Falling Marriages, Rising Cohabitation and a Claim about 'Family Values'
The local council of a Maltese town is drafting a report on 'the family'. The statistics officer presents three figures for the town over 20 years: marriages down from 210 to 130 a year; births down from 340 to 200; and households recorded as cohabiting up from 90 to 260. One councillor concludes: 'This proves family values are collapsing - people no longer want to commit, and that is why we have fewer children.' Another councillor notes that the town's population has fallen by 8 per cent over the same period and that the mean age has risen from 36 to 44. A third points out that from 2017 the national statistics office changed the way it recorded marital status in the census, creating a separate cohabitation category. The council must decide what to say in its report.
1. Identify what is wrong with the councillor's inference, using at least three problems with the data.
Read the worked answer
Counts, not rates. All three figures are raw numbers for a population that shrank by 8 per cent. Marriages per 1,000 residents, and especially births per 1,000 women aged 15–49 (the general fertility rate), are the appropriate measures. Some of the fall is simply fewer people.
Age structure. The mean age rose from 36 to 44, so there are proportionally fewer women in the childbearing years and fewer people in the ages at which first marriages typically occur. An ageing population will produce fewer births and marriages without any change in values.
A definitional break. The 2017 change in census recording created a cohabitation category that did not previously exist as a separate code. Part of the rise from 90 to 260 is an artefact of measurement: households now counted as cohabiting may previously have been recorded otherwise.
Postponement versus decline. Falling annual marriage counts are compatible with people marrying later rather than never; period figures cannot distinguish these. Cohort data would be needed.
Ecological fallacy. Town-level totals say nothing about the commitments or attitudes of individual residents. No data on values has been presented at all.
2. Explain the difference between a value judgement and a sociological claim here, and redraft the councillor's sentence for the report.
Read the worked answer
The distinction. 'Family values are collapsing' evaluates a change against a normative standard and treats change as deterioration. It is not a factual statement that data could confirm or refute: no measure of 'values' has been presented, and 'collapsing' already contains the verdict. A sociological claim would be descriptive, specify its measure and remain open to competing explanations - demographic, legal, economic and definitional.
Redraft. 'Over twenty years the town recorded fewer marriages and fewer births and more households recorded as cohabiting. Part of this reflects a shrinking and ageing population, part a change in how unions are legally and statistically recorded since 2017, and part a shift in partnership patterns, with marriage later or replaced by cohabitation. The figures describe changing household and partnership arrangements; they do not measure commitment.'
Why this is better. It distinguishes what was measured from what was inferred, names the confounders raised by the other councillors, and identifies candidate explanations - later marriage, secularisation, changing gender roles, housing costs - rather than asserting one. It also avoids the implication that a legal status is the only valid form of commitment, which is a value position the council has no mandate to adopt.
3. Which additional indicators would let the council distinguish postponement from a lasting decline in partnership and childbearing?
Read the worked answer
Age-specific first-marriage rates. If the rate for women aged 25–29 has fallen but risen for 30–34, partnership is being postponed. If rates fall at every age, that suggests a genuine decline.
Mean age at first birth and cohort completed fertility. The mean age at first birth shows timing directly. Completed fertility for women who have passed their childbearing years shows whether postponement was later made up or lost.
Cohort measures rather than period measures. Period totals in a single year are distorted when births are delayed, because women of several cohorts are simultaneously having fewer children later. Following one cohort over time resolves this.
Separation and divorce rates, and household size. These show whether unions are less stable as well as differently formed, and whether single-person households are growing.
Attitudes data. Survey questions on ideal family size and on the importance of marriage would measure the 'values' the councillor invoked, rather than assuming them. Together these indicators would let the council report change without deciding in advance what it means.