Perceived Safety

Outcome

Impact

% of individuals / target population in a given context reporting an improved feeling of safety (with dignity) by the end of the intervention compared to at the beginning

This DG ECHO Key Outcome Indicator measures a protection outcome: the proportion of direct beneficiaries who report that their feeling of safety (with dignity) has improved by the end of the action compared to the beginning.

The DG ECHO methodology is described in detail in the ECHO guidance document below. Given its complexity, this page provides a summary of the key steps to help you navigate the calculation of the indicator value.

 

1) Start with a protection risk analysis:

• Review or conduct a protection risk analysis (using assessments, KIIs/FGDs, protection monitoring, etc.) to identify the main protection risks affecting the target group. Make sure it reflects differences by sex, age and disability at minimum.

• From this analysis, list the concrete situations in which people feel unsafe or experience unsafe treatment (e.g., when travelling to services/markets, collecting water, moving at night, crossing checkpoints, accessing documentation). This is what ECHO means by “qualifying and contextualising” safety.

 

2) Select what your action can realistically influence:

• For each key risk / situation identified in Step 1, note whether your activities are intended and capable of reducing that risk. ECHO expects this link between risk analysis → activities → survey questions / indicator focus.

• Select those risks / situations that (a) are important for people’s safety, and (b) your activities are likely to influence effectively. These will become the focus of your baseline / endline questions for this indicator.

 

3) Define who will be surveyed:

• People’s feelings of safety in relation to the prioritised risks / situations should be measured through a survey. The survey sample must be representative of the direct beneficiaries of the protection activities addressing those risks.

• ECHO requires results to be disaggregated at least by sex, age and disability. Because some groups may be under-represented in the beneficiary population, you may need to increase the sample size beyond the standard 95% confidence / 5% margin of error in order to obtain sufficiently robust subgroup results.

 

4) Formulate baseline questions:

For each selected situation (from Step 2), write a baseline question asking how safe people currently feel in that particular situation. For example: “How safe do you feel when you must go to fetch water to the nearest water source?”

For answers, use a non-comparative Likert scale that captures how safe people currently feel in the selected situations: I feel very safe / I feel somewhat safe / I feel neither safe nor unsafe / I feel somewhat unsafe / I feel very unsafe / I don’t know / Prefer not to answer. If a respondent answers “somewhat unsafe” or “very unsafe”, include a follow-up question asking why to provide additional insights into your risk analysis.

 

5) Formulate endline questions:

For endline, prepare questions that explicitly ask about changes in people’s feeling of safety in specific situations.

ECHO requires including a non-leading attribution element linking the perceived change to the project. However, the questions suggested in its guidance increase the risk of social desirability, courtesy, and attribution bias. To reduce the risk of such biases, IndiKit suggests using the approach listed below.

First, ask a comparative question regarding each prioritised risk / situation with no mention of the project or its activities. For example: “Compared to [specify the time before your protection support started], how safe do you feel now when [specify the situation]?” For responses, use a Likert scale: I feel much safer / I feel somewhat safer / I do not feel any safer / I feel less safe / I feel much more unsafe / Don’t know / Prefer not to answer. If a respondent answers “less safe” or “much more unsafe”, include a follow-up question asking why.

Then, ask about the attribution: What do you think are the main reasons for this change (or lack of change)? Allow multiple responses, which might include:

   - support or activities provided through this project

   - actions by other organisations or authorities

   - changes in the general situation (e.g. security, economy, environment)

   - personal or household circumstances

   - no change occurred

   - other (please specify)

The enumerators must be instructed and trained to ask non-leading probing questions to reduce the risk of underreporting the benefits of the project's support (or other responses).

 

6) Conduct interviews with a representative sample of your direct beneficiaries, ensuring proportional representation of sex, age, and disability groups (in line with what you defined in Step 3).

 

7) For endline survey only: Count respondents who report feeling “much safer” or “somewhat safer” across all surveyed risks and whose explanation links the change to the project, either by mentioning:

   - specific project activities (e.g., improved lighting, separate queues for women and men, closer water points); or

   - by generally acknowledging the presence / support of the organisation

The respondent does not need to explicitly name the project or identify who provided the support.

Enumerators should therefore be trained on the project’s specific activities and outputs to classify answers under “Support or activities provided through this project” when the respondent describes an improvement that corresponds to what the project delivered. Enumerators should not suggest that the project caused the change; they should ask neutral prompts (e.g., “What changed?” “What makes it safer?”) and then code the response consistently.

 

8) To calculate the indicator value:

   - count respondents who report feeling “much safer” or “somewhat safer” and whose reason is linked to the project (via specific activities or general acknowledgement of the organisation’s support)

   - divide it by the number of interviewed respondents (exclude those who refused to answer or didn’t know)

   - multiply the result by 100 to convert it to a percentage

Note: The indicator value is calculated using endline data only and is not derived from baseline results (see Important Comment 1 for how baseline data are used).

The data can be disaggregated by sex, age and disability groups. Report both percentages and absolute numbers for each group. If you expect small subgroup sizes (e.g., persons with disabilities, older persons), plan a larger sample to reduce imprecision and allow meaningful subgroup comparisons.

1) Use the DG ECHO guidance below as your primary reference. The IndiKit guidance only summarises the main steps in a simpler, easier-to-follow format.

2) Under DG ECHO’s PKOI methodology, the indicator value is calculated from endline responses only (i.e., the % reporting they feel “much/somewhat safer”); baseline data are not used in that calculation. Baseline data are still important because they:

   - document the starting situation;

   - help verify that the selected “safety (with dignity)” situations are relevant; and

   - provide the context needed to interpret endline results credibly

 

3) Participatory protection risk analysis is essential for using this indicator and measuring changes in perceived safety (with dignity). It helps you identify (i) who is at risk, (ii) from what/whom, and (iii) why. You can draw on existing resources such as the Protection Mainstreaming Toolkit (see page 28) and any available secondary analyses (e.g., regional or national analysis produced by the cluster). The prioritised risks should be defined primarily on the basis of the target population’s perceptions and experiences. Ensure the analysis covers not only factors that threaten physical safety, but also those that undermine dignity. See the examples below of two safety and two dignity issues:

   - Girls and women in the camps feel uncomfortable about going to bathrooms after dark.

   - Children are unsupervised because their parents are looking for income generating activities. 

   - People have to wait for distributed aid in the sun on a hot day.

   - People receive food assistance instead of cash where markets are still working and people want and could shop.

 

4) Measure what you can influence only: Ask about specific prioritised risks / situations your action aims to address (not “overall safety”).

 

5) Enumerator training is critical: Train enumerators on neutral probing and consistent coding, especially for mapping respondent reasons to project-supported changes.

 

6) At baseline, the guidance above suggests using “current safety” scale (e.g. very safe → very unsafe) because there is no meaningful reference point for comparison at the start of the action. The purpose of the baseline is to describe the starting situation and confirm the relevance of the selected risks and situations. At endline, a comparative “safer than before” scale, as recommended by ECHO, is used to capture perceived change over time. Because the baseline and endline scales measure different constructs (current level vs perceived change), baseline results are used for contextual interpretation only and are not used to calculate the indicator value or for direct numeric comparison with endline results.