# Analyze Microsoft Forms Responses: Excel Copy or Connected Workbook? Analyze Microsoft Forms responses: choose an Excel copy or connected workbook, preserve useful IDs, and trace comments to their source. - Canonical URL: https://www.harmate.com/en/blog/microsoft-forms-data-analysis - Author: Harmate Team - Published: 2026-10-25 - Updated: 2026-10-25T22:47:57.743215+00:00 - Language: en ## Content # Analyze Microsoft Forms Responses: Excel Copy or Connected Workbook? Before analyzing Microsoft Forms responses, identify the file you are using. A connected Excel workbook can receive new responses; a downloaded copy is a snapshot. Two people working from different versions may reach different totals even when neither has made an analytical mistake. This guide shows how to hand over a usable response dataset: choose its cutoff, retain useful identifiers, separate questions from metadata, and connect findings to the original comments. It covers form responses, rather than grading a quiz. ## Decide whether you need a live view or a fixed dataset In Microsoft Forms, open the form and select **View responses**. The response overview provides counts and question summaries, while individual results let you return to a particular submission. This is useful for monitoring collection. It does not define which responses belong in a report. Microsoft’s documentation distinguishes two routes: | Your task | Option to examine | What it means for analysis | |---|---|---| | Keep following incoming responses | **Open results in Excel** | A connected workbook can receive new data. Record when you stopped the dataset used for your analysis. | | Give a stable version to a colleague or another tool | Arrow next to the button, then **Download a copy** | The copy is offline and has no connection to the form. Later responses do not appear in it automatically. | For the work or school accounts described in the help page, the connected workbook is stored in OneDrive; group forms use SharePoint. Microsoft also states that the form owner and people in the organization with workbook edit permissions can trigger a refresh, while external users cannot. The available options depend on the account and environment. Check the route visible in your interface rather than assuming that a file received by email is still connected. See Microsoft’s official [Check and share your form results](https://support.microsoft.com/en-us/forms/check-and-share-your-form-results). Choose according to the job. Ongoing monitoring and a report with a stated cutoff may use the same form without using the same set of responses. ## Worked example: two reports, two versions of one form Consider an entirely fictional situation. After a professional event, an organizer reviews feedback from exhibitors. At 10 a.m., she downloads a copy containing 12 responses. At 2 p.m., a colleague opens the connected workbook and sees 15. Three responses arrived in between. Their totals can differ without any missing rows. Before comparing findings, they check extraction time, filters, and the response IDs present. They then choose either to report the 10 a.m. dataset or to repeat the analysis using the 2 p.m. dataset. They do not silently combine new comments with the earlier totals. A handover note can be brief: | Field | Fictional example | |---|---| | Form and question examined | Exhibitor feedback — “Describe a moment when meeting someone was useful or difficult.” | | File type | Downloaded XLSX copy, offline | | Dataset cutoff | October 24, 10 a.m.; 12 responses present | | Filters | None applied to the copy | | Identifier retained | Form response ID | | Collection after cutoff | Three additional responses seen at 2 p.m.; excluded from this version | | Intended use | Prepare changes to the next event’s meeting arrangements | This describes scope, not quality. The 12 responses do not automatically represent all exhibitors, and the three later responses are not less legitimate. ## Separate questions from context columns In the export Microsoft describes, each response is a row and questions appear as columns. The initial columns include the response ID, start and completion times, and name and email where these are collected. The help page states that names and email addresses are not displayed when the form accepts anonymous responses. The response ID helps you locate a submission within that form. It does not, by itself, prove that two submissions came from two different people. A spreadsheet row number is not a stable identifier after sorting either. Review columns by purpose: - **Questions:** what the person answered, including the exact wording and instructions. - **Useful context:** collection period, form version, or another detail needed to interpret an answer. - **Identity:** contact information and personal identifiers to retain or share only when needed for the task. Timestamps help situate a response. They do not diagnose attention, honesty, or response quality. The interval between starting and finishing can include an interruption. If you need to combine several copies, first check their provenance and the questions actually asked. The guide to [preparing survey responses before analysis](/en/blog/prepare-survey-responses-before-analysis-a-practical-method-and-example) covers that review. A more recent file does not automatically represent a new population to add to the previous one. ## Read a comment with its question and ID In the fictional example, response ID 8 says: “I met two useful contacts, but only after the designated meeting area had closed.” Summarizing this as “the meetings worked well” would erase a practical difficulty. Saying “the meeting area was useless” would go beyond the evidence. A more faithful working note would preserve: | Element | Fictional analysis note | |---|---| | Source | ID 8, 10 a.m. copy, question about meetings | | Provisional reading | Useful conversations happened, but outside the planned arrangement. | | Next check | Reread other comments and check the event schedule before changing the hours. | Keep responses describing a different experience visible too. Categories and themes organize reading; they do not replace the full wording. The method for [analyzing an open-response corpus](/en/blog/analyzing-200-open-responses-without-bias-a-method-for-actionable-decisions) develops this next step. Do not treat the most vivid comment as the most common experience. If you report a count, state which version and question it covers. A response missing from the 10 a.m. copy may simply have arrived later, and an empty answer does not explain why the person left it blank. ## Hand over a usable copy Before transferring the file, preserve the original, prepare a working copy, and record columns removed or renamed. Check that a free-text question has not been mistaken for a name, date, or status field. Retain the context needed for interpretation without transferring every contact detail by default. You can then [import the response copy into Harmate](https://www.harmate.com/en/product/response-import) as a CSV, JSON, or XLSX file, review its columns, and examine themes alongside their source excerpts. This is a manual file workflow; it does not assume a direct Microsoft Forms connection to Harmate. A useful report lets a colleague find the file, its cutoff, and the responses supporting each finding. That makes it possible to distinguish a real change in feedback from a change in the dataset being analyzed.