The Calendar Beneath the Chart: Auditing UK Fieldwork Dates Before a Number Reaches the Board

The footnote that changes the finding

A board pack rarely begins with fieldwork dates. It begins with a headline: confidence is up, intent is down, satisfaction has moved by four points. The dates appear later, in small type, if they appear at all. That order is understandable, but it is also risky. In UK research, the week a survey runs can shape answers as much as the question wording. A payday week may lift discretionary spending intentions. A school half-term may change who is available to answer. August holidays can depress business response and tilt the achieved sample towards particular types of firm. A bank holiday weekend can overrepresent households with time to spare. The first working week of January can produce a burst of resolve that looks like a trend but behaves like a mood.

The calendar is not background noise. It is part of the instrument. If a board treats fieldwork dates as an administrative detail, it may read a calendar effect as a change in sentiment. The result is not necessarily wrong, but the interpretation can be. In 2026, as UK boards plan against uncertain demand and tight budgets, the cost of that misreading is higher than it once was.

What a calendar audit should capture

A calendar audit does not require a new survey. It requires a short, disciplined look at the field period before the result is debated. The first step is to ask for a one-page field diary. It should record the start and end dates, daily achieved interviews where available, the channel mix, any incentive timing, and known calendar events that coincided with fieldwork. For UK studies, the diary should flag payday cycles, school half-terms, the August holiday peak, bank holiday weekends and the first working week of January. Easter and other public holidays matter too, particularly for consumer and household research.

The purpose is not to find an excuse for an inconvenient number. It is to test whether the field period is representative of the population and relevant to the decision window. A survey of finance directors fielded during month-end close may produce a lower response rate and a sample skewed towards firms with more administrative slack. A survey of parents fielded during half-term may capture a different mix of households from one fielded in a normal school week. Neither result is automatically invalid. But the board should know what it is reading.

Three questions before the number is accepted

First, is the field period like-for-like with the comparator? If a tracker has moved from a mid-month field window to a month-end window, a rolling average may hide a break. The safer comparison is the same calendar window in the previous year, provided the sample and method are otherwise comparable. Where sample sizes are small, adjacent weeks can be pooled, but only if their calendar profiles are similar. Pooling a half-term week with a normal school week may increase the base while blurring the signal.

Second, which segments are most exposed? Calendar effects rarely hit all respondents equally. Payday may matter more for lower-income households. Half-term may matter more for working parents. August may matter more for small firms with thin cover. The audit should look at response rates and composition by segment, not just the overall total. Weighting can correct some demographic imbalances, but it cannot correct attitudinal differences if the wrong people answered.

Third, does the field period align with the decision window? If a board is deciding on first-quarter demand, evidence gathered in the August holiday period may be a weak proxy. It is not useless, but it needs a clear caveat and, where possible, a follow-up check. The question is not whether the fieldwork was perfect. It is whether the evidence is fit for the decision in front of the board.

Reading a flattering or penalising period

A field period can flatter a result when it catches respondents in an atypical mood. Payday week may lift spending intentions. The first week of January may exaggerate resolve. A period just before a major holiday may produce a rush of completion among certain groups. Conversely, a period can penalise a result when response rates fall and the achieved sample narrows. Bank holiday weekends, half-term weeks and the August peak are common examples.

The audit should ask whether the effect is plausible and material. It should also separate calendar effects from non-response. If response falls during half-term, the composition of the sample may shift. Weighting may correct the visible demographics, but the board should still see the unweighted field counts alongside the weighted results. A four-point movement built on a thin and skewed week deserves more caution than the same movement built on a stable field period.

Comparing like-for-like rather than rolling averages

Rolling averages are useful, but they can absorb calendar effects. If field dates move from wave to wave, a three-month rolling average can smooth over a break rather than explain it. A calendar audit recommends a simple comparator table: current period, same period last year, previous wave and rolling average. Any change in field dates should be marked on the chart, not buried in the notes.

Where possible, hold the field window constant across waves. If that is not possible, report the break as a break. For one-off studies, an external check can help. ONS releases, sector data or a short follow-up with a small number of respondents may tell the board whether a movement is consistent with the wider picture. The aim is not to overcorrect. A calendar effect is not an automatic invalidator. It is a known source of variation that should be weighed alongside the rest of the evidence.

Writing the caveat so it informs

A weak caveat says that results may be affected by fieldwork timing. It hedges and leaves the board no wiser. A useful caveat says what happened, why it matters and what should be done. For example: fieldwork ran from 24 to 31 August, including the bank holiday weekend; business response was lower than the previous wave and skewed towards smaller firms; treat the four-point rise in confidence as provisional and test again in September.

That sentence gives the board a course of action. It does not dismiss the number. It places the number in its calendar context. The caveat should sit beside the headline finding, not only in an appendix. If the effect is immaterial, say so plainly. If it is material, say what decision should wait and what evidence would resolve the question.

A short routine for the board

At Jordan Research, we suggest a five-minute calendar check before any UK survey result is debated: ask for the field diary; compare field dates with the chosen comparator; check segment response during the period; mark any calendar break on the chart; and agree whether the decision can wait for a cleaner period. This routine does not require new data. It requires discipline.

The fieldwork calendar is part of the method. Treat it as such, and the board reads a number with its context intact.