DATA COLLECTION TOOLS · PRACTICAL GUIDE
WATCH THE WALKTHROUGH
Watch episode 1 of Sopact’s data-collection series. The three companion episodes below cover analysis, change over time and reporting.
When you run several programs, each intake form can work perfectly while the combined evidence remains difficult to use. Attendance lives in a spreadsheet, participant feedback sits in a survey project, and staff notes stay in a folder. A prospect described the consequence plainly: “We can’t tell a cohesive story across all of it.”
This guide helps program teams choose the collection instruments and connections they need. In Sopact’s approach, the starting point is the participant record: decide which evidence belongs together before launching another form. For product shortlists, continue to the data collection software guide or the nonprofit software comparison.
Use this table to choose the next instrument in an Outcome Thread. The missing evidence determines the choice; a software subscription does not decide what the program should ask.
| Evidence needed | Tool to use | Check the handoff |
|---|---|---|
| Who joined each program? | Intake or enrollment form | One person can have several enrollments. |
| What does the participant report? | Questionnaire or check-in | Keep scale, date and question version. |
| Why did participation change? | Interview guide and notes | Separate their words from interpretation. |
| What happened during delivery? | Observation checklist or log | Distinguish visits, events and people. |
| What evidence already exists? | Document extraction template | Retain the file and supporting passage. |
| Did the change persist? | Follow-up form | Keep missing responses visible. |
Sopact calls the connected participant history the Outcome Thread: one persistent Contact ID linking intake, program participation, supporting evidence and later follow-up without replacing earlier observations. A submission becomes an event in that history. It does not become a new person simply because another department collected it.
Sopact documents contact-to-form relationships and analysis of open-ended responses and PDF attachments while retaining that relationship. Those are foundations for a connected toolkit; the team still has to configure the program structure, define fields and decide which sources belong in an analysis.
The useful distinction is between a collection process that restarts identity at each submission and one that preserves identity across stages. Other products also support contacts or linked records. Evaluate whether your configured workflow carries the right context into the answer, rather than assuming a unique-ID feature settles the entire problem.
Paper forms and spreadsheets made it possible to capture local information with little setup. Online survey platforms such as SurveyMonkey and Qualtrics expanded distribution and analysis. Case systems such as Apricot and Casebook organize information around ongoing service work. These approaches overlap; a mature survey platform can have contact management, and a case platform can have configurable forms.
Sopact’s emphasis is the connection between collection and recurring program analysis: which participant, which program, which stage and which source support the answer? Adding AI to a disconnected folder does not establish those relationships. A well-designed existing system may already preserve some of them.
The evaluation test is to follow one person from intake through a later update and explain a result using the original sources. Include a correction and a missed follow-up. A summary chart alone cannot demonstrate that continuity.
These are six practical instruments, not six mutually exclusive products. One application may support several. Build only the instruments needed for your decision, then connect them through the Outcome Thread. For research-method selection and methodological tradeoffs, use the data collection methods guide.
Use an intake form to establish who is joining and the information needed to deliver the service. Keep the participant identity separate from enrollment: the same person may join mentoring and employment support, or return next year. In Sopact, start with the contact relationship and decide which details belong to the person versus each program entry.
Capture program, enrollment date and the agreed baseline only where needed. Test a returning participant before launch. A new enrollment should not erase the earlier enrollment or silently create a duplicate contact.
Use a questionnaire when people can answer a defined set of questions consistently. A rating plus an optional explanation can reveal both a reported change and a reason worth investigating. In an Outcome Thread, preserve the question version, response scale and collection date alongside each answer.
A confidence rating is not a skills assessment. Keep those measures distinct, and do not substitute one for the other in a report. Choose relevant items from impact survey questions rather than sending every participant the full question bank.
Use an interview guide when the team needs to explore how something happened or why a short answer is unclear. Include a small common core and space for follow-up questions. Keep a participant’s words separate from the interviewer’s interpretation; those are different kinds of evidence.
For Sopact analysis, retain the source reference and the approved participant or program relationship. An interview about a household or organization may need a different record structure. Do not attach every speaker’s statement to one person merely because that person arranged the conversation.
Use a checklist to record a defined event or behavior: attendance, a completed practice task or an observed service condition. Specify what counts, who observed it and when. Sopact’s Outcome Thread can connect those dated observations to subsequent participant feedback when the relationship is configured.
Six attendance entries represent six visits, not six people. Preserve event records and calculate unique participants separately. If two staff members interpret “completed” differently, fix the definition and train the observers before treating their records as comparable.
Use existing documents when they contain evidence the team would otherwise ask people to submit again. An extraction template identifies the fields to read, the source date and the passage supporting each finding. In Sopact, supported PDF evidence can be analyzed alongside responses within the configured contact relationship.
File storage, reading a file and producing a signed document are separate requirements. Test the actual format and scan quality. If a document does not state an outcome or a date, record that gap; do not let a plausible AI summary supply the missing fact.
Use a follow-up form when the question requires another observation after service, training or an intervention. Keep the participant identifier, collection window and comparable questions explicit. Sopact’s documented collection workflow provides participant-specific links for configured follow-up forms; verify delivery and reminder requirements separately.
Track who was eligible, invited and reached, as well as who responded. An unanswered form is missing evidence, not a zero outcome. The detailed design belongs in longitudinal data collection.
Sopact’s Outcome Thread depends on a small shared structure. For each collection event, retain the appropriate participant or entity ID, program or enrollment reference, event date, collection stage, question or field version, and source reference. Add the person responsible for review and a clear status for missing or corrected information.
A shared ID answers “whose record is this?” It does not answer “what does this number mean?” Two programs may both record employment, but one counts any paid work and another requires sustained work for a defined period. Keep their measures separate unless an agreed definition makes aggregation valid.
Write those agreements in a shared data dictionary. Include units, allowed values, missing-value rules and ownership. A dictionary helps people interpret the data; the system must also preserve the source and correction history. Adding a column called “verified” does not create either one.
Choose a connected set of data collection tools by naming the decision, identifying the evidence gap, selecting the smallest suitable instrument and testing how its records join the next collection stage. In Sopact’s Outcome Thread, the test follows the same participant across forms and sources, with program context and dates intact.
Start with a question someone will act on: “Which participants need help staying engaged this month?” Identify what you already know from enrollment and attendance. Add a short check-in or a few interviews only if they answer an unresolved question. Avoid asking participants to reconstruct information your team already holds.
Next, test the full path with a small authorized sample. Include one person in two programs, one missing follow-up, a corrected answer and a document. Ask the person who will maintain the toolkit to explain a combined result and open the evidence behind it. Record any step that depends on an administrator or manual reconciliation.
Finally, test accessibility, language and connectivity under real collection conditions. Preserve the collection mode where it affects interpretation. Mixed-mode collection explains the tradeoffs between web, telephone, paper and in-person responses. Confirm actual offline support instead of inferring it from a product’s general positioning.
Consider a fictional organization running job training and mentoring. Participant P-104 enrolls in both. One intake establishes the person’s contact record; two enrollment records preserve the program-specific context. Attendance logs capture visits. A check-in records a scheduling problem. A later interview explains that shift work conflicts with the mentoring time.
In a proposed Sopact Outcome Thread, those events remain connected without becoming interchangeable. The team counts one person, two enrollments and the actual visits. The participant’s interview may inform a schedule change; it does not establish how common the problem is across everyone who missed sessions.
At follow-up, P-104 reports using a new skill. The team retains the earlier answers, the new response and its date. If P-104 skips the next wave, the record remains in the eligible group with a missing response. This is an illustrative design to demonstrate, not a customer result or a claim that configuration happens automatically.
Often, yes. The recurring prospect question is whether Sopact can be “an AND” alongside an existing system. Keep a working case, donor or finance system where it serves its purpose. Agree how relevant records will reach the collection and analysis workflow through supported imports or integrations.
Demonstrate how identifiers, changed records and permissions survive that handoff. Historical records without reliable identifiers may require reviewed matching; an import does not recover certainty that the source never captured. Sopact’s Outcome Thread is a continuity design, not a promise that every old spreadsheet joins automatically.
Sopact’s Loop methodology connects three recurring moves: collect clean at the source, analyze on arrival and improve in time to act. For a toolkit, that means reviewing new information while staff can still clarify a response, adjust support or repair a confusing question.
Assign an owner and review date to the change you choose. Keep the measure stable enough to examine the next cycle, and document any revision. Use an evidence trail so a colleague can inspect the records and reasoning. Connected evidence supports learning; it does not by itself prove that the program caused an outcome.
Preserve identity, dates and meaning.
Review new evidence and missing responses.
Assign a change and check the next cycle.
Keep the Outcome Thread usable for the next decision. Put the Loop into practice →
The Loop Sessions library pairs the opening video with three companion episodes. Watch them as a sequence from choosing sources to examining the result:
Use these prompts to draft a plan in Sopact’s Assistant or work through the questions with your team. Provide only approved information. The prompts request draft artifacts for review; they do not configure your system or authorize changes to participant records.
What: Specify the fields that connect the toolkit before building its forms.
Academy walkthrough → Build a shared data dictionary
Using [PROGRAM DESCRIPTION] and [EXISTING FIELDS], draft a data dictionary for our Outcome Thread. Separate participant, enrollment and event fields. Include field definitions, units, allowed values, missing-value rules, source references and owners. Flag uncertain mappings; do not infer identities from names alone.
Expected output: A draft dictionary and a list of relationships requiring confirmation.
Tips: Have the program owner check the counting unit. Test one participant with two enrollments.
What: Connect the collection plan to the program’s intended outcomes.
Academy walkthrough → Build a logic model
Build a draft logic model from [PROGRAM PLAN]. Separate activities, outputs and outcomes. For [DECISION TO MAKE], identify evidence already available in [SOURCE LIST] and recommend the smallest additional collection instrument. Explain how each source would connect to the Outcome Thread. Mark assumptions and unsupported outcome claims.
Expected output: A program model with an evidence gap and a justified instrument choice.
Tips: Remove questions without a named use. Check whether an existing document already contains the answer.
What: Decide which evidence is needed when, and who owns its review.
Academy walkthrough → Build a results framework
For [OUTCOME] in [PROGRAM], draft a results framework using [AGREED DEFINITIONS]. Specify each indicator, source, collection window, eligible population and responsible person. Map baseline and follow-up to the same Outcome Thread. Distinguish missing, declined, not applicable and zero. Flag measures that cannot be compared across stages.
Expected output: A collection schedule with sources, denominators and review responsibilities.
Tips: Demonstrate follow-up with the actual delivery process. Do not assume reminders or offline capture are included.
What: Use connected evidence to examine an assumption and choose a bounded next step.
Academy walkthrough → Build a theory of change
Using [PROGRAM THEORY] and [APPROVED EVIDENCE SUMMARY], identify one assumption we can examine next. Separate observed change, participant explanation and causal inference. Propose one follow-up question, its collection tool, the Outcome Thread fields needed and a review date. List evidence gaps without inventing outcomes or participant quotations.
Expected output: One learning question, a proposed collection step and the limits of the current evidence.
Tips: Review the source records before acting. An improving score alone does not establish program impact.
Examples include intake forms, questionnaires, interview guides, observation checklists, document extraction templates and follow-up forms. Sopact’s Outcome Thread provides a way to connect their results to the same participant while retaining the dates, program context and sources needed to interpret each observation.
A method is the approach, such as interviewing or observation. A tool is the instrument or application used to carry it out, such as an interview guide or checklist. Sopact’s Outcome Thread concerns how the resulting evidence stays connected across those instruments.
Choose the smallest set that answers the program’s decisions and fits its participants and staff. A registration form may cover a one-time event; recurring outcomes need later observations and reliable connections. Sopact’s Outcome Thread is relevant when several instruments must contribute to one participant history.
Yes, when the team can maintain the required collection, access, linking and review process. Assess staff work as well as subscription cost. Sopact’s Outcome Thread offers a comparison point: demonstrate how each option preserves the same participant, definitions and source evidence at the next collection stage.
Establish a persistent participant identifier and configure later forms to reference that participant. Preserve each response’s date and stage. Sopact’s Outcome Thread uses contact-to-form relationships for this continuity; test returning participants, duplicates and missing responses before launching the full collection.
Yes, when the evidence is connected at the appropriate person, program or other unit and interpreted with clear definitions. Sopact’s Outcome Thread can put a participant’s rating beside the words explaining it. A quote provides context; it does not establish prevalence among people who did not respond.
No. Anonymous feedback must remain anonymous, and some evidence belongs to a group, site or organization. Sopact’s Outcome Thread is useful for authorized participant-level follow-up. Choose a different record structure when the purpose or collection promise does not support a person-level relationship.
AI can help interpret supported text, but it cannot reliably reconstruct missing identities, unstated definitions or evidence that was never collected. Sopact’s Outcome Thread supplies collection context before analysis. Staff still need to review source coverage, ambiguous findings and the limits of any conclusion.
BRING YOUR REAL WORKFLOW
Bring the question your team needs to answer and the information you already collect.
See it with your workflow ↗
