Lesson 7 of 8 · 20 min

Twin, family and adoption studies

Read correlation coefficients from twin, family and adoption studies, say what each comparison of relatives shows, and evaluate the Minnesota study of twins reared apart (Bouchard et al. 1990).

What you will be able to do

  • Explain the logic of family, twin and adoption studies of intelligence.
  • Interpret correlation coefficients between relatives' IQ scores, saying what each comparison isolates.
  • Describe the aim, method, results and conclusion of Bouchard, Lykken, McGue, Segal & Tellegen (1990).
  • Evaluate twin and adoption research, including its design, sampling and ethics.

Family members are often alike in IQ. That could be because they share genes, or because they share a home, a school, books, meals and conversations, or both. Psychologists cannot assign babies to families to find out, so they use the "natural experiments" that families already provide: relatives who share different amounts of genes and different amounts of environment.

What relatives share
RelationshipGenes shared (on average)Same home?
Monozygotic (MZ, identical) twins reared together100%Yes
MZ twins reared apart100%No
Dizygotic (DZ, fraternal) twins reared together50%Yes
Siblings reared together50%Yes
Parent and biological child, living together50%Yes
Parent and biological child adopted away50%No
Unrelated children adopted into the same home0%Yes
Adoptive parent and adopted child0%Yes

Key idea

The logic

If genes matter, IQ correlations should rise with the proportion of genes shared, even for relatives raised apart. If the shared environment matters, people raised in the same home should be more alike than people with the same genes raised apart, and unrelated children raised together should be alike at all.

Three designs. A family study compares relatives of different closeness. A twin study compares MZ twins with DZ twins of the same sex, both raised together: if MZ pairs are more alike, the extra similarity is put down to their extra shared genes. An adoption study compares adopted children with their biological parents (shared genes, no shared home) and with their adoptive parents (shared home, no shared genes). All three are correlational: nothing is manipulated, so they measure how closely two variables, here the IQs of two relatives, go together.

Note

Describing a correlation

Give the direction (positive or negative) and the strength. On the scale used across Study: ∣r∣≥0.75|r| \geq 0.75 strong, 0.5 to 0.75 moderate, 0.25 to 0.5 weak, below 0.25 little or no correlation. Your textbook may use slightly different cut-offs; what matters most here is comparing the coefficients.

In 1981 Thomas Bouchard and Matt McGue pooled 111 studies of the IQs of relatives and averaged each kind of correlation.

Average correlations between the IQs of relatives (Bouchard & McGue 1981)
RelationshipGenes sharedSame home?r
MZ twins reared together100%Yes0.86
MZ twins reared apart100%No0.72
DZ twins reared together50%Yes0.60
Siblings reared together50%Yes0.47
Parent and child living together50%Yes0.42
Siblings reared apart50%No0.24
Parent and child, child adopted away50%No0.22
Unrelated children adopted into the same home0%Yes0.34
Adoptive parent and adopted child0%Yes0.19

Read these in pairs that differ in only one thing:

  • Same home, different genes. MZ twins together (0.86) against DZ twins together (0.60); siblings together (0.47) against adopted children together (0.34). Similarity rises with genes shared: evidence for genetic influence.
  • Same genes, different homes. MZ twins together (0.86) against MZ twins apart (0.72); siblings together (0.47) against siblings apart (0.24). Sharing a home adds similarity: evidence for environmental influence.
  • No genes shared at all. Adopted children raised together (0.34) and adoptive parents and their children (0.19) are still positively correlated, which can only come from the environment (or from selective placement, below).

Both kinds of influence show up. The pattern fits genes and environment working together.

Interactive

Compare two relationships

Pick two relationships. The lab says what differs between them and what the gap in r suggests.

Worked example

Read IQ correlations between relatives

  • Simple familiar

The table gives average correlations between the IQ scores of relatives (Bouchard & McGue 1981).

Relationshipr
Identical (MZ) twins reared together0.86
Fraternal (DZ) twins reared together0.60
Siblings reared together0.47
Siblings reared apart0.24

(a) Which two correlations show the effect of genes, with the home held the same? What do they suggest? (b) Which two show the effect of the home, with genes held the same? What do they suggest?

4 steps

Try it yourself first, then check the working one step at a time.

Method

  1. Find two relationships that differ in only one thing: genes or home.

  2. Compare their correlations and say what the difference suggests.

Watch out for

  • Comparing MZ together with siblings apart: they differ in both genes and home.

  • Saying a correlation of 0.86 means 86% of intelligence is genetic.

Which comparison best isolates the effect of sharing a home, with genes held constant?

Choose one answer

The Minnesota study of twins reared apart

Identical twins separated early in life and raised in different homes are the closest thing to a natural experiment on intelligence: same genes, different environments. They are also rare.

Bouchard, Lykken, McGue, Segal & Tellegen (1990)
PartDetails
AimTo find how much of the variation between people in intelligence, personality and other characteristics is associated with genetic differences, by studying twins separated early in life and reared apart.
ParticipantsFrom 1979, more than 100 sets of twins or triplets reared apart, found through media coverage, word of mouth and professionals. Monozygotic (MZA) and dizygotic (DZA) pairs, with zygosity checked by biological tests.
ProcedureAbout a week (some 50 hours) of medical and psychological assessment in Minneapolis, including the WAIS and other ability tests and measures of personality, interests and attitudes. Each twin was tested separately by a different examiner. Information was gathered on how much contact the twins had had and on their adoptive homes.
DesignCorrelational (a natural, or quasi-, experiment): no variable was manipulated. The IQs of the two twins in each pair were correlated.
ResultsThe WAIS full-scale IQs of MZA twins correlated 0.69 (48 pairs): higher than the 0.60 usually found for fraternal twins raised in the same home, though lower than for identical twins raised together. MZA twins were about as similar as MZ twins reared together on personality, interests and social attitudes. Similarity was not explained by how much contact the twins had had or by how similar their adoptive homes were.
ConclusionAbout 70% of the variation in IQ in this population was associated with genetic variation. Because MZA twins share genes but not a home, their correlation estimates this proportion directly.

Worked example

Evaluate the conclusion of the Minnesota twins study

  • Complex unfamiliar

Bouchard, Lykken, McGue, Segal & Tellegen (1990) found that the WAIS IQs of identical twins reared apart correlated 0.69, and concluded that about 70% of the variation in IQ was associated with genetic variation.

Evaluate this conclusion. (5 marks)

5 steps

Try it yourself first, then check the working one step at a time.

Method

  1. Explain why the design supports the conclusion.

  2. Name strengths of the method.

  3. Name factors that could inflate or bias the estimate.

  4. Judge how far the conclusion holds, and for whom.

Watch out for

  • Only describing the study. Evaluate means judging the conclusion against the evidence.

  • Saying the study proves intelligence is 70% genetic for each person.

Strengths

  • Separated MZ twins separate genes from the shared home better than any other design.
  • Intensive, standardised testing with well-established instruments such as the WAIS.
  • Twins were tested separately by different examiners, reducing experimenter bias.
  • Zygosity was confirmed biologically, not by appearance.

Limitations

  • Small sample: 48 MZA pairs gave the WAIS correlation, so the estimate has wide uncertainty.
  • Sampling bias: volunteers who answered publicity may be unusually similar twins.
  • Shared environment remains: the same womb, often months together before separation, and some contact before testing.
  • Selective placement: adoption agencies often placed children in homes like their birth families'.
  • Limited range of homes: mostly Western homes of adequate income, which narrows environmental differences and inflates heritability.

Note

Ethics

Twins gave informed consent and could withdraw, but the study raised ethical issues worth naming: confidentiality (separated twins attracted media attention, and the twins' life histories are personal), the emotional impact of reunions that the study sometimes brought about, and debriefing and support afterwards.

Many of the twins in Bouchard et al. (1990) volunteered after seeing stories about other reunited twins in newspapers and on television. Which limitation does this create?

Choose one answer

Exam tip

Using the evidence

Quote the numbers and say what each comparison shows: "MZ twins reared apart correlate 0.69 on the WAIS (Bouchard et al. 1990), more than DZ twins raised in the same home (0.60), which suggests a substantial genetic influence; however, MZ twins reared together (0.86) are more alike than those reared apart (0.72; Bouchard & McGue 1981), showing an environmental influence too."

Summary

  • Family, twin and adoption studies compare relatives who share different proportions of genes and of environment. All are correlational.

  • IQ correlations rise with genes shared (MZ together 0.86, DZ together 0.60), which suggests genetic influence.

  • People with the same genes are more alike raised together than apart (MZ 0.86 against 0.72), and unrelated children raised together still correlate (0.34), which shows environmental influence.

  • Bouchard et al. (1990) found MZA twins' WAIS IQs correlated 0.69 and concluded that about 70% of IQ variation in their population was associated with genetic variation.

  • Limitations: small, self-selected samples, shared prenatal and early environment, contact, selective placement and a narrow range of homes.

Finished the lesson?

Try the 2 quick checks above and the lesson completes itself, or mark it complete now.

Lesson 7 of 80% read

Save your progress with a free account