03 Section B Option 3, Key question 1 - Measuring religiosity and belonging (part 3)

Dimensions of religiosity

Religiosity has several dimensions, so one indicator cannot describe every way religion matters in a person's life.

Religious belief concerns what someone accepts as true about religious matters: for example, a divine being, an afterlife or the authority of a sacred teaching. Faith can involve trust or commitment as well as agreement with propositions. Asking whether someone believes in God measures something different from asking how important faith is when making a difficult decision. A yes-or-no answer can also hide certainty, doubt and changing interpretations.

To measure an abstract idea, researchers operationalise it: they decide which observable answers or actions will count as indicators. Their choices shape the result. A survey that defines a religious person as a weekly worshipper will classify some private believers as non-religious. A survey using only a religious identity may include people who reject that religion's teachings.

Institutional influence is a property of organisations and social arrangements, rather than simply another score for an individual. Researchers might examine whether a religious organisation actually shapes a school policy or provides a service. Counting how often its leader appears in the news would measure visibility; tracing an accepted policy proposal would provide stronger evidence of influence. An organisation may retain influence while its membership declines.

Consider this original example: Leena identifies as Hindu, prays at home and attends a temple only occasionally because it is far away. An attendance-only measure would underestimate her private commitment. That does not make attendance useless: it remains relevant if the research question is about participation in temple congregations. The error comes from treating that specific indicator as the whole of religion.

Counting affiliation and practice

A census and a religious organisation's records can both provide useful evidence while counting different things.

The Office for National Statistics describes the religion measure in the England and Wales Census 2021 as affiliation, rather than a test of belief or practice. Of all usual residents, 46.2% were recorded as Christian, compared with 59.3% in 2011; 37.2% were recorded as having no religion in 2021, compared with 25.2% in 2011. The religion question was voluntary and 94.0% answered it in 2021. The published percentages use the whole usual-resident population, including people whose religion was not stated, as their denominator. ONS, Census 2021 religion bulletin, released 29 November 2022.

The Christian share fell by 13.1 percentage points: 59.3 minus 46.2. This is a change in the proportion identifying as Christian in England and Wales. It does not show that every person who ceased identifying abandoned belief, that 46.2% attended church, or that the whole United Kingdom had the same distribution. These are aggregate snapshots, not a record of the same individuals changing their answers. Population composition can change between censuses too.

A census offers broad coverage and detailed local comparisons, including groups too small for reliable estimates in a modest survey. However, a very large dataset does not repair a mismatch between concept and indicator. Someone who ticks a Christian identity because of family heritage is not necessarily misreporting affiliation; the researcher makes the mistake if they relabel the response as doctrinal belief. Equally, no religion is not a direct declaration that God does not exist. Nor can the unanswered category simply be added to no religion. Its members' answers are unknown.

Organisational evidence answers other questions. The Church of England's annual Statistics for Mission collects parish returns covering attendance and participation, including baptisms, marriages and funerals. These are records of activity within that institution, rather than a count of all religion in England or Britain. Church of England Data Services.

Records can avoid asking an individual to remember a year's attendance, but their units still need inspection. Attendances count visits; attenders count people. In an original counting example, 30 people who each attend four services produce 120 attendances, not 120 different worshippers. Membership lists may include inactive members, while visitors may attend without joining. An unusual festival or a building closure can also alter the count. Researchers need comparable periods, clear rules and information about missing returns before interpreting change.

The distinction also protects comparison: two sources need not disagree merely because their totals differ. They may be answering different questions.

Believing without belonging

Belief and organisational participation can become separated, but the existence of that separation does not prove that belief remains unchanged over time.

Grace Davie's believing without belonging argument directs attention to religious belief that persists despite weak engagement with organised religion. She developed it in work including her 1990 article and her 1994 book on religion in Britain since 1945. Treat it as an interpretation of patterns of religion, rather than as one experiment that established an unchanging national law. Davie, 1990, publisher record.

The argument challenges the inference that empty places of worship necessarily mean the disappearance of religious convictions. People may retain ideas about divine help or life after death while rejecting an organisation's authority, finding its services unhelpful or practising privately. In an original example, a woman believes in God and prays during a family illness but has no congregation and never attends services. Her beliefs would be invisible in a church-membership register. A survey asking her separately about belief and participation could capture the gap.

Here belonging needs an explicit indicator. Formal membership, attendance and a self-chosen religious identity are not identical. Someone may have no organisational involvement yet still describe themselves as Christian. The same person could therefore count as believing without belonging under an attendance definition, but as both believing and belonging under an affiliation definition. State which meaning is being used.

There are two distinct claims to examine. One is that some individuals believe without participating. The other is that belief is resistant to the social changes reducing participation. The first needs evidence linking belief and participation in the same individuals. The second needs comparable evidence across time. A gap at one date does not establish different rates of decline.

Voas and Crockett (2005) tested the more general claim using secondary analysis of the British Household Panel Survey and British Social Attitudes surveys. They reported erosion of belief alongside affiliation and attendance, and interpreted decline as generational: less religious younger birth cohorts replace older cohorts in the population. Their findings challenge the claim that belief has generally remained robust while belonging declines. They do not show that no individual believes without participating. Voas and Crockett, University of Manchester research record.