Comparing Condition Scores When the Rating Scale Has Changed
Compare condition surveys across changed rating scales by preserving definitions, assessment methods and evidence before claiming improvement or decline.
Can this year's condition scores be compared directly with last year's scores?
Only after checking that the scales, definitions, assessment scope and scoring methods are comparable. Preserve both original scales and build a reviewed translation where justified. If the methods differ too much, report a break in the trend and compare underlying observations instead of presenting changed numbers as evidence of physical improvement or deterioration.
A facilities dashboard may show a building improving from a score of four to a score of two after a new consultant takes over. The apparent improvement could be real, or it could reflect a reversed scale, revised category definitions or different weighting. A credible comparison begins with the assessment method before interpreting the direction of the number.
Recover the definitions behind each score
Obtain the rating key, assessment instructions and report version for each survey period. Record whether a higher number means better or worse condition and whether scores describe the whole asset or selected components. Check how assessors handled inaccessible areas and missing evidence. A numeric field without its definitions is not a complete assessment record.
Compare the observations required for each category. A label such as fair may cover different defects or service limitations in the two methods. Ask the technical reviewer to identify where categories overlap and where they do not. Do not assume that similarly named categories are interchangeable simply because both surveys use a five-point scale.
Build a defensible comparison method
If a translation is feasible, create a crosswalk showing the original category, proposed comparison category and rationale. Keep the raw scores unchanged. Use the translation in a separate reporting field so reviewers can see how the trend was produced. Document categories that cannot be mapped reliably and prevent them from being silently assigned an average.
Review changes in aggregation as well as labels. One survey may use the worst component score while another calculates a weighted average. Those methods answer different questions. A building can appear to improve when the calculation changes even if no component has improved. Compare selected underlying defects and photographs to test whether the headline trend is plausible.
Explain the limits in management reporting
Show the method change beside the trend, with the periods affected and the approved comparison basis. Where comparability cannot be established, start a new baseline rather than joining incompatible series with a continuous line. Management can still use the latest assessment for current decisions while understanding that a year-on-year percentage change is unsupported.
Retain the scales, crosswalk, reviewer and sample checks with the dataset. Require future survey briefs to specify whether continuity with the existing method matters. If a better method is adopted, plan the transition deliberately and capture enough overlap evidence to support comparison where practical. Consistent interpretation is more valuable than an uninterrupted but misleading chart.
Practical Example
Illustrative example: an estate previously used one as good and five as poor, while a new survey uses the reverse direction and changes the definition of the middle categories. The team reverses no values automatically. A technical reviewer supports comparison at broad good, intermediate and poor bands, and the report states where finer trends cannot be defended.
Action Checklist
- 1.Obtain each survey's rating definitions, scope and aggregation method.
- 2.Check scale direction and treatment of missing or inaccessible observations.
- 3.Preserve raw scores and document any approved comparison crosswalk.
- 4.Test the apparent trend against underlying defects and dated evidence.
- 5.Disclose a method break where reliable comparison cannot be established.
Make condition trends traceable to their assessment evidence through Audit-Ready Asset Management.
