How to Interpret Your Analysis with Sense Making

Analysis does not end when the results appear. A frequency table, a regression output, a set of coded themes: each one answers something specific. Making sense of what they mean together, in relation to a real question and a real audience, is the part that takes time and judgment.

Sense Making is built for that step. You collect the results that matter for a particular question, describe what the study was about, and run a synthesis that reads everything together. The interpretation draws only on the outputs you have explicitly added. You can refine it with a plain-language instruction — the synthesis is rewritten and replaces the current version — then export to a Word document or copy the text directly.

01What is Sense Making, and why does it matter?

Sense Making is a module in AnalyZ Solutions that lets you collect results from across your analysis, provide context about your study, and run a synthesis that reads all of the findings together and produces a narrative interpretation. It is powered by AnalyZense, AnalyZ Solutions' built-in AI synthesis engine.

Without a structured interpretation step, findings stay fragmented. A cross-tabulation, a frequency table, and a regression each get written up separately, and conclusions are drawn from a partial reading of the data. Results that point in different directions go unnoticed. Patterns that only emerge across multiple analyses are missed. Sense Making is designed to close that gap.

The synthesis does not replace your judgment. It reads what you have added, identifies patterns and tensions across the results, and produces a draft interpretation based on the outputs you provided. You review it, edit it, and decide what to include in your report.

02Key concepts

Projects

A project is the container for a Sense Making session. Each project holds a set of analyses and an optional context description. You can have multiple projects: one per research question, one per dataset, or one per reporting period. Projects are independent of each other. Adding an analysis to one project does not affect any other.

Analyses

An analysis is a result you have run in any module in AnalyZ Solutions and added to a Sense Making project. It could be a frequency table from Descriptive Analysis, a cross-tabulation, a regression output from Inferential Analysis, a code frequency summary or co-occurrence table from Qualitative Analysis, or any other result. Each analysis is tagged with the module it came from and the title you gave it. Quantitative and qualitative results can be combined in the same project. You can add up to 20 analyses per project.

Context

Context is a free-text description you provide about the project. It tells AnalyZense what the study is about, how the data was collected, what the research questions are, and who the intended audience is. Context is optional but strongly recommended. A synthesis run without context produces a more generic interpretation than one where the research background has been specified.

Synthesis

A synthesis is the narrative interpretation AnalyZense produces by reading all the analyses in a project together, informed by the context you have provided. It identifies what the findings show, where they converge, where they diverge, and what they collectively imply for the research question. A synthesis can be run and re-run at any point as you add or remove analyses.

Before running, you choose one of four synthesis modes and a target length. The mode controls the tone, structure, and vocabulary of the output. The underlying findings are the same; the interpretation is shaped for a specific audience and purpose. The length selector gives you three options per mode — Short, Standard, or Detailed — so you can match the output to the space you have in your report.

03What belongs in a Sense Making project

Add the results that are most relevant to the research question you are trying to answer. Results from Descriptive Analysis, Inferential Analysis, Data Visualisation, and Qualitative Analysis can all be added, individually or in combination. A project that mixes a frequency table, a regression output, and a set of qualitative code frequencies gives AnalyZense a much richer base to interpret than one that contains a single result. Aim for coverage of the research question rather than completeness of every analysis you have run.

04How to Do It in AnalyZ Solutions

Adding results to a project

Results are added from within the analysis module where you ran them. After running an analysis, look for the + Add to AnalyZense button in the result panel. Clicking it opens a short form where you can give the result a descriptive title and select the project to add it to. If no project exists yet, you can create one from the same form.

1. Run the analysis. Complete the analysis in any module. Make sure the output reflects the final configuration you want included: the right variables, the correct filters, and the subset you intend to interpret.
2. Click + Add to AnalyZense. The button appears in the result panel after a successful run. If you do not see it, scroll down within the result panel.
3. Give the result a clear title. The title is what you will see in the project's analysis list and what AnalyZense uses to understand what each result represents. A title like "Cross-tabulation of food security by region" is more useful than "Table 1".
4. Select or create a project. Choose an existing project from the dropdown, or type a new project name to create one on the spot. Group results by research question rather than by module: put all findings relevant to a single question in the same project.
5. Repeat for each relevant result. Go back to the analysis module and run additional analyses. Add each one to the project using the same process. You can add results from different modules and different runs in any order.

Opening a project

Navigate to Sense Making from the main menu. The projects list shows all your projects with the number of analyses in each, the creation date, and the last modified date. Click any project to open it. If a project has been synthesised before, it is marked as Interpreted.

Adding context

Once inside a project, the Context panel appears at the top. This is where you describe the study in plain language. A useful context entry answers four questions:

• What is the study about? (Research question or evaluation objective)
• How was the data collected? (Survey, interviews, administrative records, mixed methods)
• Who are the respondents or units of analysis?
• Who will use the findings, and for what purpose?

Context does not need to be long. Two or three sentences covering those four points is enough to meaningfully improve the synthesis. Click Save context after writing it.

Reviewing the analysis list

Below the context panel, the Analyses section lists every result you have added to the project. Each entry shows the module it came from, the title you gave it, and the date it was added. Click any analysis row to expand a preview showing the variables used and a summary of the output. Before running a synthesis, scan this list to confirm it contains the right results. Remove any that are outdated or irrelevant using the trash icon on the right of each row.

Running the synthesis

With your analyses in place and context saved, choose your mode and length, then click Interpret project. AnalyZense reads all the analyses in the project together with the context you have provided and streams a written interpretation. This typically takes 15 to 30 seconds depending on the number of analyses and the length setting you selected.

The synthesis is structured into sections: an overall summary of what the findings show, an account of the key patterns and relationships across results, any tensions or apparent contradictions between findings, and implications for the research question or evaluation objective. Each section draws directly from the outputs you added.

Refining the interpretation

Once the synthesis is complete, a Refine interpretation panel appears below the text. Type a plain-language instruction — for example, "make this more concise", "add more on perceived behavioral control", "include more participant quotes", or "rewrite for a non-technical audience" — and the synthesis is rewritten incorporating it. The revised version replaces the original. Each refinement counts as one interpretation from your monthly allowance.

Version history

Each time you interpret or refine, the previous version is saved to a version history, accessible via a small toggle below the synthesis panel. Up to three previous versions are kept. Clicking Restore on any entry replaces the current synthesis with that version in the session — no Supabase write occurs until you run a new interpretation or refinement.

Exporting and copying

When the synthesis is ready, click Export to Word to download it as a formatted Word document, or Copy text to copy the plain text directly to your clipboard for pasting into any tool. The Word document includes the project name, interpretation mode, context, and the full synthesis text.

05Getting better results from Sense Making

Organise by research question, not by analysis type

A project that contains one frequency table, one cross-tabulation, and one regression all addressing the same outcome variable produces a more coherent synthesis than a project containing six frequency tables from different variables. Group results around a question: what drives food security? What predicts programme uptake? What characterises the treatment group relative to the control?

Use descriptive titles

The title you give each analysis when adding it is what AnalyZense uses to understand what the result represents. A title like "Regression of food security score on subsidy participation, controlling for household size and land ownership" gives the synthesis engine much more to work with than "Regression 2". Take an extra moment when adding results to write a title that describes what the analysis shows.

Add context before running the synthesis

Context shapes how AnalyZense frames the interpretation. A synthesis run with no context will describe the findings in general terms. A synthesis run with a clear research question and audience description will frame the findings in relation to that question and that audience. Write the context before the first synthesis run, not after.

Refine rather than re-run for small adjustments

If the synthesis is largely right but needs a specific adjustment — a different emphasis, a shorter conclusion, more attention to a particular finding — use the Refine panel rather than re-running from scratch. Refining is faster and preserves the overall structure while incorporating the instruction. Re-run from scratch when you have added or removed analyses, changed the mode, or want a fundamentally different take.

Re-run after adding or removing analyses

The synthesis is not automatically updated when you add a new analysis or remove an old one. If you add a result after an initial synthesis, run the synthesis again to produce an updated interpretation that includes the new material.

Use the analysis preview to check before running

Before running a synthesis, click each analysis row in the project to expand the preview. Check that the variables listed match what you intended, and that the result summary looks right. Removing an incorrect or outdated result before running is faster than reviewing a synthesis based on wrong inputs.

Refine with follow-up questions

After a synthesis is produced, a chat panel opens below the output. You can ask follow-up questions to dig into specific aspects of the findings. For example: "What are the main contradictions across these results?", "What does this suggest for programme design?", or "Summarise the key findings for a non-technical audience." Each response draws on the same project findings, not general knowledge. You can ask multiple questions in sequence, building toward a more detailed interpretation as you go.

Export to Word

When the synthesis is ready, export the full output, including any follow-up exchanges, to a Word document using the Export button in the synthesis panel. The exported document includes the synthesis text formatted for editing. Use it as a starting point for a findings section, a briefing note, or an evaluation report. Edit and expand from there rather than treating it as final text.

06Common mistakes to watch for

• Adding too few analyses. A synthesis based on one or two results produces a narrow interpretation. Add enough results to represent the key dimensions of your research question before running.
• Adding redundant analyses. Adding three versions of the same cross-tabulation with minor differences dilutes the synthesis rather than improving it. Add the final version of each analysis, not intermediate iterations.
• Skipping context. A synthesis without context produces a generic interpretation that could apply to almost any dataset. Even two sentences describing the study and the research question make a significant difference.
• Treating the synthesis as final text. The synthesis is a draft. Read it critically, check specific claims against the original outputs, and edit before including anything in a report or presentation.
• Mixing unrelated questions in one project. A project that contains findings about programme uptake, household food security, and enumerator performance all at once produces a synthesis that struggles to find a coherent thread. Keep each project focused on one research question or evaluation objective.
• Not re-running after changes. If you add, remove, or replace an analysis after a synthesis has been run, the existing synthesis no longer reflects the current project. Re-run to get an up-to-date interpretation.

• You have run analyses across multiple modules and want a coherent interpretation of what they show together.
• You are preparing a report or presentation and want a structured draft of the findings section.
• You are working with a team and want a shared record of the analyses used and the interpretation they support.
• You want to check whether your results are internally consistent before presenting them to stakeholders.
• You are returning to a dataset after a break and want to reconstruct what was found and why.
• You want to try different interpretation styles — academic for a paper, plain language for a community report — from the same set of analyses.
• You want to iteratively improve a draft interpretation through refinement without losing earlier versions.