Through the Window

Posted on

Looking at the Numbers · WHS Elders Consultation Survey (Download the PDF version | Download the PDF Summary | Download the Committee Report and Survey Results)

Earlier, I wrote about the case for building the door — before we change the lock — and this rested on what we could learn from other churches and secular wisdom. The data from the consultation survey lets us stop guessing and look: it confirms the door has to exist. This is an argument about impact, not the theology of eldership. Make the change if it must be made — but Scripture, and our own tradition, ask us to build what it requires first.

Let me say at the outset what this is not. It is not a case for or against male-only eldership. Someone can be a convinced complementarian, sure the change is biblically right, and still owe an honest account of what it will actually do. Deciding whether to act and seeing clearly what acting will cost are different things, and the second doesn’t go away once the first is settled. The survey was never a referendum on the theology, and neither is this. It asks one question: if the change is made, what does the denomination’s own data say will happen to its churches — and is it ready for that?

The committee’s report opens with a reassurance: 91% feel safe always or most of the time, steady whether or not a congregation has women elders. That is good news, and we should want it to be true. It is also worth being clear about which question it answers. Nobody proposed this change expecting people to feel unsafe the next morning; how safe people feel on an ordinary Sunday was never really what was in dispute. The worry raised about the change was structural — what avenues exist, who gets heard, whether someone would speak up when something goes wrong. That kind of risk doesn’t show up in a question about present-day safety. It shows up in what people say they would do: whether they would report, whether they feel supported, what the churches that have actually lived this arrangement say about it. The safety headline is true and reassuring. The questions closer to the actual concern sit a row or two down in the same tables, and they read less reassuringly — which is why the rest of this looks there.

First, how big is this response?

Before reading the signal, it’s worth knowing whether 1,787 voices is a lot or a little — because both cheap dismissals are wrong, and naming the real base disarms them. NSW has roughly 153,000 people who tick “Presbyterian and Reformed” on the census. But census identity is a cultural ghost, not a churchgoer. The PCNSW is a community of about 35,000; of those, around 14,000 adults are in average attendance across 183 congregations. Those are the people a survey about church life can actually reach.

What 1,787 is measured againstFigure1,787 is…
NSW census “Presbyterian and Reformed” (2021)~153,000~1% — the affiliation ghost, not the church
PCNSW “community” (self-described)~35,000~5%
Average adult attenders~14,000~13% — roughly one in eight
Congregations represented117 of 184~64% of all charges

Census figure from ABS 2021; attendance and congregation counts are the denomination’s own current figures. Measured against the people who actually attend, this is a substantial, denomination-wide response — not a fringe, and not “barely 1%.”

So set both dismissals aside. This is not a vocal minority — one in eight regular attenders, across nearly two-thirds of congregations, is a serious response by any church-survey standard. Nor is it a census; at 13% of attenders it is exactly the size where who answered decides what the numbers mean. That is the hinge the rest of this turns on.

One caveat sits over everything below. Every figure here — the survey responses, and the denomination’s own counts of elders, attenders and vacancies — comes from partial returns. The survey heard from some attenders, not all. The denominational counts come from the congregations that reported, not every charge. And like every voluntary consultation, this one is subject to self-selection: those with strong views, either way, are likelier to answer. That doesn’t make the sample random. But two things make a small mobilised faction an unconvincing explanation for what it found. The first is sheer volume: around one in eight adult attenders responded, a high rate by any church-survey standard, and a fringe is rarely that large. The second is internal consistency — the same ordering across differently-worded questions, which a simple over-representation of the agitated on one side would not produce. Neither makes the result a population proportion, and I don’t read it as one; together they make it hard to dismiss as noise. So I’m reading what the respondents said, not pronouncing on settled populations — and the reading doesn’t need a precision the returns can’t give.

The survey’s own demographic chart shows roughly how the 1,787 split by role. Reading it off the page gives approximately:

Who answered (approximate)Share~Respondents
Members, regular attenders, volunteers~88%~1,570
Elders~6%~110
Ministers / pastoral workers / CGW~3%~50
Paid staff (incl. pastoral assistants)~1.5%~27
Deacons~1.5%~27

Read off the survey’s demographic graph, so approximate. The lay cohort is the bulk of the response; the office-bearer columns are smaller, and two of them are small enough to need care.

How to read these numbers before we start

A percentage from a small group is not a precise point; it is a band, and the width of the band comes straight from how many people answered. When only about 27 people are in a column, each one is a twenty-seventh of the total — one person in twenty-seven is nearly four points — so if just a handful had answered differently, the figure would swing several points at once. That movement is what makes the band wide: around nineteen points either way. Put more people in the column and each one counts for less. One in 110 elders is under a point; one in the 1,500-plus lay respondents is barely a twentieth of a point. So those figures hardly move, and the band tightens — to roughly nine points for elders, and just two or three across the lay column. Same survey, same logic: the more people answered, the less any one of them can shift the number, and the steadier it is.

Statisticians call that band a 95% confidence range. Read the survey this way and the true figure should fall inside the band about nineteen times in twenty. So “47% of paid staff expect leadership to deteriorate” really means “somewhere in the region of 28% to 66%” — a wide band, and an honest one. That isn’t a flaw in the survey; it is just what small samples do, and once you read the figure as a band rather than a number, it behaves honestly.

Here is the part that matters. The useful question is not “what is the exact figure?” but “which side of the line does the whole band sit on?” Take that same paid-staff figure. Even at the very bottom of its band — 28% — it still outweighs the 17% who expected leadership to improve. The band is wide, but it sits on the same side of the line from end to end. That is what a small sample can honestly tell you: not the precise number, but the direction, when the direction holds right across the margin.

So read the small columns — paid staff and deacons, about 27 each — for which way they lean, not for their decimals. The precision lives in the big columns: over 1,500 lay respondents and around 110 elders, where the figures can be taken close to face value. Nothing below asks a 27-person column to carry a number on its own. It asks only whether each column leans the same way as the large ones — and for that, the small samples are entirely good enough.

One more thing worth reading off the survey’s charts before the questions: who answered, by gender. The response was not heavily distorted by either gender — overall it ran about 56% female to 42% male, and the columns that carry most of the concern are mixed, the lay respondents around 60/40 and the paid staff about 52/48. Both men and women cared enough about this to answer, which means the sample is a reasonably representative cross-section of the church rather than a single-gender bloc. That matters for how you read everything below: a negative result here is not simply one group, unhappy at losing something, speaking for itself. One limit to be honest about — there is no published breakdown of answers by gender, so we can’t see whether, on a given question, one gender in a particular role was markedly more concerned than another. That cut would be worth having. But even without it, the balance of who responded tells us this is not the voice of one gender alone.

The committee’s own summary of the results notes that the responses were not drastically different by gender, age or role, and that the sharpest variation was whether a congregation currently has female elders. That is fair, and worth holding alongside what follows. The point below is not that the roles split wildly from one another — on most questions they don’t. It is narrower: one cohort, the ministers, sits apart from all the others, who cluster together. Reading for that single break is not in tension with the committee’s summary; it is a closer look at the one place the otherwise broad agreement parts.

For orientation on the offices themselves: the denomination’s 2026 snapshot records around 700 elders (about 12% female), 178 ordained ministers and ministry workers plus 26 commissioned gospel workers, and 8 deaconesses. The office of deacon isn’t given a separate figure in the published data, which is why I read the deacons’-court column for direction rather than against any office total. So the ~50 ministers who answered are about a quarter of their office; the ~110 elders a useful slice of theirs; and the lay column simply enormous. Hold those rough sizes in mind — they shape what each survey column can and can’t tell us.

The pattern the pie charts wash out

Take Q8 — whether excluding women from eldership would affect the quality of leadership. Read by role, one cohort stands apart from all the others. Ministers expect deterioration at 25%; everyone else — paid staff, elders, deacons, members — sits between 47% and 57%. The elders are not the unconcerned ones here: a clear majority of them, 54%, expect the quality of leadership to fall. So do the deacons in their pastoral role, and the members in the pews. It is the ministers, and the ministers alone, who break away from the rest:

Q8 — “Will affect leadership quality”Minister/
PW/CGW		Paid StaffElderDeaconMember/
Adherent
Yes, positively (will improve)27%17%14%3%12%
Yes, negatively (will deteriorate)25%47%54%57%56%
No change45%28%29%29%23%
Unsure2%8%4%4%10%

Ministers are the outlier on the low end; every other role clusters high. The gradient — not the average — is the finding.

This is the single most important shape in the data, and the aggregate pie chart erases it. The deacons’-court column sits at the top, and the four non-minister columns move together — a single ordering that the “All Respondents” pie cannot show.

And it isn’t one question. The same ordering recurs across the survey, which is what turns a curiosity into a finding:

“Negative / increase / less likely” by roleMinister/
PW/CGW		Paid StaffElderDeaconMember/
Adherent
Q9 — significant negative culture impact22%28%37%53%43%
Q10 — yes, impacts relationships/teams49%61%61%74%68%
Q11 — significant negative wellbeing impact11%17%27%33%24%
Q13 — increase in conflict/bullying/harassment24%24%40%47%31%
Q14 — staff/volunteers less likely to report20%31%33%53%41%

Five differently-worded questions. The same role ordering each time: ministers lowest, deacons highest, everyone else stacked in between.

Why the pattern is the finding, not the noise
The same cohort ordering turns up across five differently-worded questions. A one-off result might be a quirk of a small sample; the same ordering recurring five times over is much harder to put down to isolated variation. When deacons, elders and members all point in the same direction on leadership, culture, relationships, wellbeing, conflict and reporting, while ministers sit lowest each time, that consistency is the evidence on its own — whatever the precision of any single percentage. One figure might sit anywhere within a margin. A pattern that holds across five differently-worded questions is not so easily explained away.

Reading honestly in both directions

First, a discipline worth holding to, since the argument only earns its weight by observing it: not every question is a warning. Some results are genuinely divided, and a divided result isn’t a negative one. Where about as many expect improvement as expect harm, the honest word is “split,” and it would be wrong to count the negative half while setting the positive half aside. So here is each question sorted by what it actually shows. Naming the split ones as split is what makes it fair to name the lopsided ones as lopsided.

QuestionPositiveNegativeReads as
Q8 — leadership quality (non-minister roles)3–17%47–57%Lopsided — negative ~3× positive
Q9 — church culture (non-minister roles)14–22%56–70%Skewed negative
Q11 — personal wellbeing (ministers)25%20%Genuinely split — near wash
Q13 — conflict / bullying / harassment3–11%24–47%Lopsided toward increase
Q14 — willingness to report3–9%20–53%Strongly asymmetric

Granting the washes doesn’t weaken the case; it clarifies it. Ministers’ own wellbeing answer (Q11) is close to even — said plainly, because it is true. The reporting and conflict questions are not even close.

That is the thing to hold onto. The argument doesn’t need every number to lean one way, and several don’t. It rests on the questions that bear on structure and safety — reporting, conflict, the loss of a safeguarding voice — and those are the ones that aren’t washes. A balanced question elsewhere doesn’t settle those, because the balanced questions are not the ones carrying the weight.

What each role’s numbers actually carry

Start with the ministers, because they sit at the low end of the gradient and it would be easy — and wrong — to wave them away. The survey’s ministry-workers column — about 178 ordained ministers and ministry workers in the denomination — drew around fifty responses, about a quarter of that office. That is a fair slice, not a thin or self-selected sliver. So the low minister concern is real, decently sampled, and sincerely held, and I am not going to explain it away. What makes the ministers worth pausing on isn’t that they are unrepresentative. It is that they are reading this change very differently from everyone else who answered — and they sit, with the elders, among those who will vote on it. That is the thread the next section picks up.

The elders sit on firmer ground than the small office columns, and for a simple reason: more of them answered. Around 110 elders responded — far more than the ministers, deacons or paid staff — and that larger sample tightens the band from the ±19 of a 27-person column to something nearer ±9. So where the small columns can be trusted only for direction, the elder figure can be read closer to its face value. Of the elders who answered, around 54% expect leadership quality to deteriorate. No special pleading is needed for that one. It is the most ordinary number in the whole exercise — a clear majority of a decent sample, read at something like its actual value rather than as a wide band — and it leans the same way as the rest.

Deacons need care for one straightforward reason: the size of the sample. About 27 people identifying with the deacons’ court answered — a small base, read as a band, not a figure. That is enough to read which way the column leans, not enough to trust its decimal. And of those who answered, the lean is the same as everyone else’s: around 57% expect things to get worse. Don’t take the 57% to the decimal, and don’t bin the column either. Read it for direction, and the direction puts it with the rest — in fact at the head of it.

Are we all answering the same question?

Here the data turns from broadly agreed to genuinely striking — though not in the way you might expect. The ministers who answered are the outlier, lowest concern on every question, while elders, deacons, members and paid staff all read it as carrying a real cost. That gap is remarkable precisely because it is so disjointed: the people who lead these churches and the people who fill and serve them are looking at the same proposed change and reporting very different things about what it will do.

It is worth asking why, and one possible explanation is worth offering — tentatively, because it concerns how people read a question rather than anything the survey directly measured. Some who support the move to male-only eldership see it as a matter of critical biblical obedience: not a question of cost and benefit at all, but of faithfulness to Scripture. Read that way, the impact question may land differently — a thing one believes is commanded is a thing the mind is less inclined to weigh a price against. Offered as a possibility, not a diagnosis, and not of any individual, it would explain the disjoint without anyone being foolish or self-serving: one group may be answering a question about obedience while the others answer a question about cost.

And obedience is good; you do not refuse to obey because obedience is expensive. If those who support the overture are right that this is commanded, proceeding despite the cost may be exactly faithfulness, and I am not claiming to have settled that. But counting the cost and obeying anyway is not the same as obeying and seeing no cost — and we may not all agree which of those faithfulness requires. The conviction that something is right and the observation that it will be costly are not rivals; a church can hold both, and most of the survey’s respondents do.

This matters for the decision, though not in the way it would if ministers simply outvoted everyone. They don’t. At Assembly the floor is, in principle, shared: each charge is represented by a minister and an elder, two voices of equal standing in the vote. So the room that decides is not the low-concern reading writ large. It is split down the middle — and the elder half of it is among the most concerned in the whole survey, with a clear majority expecting the quality of leadership to fall. The disjoint the data found across the church runs straight through the body that will vote on it.

That is the thing worth sitting with. The elders who see a cost are not petitioners outside the room; they are full voters inside it, with every right to carry what they have seen into the decision. And the ministers, who are half that voice, have reason to listen hard to the other half — because the people who would absorb this change are telling them, in the church’s own consultation, that they expect it to be felt. The survey gave everyone an equal say; the floor gives ministers and elders an equal vote. On the evidence here, both halves of that floor have cause to take the concern seriously, and the elders in particular have the standing to act on it.

Paid staff: the people no one disputes are covered

It is worth pulling paid workers out and looking at them alone, and not out of sentiment. The Work Health and Safety duty reaches further than paid staff: in law a “worker” includes volunteers, so a great many more people in the life of a church are covered than the payroll alone. Some will argue the church shouldn’t have to count volunteers — but that is an argument with the legislation, not with this piece, and I don’t need to win it. Set the whole question of volunteers aside. Even a reader who thinks the survey cast too broad a net by asking the entire church still has to reckon with the paid staff, because about them there is no dispute at all: they are employees, and there are a lot of them — 311 paid through central payroll in 2026, up from 292 the year before and 262 the year before that. Whatever your theology of eldership, they sit inside the WHS framework this consultation was convened to address. An elder’s discomfort is a governance matter. A paid worker’s psychosocial hazard is a duty owed by a person conducting a business or undertaking. The survey gives them their own column, and it repays reading.

Paid Staff — selected responsesFigureReads as
Q8 — change will deteriorate leadership47%Near half
Q9 — negative culture impact56%28% “some” + 28% “significant”
Q10 — yes, impacts relationships/teams61%Clear majority
Q11 — some/significant negative wellbeing impact37%20% + 17%
Q14 — less likely to report issues31%Nearly a third

Paid staff also reported the highest “always feel safe” figure (72%) — safety and support are not the same thing, a distinction many respondents drew explicitly.

And paid staff are not the cohort with the governance vote, nor mostly the female-elder congregations the report flags as over-represented. They are the people who turn up to work in these churches. Their column is small — about 27 answered — so read its figures as bands, not points, and let the legal principle carry the weight rather than any single percentage. The principle doesn’t need a large sample: paid staff are employees, and the duty is owed whether 27 of them answered or 270. When a column like theirs leans, as it does here, toward expecting people to be less likely to report conflict, bullying or harassment, that lean matters. The national model code of practice on managing psychosocial hazards — the more recent of the two, even if NSW hasn’t yet adopted its wording — recognises that this kind of conduct is routinely under-reported, out of fear of blame, escalation or reprisal, and that an absence of reports doesn’t mean nothing is wrong. It treats under-reporting as a marker of poor culture and a barrier to seeing hazards at all. The survey contained responses consistent with exactly that concern, from people the framework plainly covers — and they sit below the headline rather than beside it.

The finding that deserves the headline
A consultation conducted under the banner of work health and safety — convened because of a safety concern — produced data in which, of those who answered, a plurality of several cohorts (around 53% of deacon respondents, 41% of members, 33% of elders, 31% of paid staff) expect people to become less willing to report conflict, bullying or harassment if the change is made, and that all-male sessions would be worse at seeing and handling it. In a church on the other side of a Royal Commission that examined exactly how unchallengeable authority and barriers to disclosure let harm continue, that is not a detail. Every safe-ministry structure the denomination has built assumes people will report. The summary leads with “91% feel safe,” which is true; but the reporting result is the one a safety process most needs to see, and it is worth bringing up to the light beside it.

Where the summary and the sharper reading part

The committee’s summary is accurate, and nothing in it is false. But a summary makes choices about what to foreground, and these choices lean a consistent direction — toward the reassuring reading, with the sharper edges set a little further back. Five places show it.

  • The hardest comments are summarised, not shown. The trauma and abuse responses at Q9 are withheld “to ensure such responses are not identified” — a fair privacy rationale, but the effect is that the most serious material is the least auditable. The same is true of the “don’t trust current male leadership” subset at Q8 and the female elders who said they would feel ruled “unbiblical/evil.” We are told these exist; we cannot weigh them.
  • Safe is quietly substituted for supported. The committee notes “in passing” that a large number of respondents marked themselves safe but not supported. That distinction is buried under the 91% headline, when it is arguably the more important of the two for a WHS frame.
  • The sample skew is acknowledged, and it is worth testing rather than waving past. The report notes that female-elder congregations are over-represented (20% of churches, 31% of responses) and that the strongest predictor of perceived risk is whether a church currently has women elders. That is a fair caution against reading the raw negatives at face value, and it deserves a real answer rather than a rhetorical one. So here is the test — illustrative rather than a formal weighted analysis, but enough to show which way the correction cuts. The over-representation is an excess of about 11 percentage points (31% against an expected 20%). Take the least favourable assumption: that the entire 11-point excess would vanish from a properly weighted sample, and that it fell uniformly on the negative side — the worst case for reading the data as concerning. Strip that weight out of the pooled figures and the leans still hold: a 54% or 57% drops toward the mid-to-high 40s — still a plurality, still elevated, still pointing the same way. The finding was never a bare majority clinging to the 50% line; it was a wide and consistent lean, and an 11-point adjustment on the most-exposed slice doesn’t erase it. And the correction touches only the pooled aggregate. The role gradient is computed within each office column — it is a property of which office a respondent holds, not which church-type they attend — so the over-representation doesn’t enter it at all. The skew affects the “All Respondents” pie charts; it leaves the gradient, where the real signal lives, untouched. Run the test the caution invites, and the reading holds.
  • The same skew cuts the other way, too. The female-elder churches answered at about one and a half times their share, and it is worth asking what that over-representation contains. These are the only respondents speaking from current experience of the arrangement the change would remove. Others write from past experience, or from watching other churches; these congregations have women in session today, and would feel the loss directly. So their responses may reasonably be read as particularly informative — not because they should carry extra statistical weight, but because they reflect lived experience of the very arrangement under discussion. The correction above already shows the leans survive even if their excess is stripped out entirely. What this adds is only that, set alongside the correction, their testimony is the survey’s closest thing to first-hand evidence about what removal would mean — and that is worth noticing, not discounting.
  • The aggregate is foregrounded; the role split is “texture.” Every flagship chart is “All Respondents.” The role tables are present but never assembled into the cross-question comparison that reveals the gradient. The structure of the document makes the finding hard to see.

What the change closes has a size, and it’s 12% today

Until now the place women hold in the courts of the church — the standing by which they presently sit in session as of right, rather than by an elder’s invitation — has been an abstraction for many of us in this debate. Not because the figure was unknowable; the numbers were there for anyone who went looking. It is that, like much else here, few had gone looking, and fewer still had put it in front of the people arguing the question. The official figures let us put a size on it. The denomination’s 2026 snapshot records about 12% of elders as female; against an estimated 700 elders that is on the order of eighty women — roughly one in eight — though the firmly-counted figure, from the congregations that actually returned their forms, is a little lower, around sixty-six. Call it somewhere between seventy and eighty-five, denomination-wide. Whatever the exact number within that band, it is small, and the proposed change grandfathers those women while closing the door to any more.

And the narrowing is not only a future event waiting on the vote. It has, in places, already begun. Female-elder churches are a minority to start with — about one congregation in five — and even among those, some will not appoint new women in anticipation of the change, while in some churches women who would serve as elders already cannot. None of that is necessarily wrong: if the theological frame is male-only eldership, this is simply that conviction being lived out ahead of the formal decision, and the people doing it are acting in good conscience. The point is only that we cannot discuss this standing as though it were intact and waiting. It is contracting now, by retirement and by choice, and every female elder lost — to relocation, to age, to a session that won’t replace her — is, under the proposed rule, unreplaceable. The number is small, and it has already started to fall.

That reframes what the female-elder churches are telling us. When those congregations say they would lose something real, we can now put a rough number on the something: somewhere around eighty women, about one in eight elders, with no way to replace them under the proposed rule and only one direction to travel. These are also the only respondents speaking from current experience of what is at stake — women in session now, not a memory of them — which is about as close to evidence as a survey of anticipated effects can get.

“Not enough men” was not hypothetical

Several respondents — ministers and members both — raised that in some congregations, particularly regional ones, the people suitable for eldership are sometimes women, and removing them would leave those churches short of leadership altogether. That concern is not idle. The denomination already carries real leadership-supply stress: of 183 pastoral charges, around 19% are genuinely vacant once session and CGW appointments are counted, and three-quarters of the un-appointed vacancies are regional. Whether those are the same churches the eldership change would touch is not something the figures tell us, and this piece does not claim it. But a denomination already short of available leaders, regionally, has less margin to absorb the loss of any eligible candidate — which is enough to make the “not enough men” concern a live structural risk rather than a hypothetical.

The church reading, kept honestly

None of the above settles the theological question, and it isn’t trying to. A great many respondents who expect a negative impact say so while affirming that the change may still be the right and biblical one — the elder who writes that obedience to Scripture and leadership quality are simply different axes is making a coherent point, not dodging. The survey was explicitly not a referendum on whether the change is correct, and this piece isn’t one either. A complementarian can hold that male-only eldership is commanded and that the data here describes real costs that fall hardest on women, on paid staff, and on the churches that will have to absorb the change. Those two things are not in tension; the survey is full of people holding both at once.

What the data does ask of the church is narrower and harder to refuse. If the change is made, the survey is close to unanimous on one point across every role and every theological camp: it cannot be made well without deliberate structures — protected avenues for women’s voices in the courts of the church, strengthened diaconal roles, clear reporting pathways that don’t depend on an all-male session, and genuine support (not merely assurances of safety) for the people most affected. Respondents who support the change and respondents who oppose it converged here. That convergence, like the role gradient, is a real finding — and it is the one the church can act on regardless of where it lands on the eldership question itself.

The three questions, answered

Strip away the framing and there are three plain questions a reader actually wants answered. Here is what the data supports — and, as carefully, what it doesn’t.

1. Are people concerned in ways that actually matter?

Yes — though not “statistically significant” in the p-value sense, because without published response counts nobody can compute that, and I won’t pretend otherwise. The significance that matters here is whether the concern is real and structured rather than noise, and it is. Among respondents it is the majority view in every non-minister column on the core question, around 47 to 57%. It holds across five differently-worded questions, which makes it hard to dismiss as isolated variation. It is shared right across the church — elders, deacons, members and paid staff all clustered high together — and only the ministers who answered sit apart from it. And it attaches to mechanisms rather than moods: reluctance to report, expected conflict, the loss of a safeguarding voice. Those are operational risks a board would minute, not feelings to be soothed.

The floor matters as much as the ceiling. Set the highest figures aside entirely; the argument survives on the lowest. Across every cohort but ministers, somewhere between a quarter and three-fifths expect negative consequences — and even a quarter is too many to wave off. A church that takes safe ministry seriously doesn’t need a majority to warn it. It needs a credible minority, and it has one in every room. That holds whatever you make of the sampling, because it doesn’t lean on the big numbers being precise.

2. Does the summary show what the data shows?

The conclusions are sound — and worth saying so plainly, because the committee’s report does good work: it sets out real options and calls for structures to be built. The question here is narrower, and it is about the summary of the data rather than the committee’s judgement. The safety headline is fairly reported and stable across church types. But a reader meets the reassuring number first and has to go looking for the harder one. The summary leads with safe and gives less room to the distinction respondents kept drawing between safe and supported. It foregrounds the aggregate, where the role gradient that carries the signal does not appear. And the Q14 reporting result — a plurality across several roles expecting people to become less likely to report conflict, bullying or harassment — sits below the headline rather than beside it, though it is just the kind of finding a work-health-and-safety process exists to surface. None of this is to fault the people who wrote it; summarising contested data under real pressure is hard, and a reassuring frame is an understandable place to land. It is only to say that the frame and the finding point in different directions, and the finding is the one a safety process most needs to see.

3. Are the concerns founded?

As to cost and risk, yes. The committee notes that the strongest predictor of concern is whether a congregation currently has female elders, and that those churches are over-represented in the responses. That is the reason to read the raw aggregate with caution — and earlier I set out how a reader can correct for it, on the most hostile assumptions, and find the lean still holds. What remains after that test is the most telling thing in the data: the churches most worried about losing women from session are the ones that have women in session now. They are not reacting to an abstraction or a memory; they are living the arrangement the change would remove, and they feel its loss most sharply. That makes sense, and it is about as close to evidence as a survey of anticipated effects can get. What the data does not establish is that harm is inevitable: this measures expectation, not outcome, and the committee is right that good transition structures could mitigate much of it. But respondents on both sides agreed those structures would be needed. The concern is evidenced; the harm is contingent; and what it is contingent on is whether the structure gets built.

What would sharpen this — and what only the committee holds
The survey’s own charts give us a good deal: the split of respondents by role, and by gender within each role, both of which I’ve used above. Two things would let a reader go further. First, exact respondent counts — I’ve read the bases off the demographic graph, so they are approximate, and the committee holds the precise figures. Second, and more useful, a cross-tabulation of answers by gender: not just how many men and women responded, but how each answered each question. That is the one cut that would show whether, on any given question, a particular gender in a particular role was markedly more concerned — and it is not in anything released. Both sit with the committee. Releasing them would let anyone replace my approximations with the real numbers and test every claim here. That would serve the whole church, and it would be welcomed.

From what the data shows, to what we must do

Everything to here has been analysis, and analysis only goes so far. It can tell us what the people who answered expect, how consistently they expect it, and where the weight falls. It cannot tell us what to do about it. For that a church doesn’t turn to a survey, or to the Work Health and Safety Act, or to a committee’s framing. It turns to Scripture. So the rest of this changes its footing on purpose. We have read the data carefully and held it loosely where it was thin. Now comes the harder question — what faithfulness asks of us — and the church has to answer that as the church.

Make the change to eldership if that is what the church, under God’s Word, discerns it must do. I have never contested that, and I do not contest it now. But the data — broad, consistent, leaning one way across every cohort but the ministers who answered — describes a cost that respondents across the church expect, and one they expect to fall hardest on women, on paid staff, and on the congregations that have actually lived the arrangement now being removed. Deciding in obedience is one thing. Doing it well is another, and that depends on whether we build what the change requires. Building it is obedience too.

For we are not first a workplace discharging a statutory duty. We are a people under a Lord who commands the protection of the vulnerable and the restraint of those who shepherd. “Speak up for those who cannot speak for themselves,” says Proverbs; “defend the rights of the poor and needy.” Ezekiel’s indictment falls on shepherds who ruled the flock with force and did not bind up the injured. Peter charges the elders to shepherd the flock “not domineering over those in your charge, but being examples.” The Westminster divines, reading the sixth commandment, named among its plain duties the “comforting and succouring the distressed, and protecting and defending the innocent.” None of this is borrowed from the culture. It is older than the WHS Act by millennia, and it asks more of us than the Act ever could. Where the best secular standard now says a workplace must make it safe to speak up, it is, belatedly, agreeing with our own book.

And it is easy to miss that Presbyterian government is itself a piece of safeguarding. The plurality of elders, the graded courts that hold one another to account — we have these because the tradition recognised long ago that fallen leaders need accountability, and that no one should hold unchecked power over another soul. Narrow who sits in those courts without strengthening the ways the vulnerable reach them, and you weaken the very protection the system was built to give. The answer isn’t to reopen the question of the courts. It is to stay consistent with the logic that gave us courts at all.

What faithfulness asks for here is concrete, and it is worth being precise about, because some will rightly say the denomination already has external mechanisms — the Conduct Protocol Unit, Breaking the Silence, channels that reach past the local session. That is true, and to the church’s credit. But three things must be said about them. The first is that the survey is not measuring whether a pathway exists; it is measuring whether people believe they would be heard — whether someone who is hurt trusts they will be listened to, believed, and not made to pay for having spoken. The second is that these external channels are a backstop, and a backstop that, in practice, often routes a matter back to the session to be dealt with. So they do not let the session off the hook; they lean on it. The third is that, for much of what actually wounds people, the formal machinery is the wrong size and the wrong shape. It is built for the serious and the broken-down. Most of what hurts people in a church is smaller and more ordinary than that — the concern brushed aside, the decision made without them, the sense of not being heard — and reaching for a denominational protocol unit over it can feel too heavy, or simply too alien, especially in smaller or more informal congregations. So it goes unused, and the hurt goes unspoken.

Which means the real ask is not a better way around the session. It is a session people do not need to go around. What we should want — what Scripture asks of elders directly — is that the men who sit in session be loving, kind, patient, and like Christ: shepherds who do not lord it over those in their charge, who bind up the injured rather than rule with force. We should want local congregations to be places where someone who has been hurt is listened to and cared for by their own church, where the ordinary wound is tended before it festers, and where the external supports can sit there, available and trusted, precisely because they are rarely needed. That is a far harder thing to build than a reporting form. It is the work of character, not procedure — and it is exactly the work the office of elder is for. The safeguarding the survey asks for and the eldership Scripture describes turn out to be the same thing: elders people are not afraid to come to.

Around that sit the supports that make it real. Access for women, as of right rather than at an elder’s discretion, to the bodies that handle pastoral care, conflict and discipline. Newer diaconal roles encouraged, taken up, and given time to establish, rather than left as a provision on paper. Genuine support for those most affected, not just the reassurance that most people already feel safe. None of this touches the theology of eldership. All of it is what the data says is needed, and what Scripture says is right — and for a church those are one thing, not two.

And none of this is a fringe demand. The committee running the consultation, setting out the paths open to the Assembly, names exactly this one: to link any change to formal ways of including women’s voices, so that the current standard of safety is maintained or improved. They observe it is the option that would best fit what respondents asked for, were a change to be made — while rightly noting that asking for structure alongside a change is not the same as asking for the change. That is the path this reading points to as well. Build the structure into the change, not after it. The body closest to the data has put the same option on the table.

None of this is new. When the results of this survey did not yet exist — when all we had were the questions it asked and the evidence that bore on them — I set out what I thought would need to be true before this change could responsibly be made: that a real mechanism for women’s voices in the courts of the church be built and tested before the door closes, not promised afterward; that our safeguarding be shown to work rather than merely assumed to; that there be a way to raise a concern that does not depend on the goodwill of the very people it might be about; that the church know which path a given hurt is meant to travel; and that the commitments we have already made to honour men and women alike be carried out, not left on paper. I made that case from governance research, from public inquiries into institutional harm, from the experience of churches that had already walked this road — and I was careful to call such evidence suggestive rather than proof, because perhaps we were different. That was the honest caveat. The survey has now answered it. What we know now is that the external evidence was never only external. The pattern the research described, our own people have described back to us — in their own pews, in their own words. We are not the exception the caveat left room for. The outside was a picture of the inside all along.

And I must say plainly that more has been done than I knew when I first wrote, and that it is good. At presbytery level there is now a compulsory female representative for safe-ministry, and a compulsory female support person for anyone who makes a complaint; presbyteries have been urged to keep a list of women able to support complainants, and there is provision for female presbytery affiliates. At Assembly, deaconesses and key women have been given rights described as equivalent to an associate member, a 2025 overture would extend that to women serving on committees, a committee is examining bullying at the local-church level, and the women’s engagement committee has brought overtures to widen women’s standing further still. At the session level, churches are being encouraged to establish deacons’ courts, to invite women to attend session, and to appoint both men and women as safe-ministry representatives. None of this is nothing. The denomination is not standing still, and the direction is right. I’m glad of it, and I won’t take any of it back.

And it is not only the formal measures. In many churches, ministers and sessions have taken this conversation seriously and handled it well — listening carefully, leading their congregations through hard ground with patience, making space for women’s voices long before any rule required it. Their congregations and communities have noticed, and have thanked them for it. That work is real, and good, and I don’t want a word of what follows to be heard as discounting it. If anything, it is the proof of the point: these are the churches showing what it looks like done well. The question the survey raises is not whether it can be done well — plainly it can, and is — but whether the people most affected can count on it being done well everywhere, by whoever happens to be in the chair, rather than where the right person happens to be sitting in it.

Because hold it up to the light and two things show through. The first is that almost none of it is yet finished, and much of what exists grants a voice without a vote. Some is encouragement a church may take or leave. Much of the rest gives rights “equivalent to an associate member” — and that phrasing is worth understanding, not as a slight but as a mechanism. An associate member is someone who already holds a full vote in another court of the church and is recognised, and seated, when visiting this one. The status is dignified precisely because the vote lives at home. To grant women rights equivalent to that is to extend the recognition — a real and good thing — but there is no home court underneath it, no vote anywhere for the equivalence to rest on. It copies what an associate looks like in the room, not what an associate is. The opportunity to speak is a genuine gift. The opportunity to have a say is a greater one, and these measures, real as they are, stop at the first. A good deal more, too, is still proposed rather than enacted — real intent that has not yet reached the Code where it would become something a person could rely on as of right. Being in motion to a destination is not yet having arrived. The committee itself, weighing the options open to the Assembly, makes a kindred observation: that across twenty years or more of this debate, only a very small change has so far been adopted — the granting of associate-equivalent rights to deaconesses and committee members. That is a candid measure of the distance between motion and arrival, set down by the body running the consultation.

The second is less obvious, but perhaps more telling. I showed an earlier draft of this piece to a friend, and they were able to gather these efforts together for me. Echoes of them had surfaced in earlier conversations — but this was the first time anyone had put a concrete list in front of me. And it is not that I hadn’t looked, or hadn’t asked. I am not standing at a distance from this process. If it takes someone assembling a bespoke list before an engaged member can see the progress whole, then the question is no longer only whether these things exist. It is how known they are, and how trusted, by the people they are meant to protect. A safeguard that the committee that built it understands, and the person in the pew has never heard of, is not yet doing the work of making anyone feel safe to speak. That is the same gap the survey found, one court higher: the distance between a thing that exists and a thing that is known, trusted, and reached for when it is needed. The progress is real. It is also, so far, strongest where the hurt is most distant, and least settled exactly at the session, where the survey says the hurt is closest.

I want to be honest about the limits of my own judgement. I don’t know whether what has been done is enough — whether it truly meets the concerns the survey surfaced, or only gestures toward them. It may be that the work is sound and only its visibility is lacking; if so, that is good news, and I will gladly be wrong. But I don’t think we are there yet, and I’ll say why, because the reason is one anyone can weigh. A church is being asked to trust that its courts, and the culture around them, are healthy enough to hear the vulnerable and treat them well. The clearest evidence we have about that culture is the way this very debate is being conducted — its tone, and the calibre of the discourse around it, in our courts and online. And that conduct does not yet read like a culture that has arrived. A conversation about caring for the vulnerable that is not always carried on with care is itself the answer to whether the structures can simply be trusted to work. I hold that as a sense, not a proof. But it is a sense with public evidence behind it, and it tells me there is still work to do.

I don’t have a vote in this. A good many of us don’t. But a vote and a voice aren’t the same thing. A church that remembers the difference, and makes room for the voice of those who have no vote, is only being what Presbyterians have always said the church is — a body where the weakest members are indispensable, and where no one is heard only when someone else allows it. Make the change, if it must be made. But build the structure first. The survey has told us, in the church’s own words, what its people expect it to cost if we don’t. Scripture told us long before the survey did.

Notes on method and sources. (1) All figures here are partial returns, not absolute totals — the survey sampled attenders, and the denominational counts are built only from congregations that reported. Percentages should be read as “of those who answered,” not as population proportions. Approximate respondent bases, read off the survey’s demographic chart: lay (member/attender/volunteer) ~1,570; elders ~110; ministers/PW/CGW ~50; paid staff ~27; deacons ~27. Gender split, read off the same chart and cross-checked against the ~56% female headline: overall ~56% F / ~42% M; lay ~60/40; paid staff ~52/48; deacon ~64% F; elder ~20% F; minister ~9% F. No published cross-tabulation of answers by gender exists, so no claim is made about how either gender answered any individual question. At a base of ~27 the 95% margin on a single proportion is roughly ±19 points (so a 47% reads as a 28–66% band); at ~110 it is about ±9 points, and at ~1,570 about ±2–3 points. That is why the paid-staff and deacon columns are read for direction (which side of the line the band falls) rather than for their point value, while the elder and lay columns can be taken close to face value. Survey percentages are from the WHS consultation data document, Q8–Q15; the denomination’s own current counts, from the PCNSW 2026 Snapshot to the General Assembly, run to roughly 700 elders (~12% female), 178 ordained ministers/ministry workers, 26 CGWs, 8 deaconesses, 311 central-payroll staff (2026, up from 292 and 262 in the two prior years), ~13,500 adult attenders within a total average attendance of 18,088, 183 pastoral charges, and a real vacancy of about 19% (three-quarters of un-appointed vacancies regional). The estimate of around eighty female elders is 12% of the ~700 total; the firmly-counted figure within the 144 reporting congregations is about 66. (2) The minister low-concern reading is offered as a divergence in how the cohorts read the question — some ministers framing the change as a matter of obedience rather than cost — not as a claim of minister under-representation, nor as an imputation of motive. (3) Elder figures rest on ordinary sample strength (a large office, an unremarkable sample), not on any finite-population or census claim. (4) The deacons’-court column rests on about 27 self-identified respondents — a small base, read for direction and consistency with the gradient only; no precise-value, size, or representativeness claim is made for it. (5) Net-negative figures for Q9/Q11 sum the “some” and “significant” negative bands; these additions have been checked against the survey tables (e.g. paid-staff Q9 = 28% + 28% = 56%; minister Q11 positive 25% / negative 20%). (6) The 1,787 / 117-charge figures are from the survey results document itself. (7) The reporting claim is framed to the Safe Work Australia national model code of practice on managing psychosocial hazards (2022), which treats under-reporting of bullying/harassment as a marker of poor culture and a barrier to hazard identification; the approved NSW code (2021) does not yet use this wording, and the text says so. The claim is deliberately not pinned to the NSW Regulation, which does not name reporting behaviour. (8) The central conclusion — a consistent, directional, cross-question gradient with ministers lowest and every other cohort elevated — is a property of the responses and does not depend on any population denominator; the partial-returns caveat does not weaken it. (9) The committee’s summary of the results, its four risk-management paths (including the option of linking any change to formal structures for women’s voices), and its observation that twenty-plus years have produced only associate-equivalent rights reflect the Elders Consultation Special Committee’s reporting to the Assembly (Ben Greig, Convener). Where the committee’s earlier report and the final survey results differ, the final results are treated as authoritative. The committee’s note that responses were not drastically different by gender, age or role is consistent with the reading here: most roles cluster together, and the analysis isolates the single cohort — ministers — that sits apart.

Some of this analysis was prepared with the assistance of AI tools, used for research, data-checking, and drafting. The argument, the judgements, and the position it reaches are mine, and I take responsibility for them.

Tagged
Unknown's avatar

Published by andrewyager

Andrew is the CEO of the Real World Group, a family of IT and telecommunications companies based in Sydney, Australia. Andrew loves Jesus, his wife Bess and his four awesome kids.

3 thoughts on “Through the Window”

  1. Thank you Andrew, for another excellent and insightful analysis on this topic.

    Reply

  2. Thank you Andrew for your in-depth analysis and for expressing the statistical findings in such a readable style.

    As a Ex-Student of PLC and a member of my local church I am disturbed by the lack of transparency in the debate with virtually no discussion within our congregation in the lead up to the vote regarding male only eldership.

    I filled out the survey sent out by the church regarding male eldership . Before opening it I felt glad that my opinion was being sought but to my dismay it rapidly became apparent that it was more geared towards gauging the impact of the mooted proposal.
    I had to ask Chat GPT where to find the results of the survey which led me to your work.

    Considering the crux of the survey speaks to how women are responding to the changes that will diminish our voices being heard it is somewhat ironic that the findings have been buried in a somewhat misrepresentative summary. We feel safe under the current system but we definitely don’t feel supported.

    I would like this part of your analysis to be pursued in the next 12 months.

    “Second, and more useful, a cross-tabulation of answers by gender: not just how many men and women responded, but how each answered each question. That is the one cut that would show whether, on any given question, a particular gender in a particular role was markedly more concerned — and it is not in anything released. Both sit with the committee. Releasing them would let anyone replace my approximations with the real numbers and test every claim here. That would serve the whole church, and it would be welcomed.”

    Reply

    1. While I do not have the raw data, I have been informed that the cross section data does one of two things:

      – in most cases the responses were balanced; both men and women saw concern with similar levels; which is encouraging because it means that the issues are not present simply because of gender
      – in a small number of cases the cross section data resulted in information that would reasonably identify the individuals involved and so it was not safe to publish

      The big question for the PCNSW now is what next. They believe they have made a biblical decision based on their reading of 1 Timothy 2. But that doesn’t address the practical issues of agency, safety and care – and the churches that are listening well are hearing that issue and taking positive steps to address it in their congregations – more so than ever before.

      But that doesn’t fix the denominational structural issues; which I would argue have been there prior to this decision.

      Reply

Leave a Reply