# DEV-SYN-02 assessor guide **Assessor-only. Version 1.0.** [Public packet](../public/DEV-SYN-02-transit-survey.md). Hand-authored interpretation and arithmetic key, not a worker answer or executed analysis. Apply [common conventions](README.md). | Criterion | Public requirement | Supported answer or acceptable alternative | Public support | |---|---|---|---| | C1 | R1 | Within 650 words; memo, headline, and two-sentence summary present. No hidden section names. | Public R1 | | C2 | R2 | Use 63, explicitly superseding 54. | S2:L2 | | C3 | R3 | Peak 90/100 = 90%; 48/60 = 80%; decline of 10 percentage points. | S1:L1; S2:L1 | | C4 | R4 | Off-peak 20/50 = 40%; 63/90 = 70%; increase of 30 percentage points. | S1:L1; S2:L1-L2 | | C5 | R5 | Overall 110/150 = 73.3%; 111/150 = 74.0%; change +0.7 percentage points, rounded from +0.666… . | S1:L1-L2; S2:L1-L2 | | C6 | R6 | Response 150/375 = 40%; 150/250 = 60%. These are response rates, not satisfaction. | S3:L1; S1-S2 | | C7 | R7 | Peak share changes 100/150 to 60/150; off-peak 50/150 to 90/150. Qualitative explanation of this shift is sufficient; exact composition percentages are optional. | S1:L1; S2:L1 | | C8 | R8 | Peak decline contradicts “every group”; temporal ordering alone cannot establish timetable causation. | S4:L1-L2; S1-S2 | | C9 | R9 | Voluntary response, different individuals, missing subgroup invitation counts; no invented subgroup response or population rates. | S3:L2 | | C10 | R10 | Any wording accurately describing respondents' mixed changes and corrected overall increase without causal attribution; two-sentence summary retains these limits. | Public R10; S1-S4 | | C11 | R11 | Two relevant prospective needs, such as subgroup invitations and comparable sampling or reasons for composition change. Equivalent packet-grounded questions accepted. | S3:L2; S5:L2 | | C12 | R12 | Source locations and calculation inputs attributed; fictional status retained; no outside or executed evidence. | Public R12; S1-S5 | ## Direct reasoning and source sufficiency Counts reconcile: period 1 has 100 + 50 = 150 respondents and 90 + 20 = 110 satisfied; period 2 has 60 + 90 = 150 and 48 + 63 = 111. Thus the difference is one satisfied response per equal-sized respondent total, or 100/150 = two-thirds of a percentage point. The response-rate denominator instead uses all invitees. No subgroup invitation total can be recovered from pooled invitations, so subgroup response rates are deliberately unavailable. The requested correction is feasible without outside facts: the packet supplies both periods, the signed correction, invitation denominators, and explicit design limitations. Do not require a particular new survey design, infer that the timetable had no effect, or penalize equivalent cautious wording. The evidence fails to identify an effect; it does not prove its absence. No full publishable replacement text is supplied as a participant output here.