)).
- Review boards are consulted to protect participants (5.02(d)).
- People may withdraw at any time without penalty (5.02(i)), and they should have access to supportive services (5.02(j)).
- When a participant cannot consent, the researcher explains the study, seeks the person's assent to the extent possible, and gets written consent from an appropriate proxy (5.02(g)).
- Confidentiality and anonymity are protected, with identifying information left out of reports unless consent allows it (5.02(m) and (n)).
- Results are reported accurately and never fabricated or falsified (5.02(o)).
For example, a stem might describe a social worker who wants to study outcomes using current clients' files. The keyed answer protects clients through review board approval and consent where required, and makes clear that services do not depend on taking part. Starting the study quietly because "the data is already ours" is the near miss.
How to study this competency
Learn it as purpose-to-method matching. For each concept in the table, write one question an agency director might ask that the concept answers, then check yourself against practice items. The free mixed LMSW set has 20 questions across all three areas and includes one program-evaluation item, so match purpose to method on the evaluation item in the free LMSW set, then use a research methods reference for design, sampling, reliability, and validity. Intervention and practice covers the other three competencies in the area, and theories and practice models covers the interventions whose effectiveness these methods measure.
Trust the source
Official sources
Exam policies can change. Use these primary sources for the most current details.