THE SHORT ANSWER
An aggregate figure can help frame a question, but it needs context. Who contributes to the data, when is it from and which values cannot be shown? Start there before interpreting a result.
Employer insurance statistics in Forsio are based on participating employees with current permission for the statistics and available documents. They do not map every employee. A percentage is not a grade for how well the whole workforce is insured, either. [1][2]
Keep these points in mind
- A participant group is not automatically the entire workforce.
- A suppressed value does not mean zero.
- An aggregate observation gives no right to read individual records.
STEP BY STEP
How to review the documents
Read who contributes to the data
Distinguish the organisation's declared workforce count from the group contributing to the statistics. Not every employee has to participate, and participants may not have added the same kinds of documents. Record what the population and available information actually cover.
Statistics Sweden explains how sampling and non-response affect statistical conclusions. A large group does not automatically represent everyone else. Forsio's participating group is not a random sample of the entire workforce or of Swedish policyholders.
[2][3][1]Check the date and scope of the data
Read the date the snapshot concerns and its accompanying explanations. A saved summary describes its data at that point in time. Changes in participation, documents or analysis can make two summaries differ.
A changed percentage therefore does not necessarily mean an employer action caused the change. Check whether groups, periods and definitions are comparable before describing a trend. Older information cannot simply be presented as today's insurance overview.
[1][3]Distinguish a suppressed value from a known zero
Forsio limits display for small participant groups and small subgroups in the statistics. The employer minimum group size is at least 25 people and can be higher. A sufficiently large overall group does not mean every individual row can be shown.
Certain sensitive insurance categories are excluded from separate per-line reporting. A row that is not shown cannot therefore be read as nobody having particular cover or no issue existing. Do not fill a suppressed value with a guess or try to infer which people might be behind it.
[1][5]Check what a comparison means
If a reference is shown, read who contributes to it and what the measure concerns. Forsio's combined reference is based on Forsio participants. It is not a representative average of all Swedish households or insurance policies.
An observed gap or overlap concerns the information and rules included. It does not establish that every employee needs the same action or that a new group agreement is suitable. Use the result for questions about the organisation's existing information and documents, not as an automatic shopping list.
[1][2]
Display limits are safeguards, not a guarantee of anonymity
The Swedish Authority for Privacy Protection describes anonymous information as data from which a person can no longer be identified. Removing names or combining figures does not itself prove that every data protection question is resolved. The information needs assessing in context.
Forsio's employer view does not disclose individual policies, personal answers or a list of who has joined. Permission within the product is not a determination of the employer's legal basis. The authority states that employers normally cannot rely on consent to process employee information.
[4][5][1]Keep the observation separate from the decision
An aggregate observation can point to a question about how an existing benefit is described or which document needs clarification. It does not assess every person's agreement or decide entitlement to compensation. Several personal insurance policies can also serve different purposes.
Keep the limitations when sharing results within the organisation. Do not turn a suppressed row into zero or a participant percentage into a share of all employees. A walkthrough can clarify definitions and documents without opening anyone's individual insurance overview.
[1]Fictional example
A row that is not shown is not a zero.
An employer can see an overall picture, but a particular part of the statistics is not shown.
- Overall group
- Sufficient data for some statistics
- Individual row
- Not shown
- Correct interpretation
- No published value, not zero
The example says nothing about which people contributed or the value behind the display limit.
Common questions
Can we extrapolate the percentage to the whole workforce?
Not without support for the groups and data being comparable. The participant group is not automatically representative of all employees. State which group the figure actually describes. [2][3]
Can a large group have suppressed rows?
Yes. Display limits apply both to the overall group and to individual parts of the statistics. Sensitive insurance categories are excluded from separate per-line reporting even when the overall group is large. [1]
FOR ORGANISATIONS
Review the information you already have.
Book a walkthrough of the employer view, your registered benefits and the limits that apply to its statistics.
Sources and references
Sources checked
This guide explains general principles. Check the current agreements, roles and permissions for your organisation.
- Forsio: employer aggregate view and display limits
- Statistics Sweden: sampling and uncertainty in a survey
- Statistics Sweden: non-response and statistical bias
- Swedish Authority for Privacy Protection: anonymous information and anonymisation
- Swedish Authority for Privacy Protection: sensitive information and consent at work