1. Purpose of this summary
This document reports the descriptive results of a synthetic community safety survey. It is written the way a real summary would be: what was received, what was done to it, what it shows, and what it does not support.
2. The dataset as received
A single export of 441 rows was supplied, with 28 variables. Before analysis the file was checked for duplicates, out-of-range values, inconsistent coding and missing data. Every change is listed below, and the original file was left untouched.
| Issue found | Rows | Action taken |
|---|---|---|
| Exact duplicate submissions | 17 | Later copy removed, first retained |
| Test entries from pilot | 9 | Removed, flagged in log |
| Age recorded as 0 or above 110 | 3 | Set to missing, not imputed |
| Ward name spelled four different ways | 62 | Standardised to one spelling |
| Sex coded as 1/2 and M/F in same column | 188 | Recoded to a single scheme |
| Rows retained for analysis | 412 | Analysis base |
Nothing was imputed. Where a value was missing or implausible it was set to missing and excluded from the relevant table, with the base stated each time. This matters: a reader who does not know the base cannot tell whether a percentage is meaningful.
3. Perceived safety after dark
| Response | Count | Per cent |
|---|---|---|
| Very safe | 54 | 13.1 |
| Fairly safe | 143 | 34.7 |
| Neither safe nor unsafe | 61 | 14.8 |
| Fairly unsafe | 102 | 24.8 |
| Very unsafe | 48 | 11.7 |
| Not answered | 4 | 1.0 |
| Total | 412 | 100.1 |
Percentages are rounded to one decimal place and do not total exactly 100.
Combining the two negative categories, 150 respondents (36.4 per cent) reported feeling fairly or very unsafe after dark. Just under half, 197 respondents (47.8 per cent), reported feeling fairly or very safe.
4. Perceived safety by sex
The cross-tabulation below excludes the four respondents who did not answer the safety question, giving a base of 408.
| Sex | Respondents | Feels unsafe | Per cent |
|---|---|---|---|
| Male | 198 | 61 | 30.8 |
| Female | 210 | 89 | 42.4 |
| All | 408 | 150 | 36.8 |
In this synthetic dataset a higher proportion of female respondents reported feeling unsafe, a difference of 11.6 percentage points. Whether a difference of that size would be meaningful in a real survey depends on how the sample was drawn, which is addressed below.
5. What this summary does not show
This section is not a disclaimer. It is the part of a data summary that stops a reader drawing a conclusion the data cannot carry.
- These are perceptions, not crime. The survey measures how respondents say they feel. It does not measure how much crime occurred, and the two can move in opposite directions.
- Nothing here establishes cause. Descriptive analysis shows what is in the dataset. The difference between male and female responses is a pattern in the data, not an explanation of it.
- Sampling governs generalisability. If respondents were not selected randomly, these figures describe the people who answered and cannot be projected onto the wider population. No confidence intervals are reported here for that reason.
- One point in time. A single round shows no trend. Any statement about things getting better or worse needs at least a second comparable round.
- Non-response is not random. People who feel least safe may be less likely to answer, which would pull these figures in a direction the data cannot reveal.
6. What a client receives
For a real assignment this summary would be accompanied by the cleaned dataset in an agreed format, the full cleaning log, the tables as a spreadsheet for reuse, and the charts prepared for the client's own report template.