(2026 Exams) Reliability (A2 only)

Test yourself on (2026 Exams) Reliability (A2 only)

Test your knowledge with free interactive questions on Seneca — used by over 10 million students.

Assessing Reliability: Internal and External

Reliability is how consistent a test or study is. Reliability can be assessed in several ways: test-retest, the split-half method and correlating inter-observer reliability.

Reliability

Reliability is how consistent a test or study is.
In other words, does the test, when repeated in identical conditions or with similar participants, come up with the same results?

Test-retest

Test-retest is when a study is repeated.
If the results are similar both times, the test is said to have external reliability.
- External reliability is when a test or study should consistently produce the same results no matter the time given.

Examples of external reliability

A simple example would be if a doctor gave two pregnancy tests to a woman and the results came back positive both times. This means the test is reliable.
Another example would be a maths test. If a student took a maths test twice and got similar results, the test is said to have external reliability.

Split-half method

The split-half is used to assess the internal reliability of a test.
The split-half method takes one test and divides it into two sections.
- This can be even/odd numbered questions or first half/second half.

Internal reliability

If there is a strong positive correlation between both halves, the test is said to have internal reliability.
In other words, a participant scores similar results in both sections of the tests.
Internal reliability is looking at the same test - do all parts of the test have consistent results?

Assessing Reliability: Inter-Observer Reliability

Reliability is how consistent a test or study is. Reliability can be assessed in several ways: test-retest, the split-half method and correlating inter-observer reliability.

Inter-observer reliability

In a highly idealised study, the exact same researcher would administer the test to all participants.
This is to avoid inconsistent results due to the extraneous variable of different researchers.
Obviously, this can be challenging to do. So researchers must follow a specific and standardised procedure to avoid this.

Correlated

To make sure the tests have inter-observer reliability, tests from different researchers are correlated to check if the results have high positive correlation.

Example - aggressive five-year-olds

For example, if two separate researchers are observing aggressive behaviour in five-year-olds, they should give the same score (1-10) for the same level of aggression.
A five-year-old who hits another child would receive a score of 10.
If both researchers observed the behaviour and gave the same score of 10, then the test is said to have inter-observer reliability.

Improving Reliability and Validity

Reliability and validity are essential for making sure that psychological tests are accurate, fair and meaningful. Standardising research and operationalising variables improve both validity and reliability.

Standardising research

Standardising research is crucial for having reliable and valid results.
Scientific studies follow strict guidelines and specific procedures.
- An example of this would be a study on short-term memory recall. The aim of this study is to look at the effect of age on short-term memory.

Conditions and procedure

Psychologists would decide on the conditions of the study.
- For example: a well-lit room, started at 9:00AM, have same race, same-gendered participants and one researcher in the room.
The study would follow a set procedure of handing out papers, giving instructions from a set script, and setting a certain time limit on memory questions.

Extraneous variables

The procedures must be as specific as possible to make sure that there are as few extraneous variables as possible.
- Extraneous variables are anything that could affect your results.
These set procedures make sure that external reliability and inter-observer reliability are improved.

Operationalising variables

When setting up a study, psychologists determine their aim and from there, establish their independent and dependent variables.
They will also try to minimise extraneous variables which could affect their results.
The variables must be clearly defined (to operationalise).

Guidelines

In the Kanner et al. (1981) study examining how daily hassles link to stress and health, they had to clearly define stress, daily hassles and health.
The participants had to rate their hassles, making it clearer.
If they were just to list hassles, but with no guidelines or descriptions, the results could have potentially been less valid.
Making the variables crystal clear improves the validity and reliability of the test.