What you'll learn

• Why depression can involve negative interpretation biases.
• How Williams et al. (2013) tested an online treatment combining internet-CBT and cognitive bias modification.
• The study’s aim, design, participants, procedure, findings and conclusion.
• How to evaluate the study for AO3, including methodology, ethics and real-world application.

First: what problem were Williams et al. trying to solve?

In A-Level Psychology, you need to be able to describe studies using AO1 (what the study did and found), apply them using AO2 (linking ideas to scenarios or the topic), and evaluate them using AO3 (strengths, weaknesses, ethics and usefulness).

Williams, Blackwell, Mackenzie, Holmes and Andrews (2013) focused on depression and whether an internet-based treatment could be improved by combining two approaches: reason and imagination.

A key cognitive feature of depression is that people may interpret ambiguous events negatively. For example, if a friend does not reply to a message, someone with depression might think, “They are ignoring me because they dislike me,” rather than “They might be busy.”

The treatment ideas: “reason” and “imagination”

Cognitive behavioural therapy: using reason

CBT is the “reason” part of the study title. It encourages the person to step back from a thought and ask: What evidence supports this? What evidence goes against it? Is there a more realistic explanation?

Cognitive bias modification: using imagination

In Williams et al., the CBM-I was based on positive mental imagery. Participants imagined themselves in ambiguous situations that resolved positively. The idea is that repeated practice may make positive interpretations more automatic.

Williams et al. (2013): aim and design

Williams et al. (2013), in “Combining imagination and reason in the treatment of depression”, aimed to test whether adding internet-based CBM-I to an existing internet-CBT programme would improve outcomes for adults with depression.

The study compared an online CBT treatment with an online condition that also included positive imagery-based CBM-I. Random allocation matters because it helps make the groups similar at the start, so any later difference is more likely to be due to the treatment.

Variables

In this study, the IV was the type of online treatment condition. The DVs included depression symptoms and cognitive interpretation bias, measured before and after treatment, with follow-up assessment.

AO1: what happened in the study?

Participants

The study used a relatively small clinical sample, commonly reported as around 70 adults with depression who were recruited for online treatment. Participants were screened before taking part, which is important because depression is a clinical condition and researchers needed to consider risk and suitability.

Procedure

Participants were randomly allocated to treatment conditions. The key elements were:

• Internet-CBT: an online CBT depression programme delivered over several weeks, including structured lessons and activities designed to challenge negative thoughts and change unhelpful behaviour patterns.
• CBM-I with positive imagery: online training in which participants imagined ambiguous scenarios resolving positively, encouraging a shift away from negative interpretations.
• Repeated assessment: participants completed standardised measures of depressive symptoms and interpretation bias before and after treatment, with follow-up to see whether changes lasted.

Measures

The study mainly produced quantitative data, meaning numerical data such as questionnaire scores. Depression studies often use standardised self-report scales, where higher scores usually indicate more severe symptoms. Interpretation bias can be assessed by asking how positively or negatively participants interpret ambiguous material.

Findings and conclusion

Overall, participants improved across treatment, showing that internet-based CBT could reduce depressive symptoms. The group receiving positive imagery CBM-I alongside internet-CBT showed evidence of more positive interpretation bias and promising improvement in depression outcomes.

The conclusion was that combining “imagination” through CBM-I with “reason” through CBT may be a useful way to treat depression, especially because the intervention can be delivered online.

Research-methods link

This study is useful for research-methods AO3 because it was an RCT with repeated outcome measures.

Researchers may summarise depression scores using a mean as a measure of central tendency and standard deviation as a measure of dispersion if the data are roughly normally distributed. If scores are very skewed, meaning clustered unevenly with a long tail at one end, medians and ranges may be more appropriate. Graphs such as bar charts or line graphs can show changes in symptom scores over time.

AO3: strengths

Stronger causal inference

Because Williams et al. used random allocation, the study has stronger internal validity than a simple before-and-after study. If one group improves more than another, it is more reasonable to link that improvement to the treatment condition.

Standardised and replicable

The online treatment materials and symptom measures were standardised. This improves reliability because each participant receives a similar procedure, and future researchers can repeat the study more easily.

Real-world application

The study has clear practical value. Online interventions can reach people who face barriers to face-to-face therapy, such as cost, location, waiting lists, mobility difficulties or stigma. This makes the study relevant to modern mental health services.

AO3: weaknesses

Sample and generalisability

The sample was relatively small and self-selected. People willing to take part in internet treatment may be more motivated, more comfortable with technology, or less severely depressed than other clinical populations. This limits how far the findings can be generalised.

Attrition risk

Online interventions often have attrition, meaning participants drop out before the end. If people who do not improve are more likely to leave, the results may make the treatment look more effective than it really is.

Self-report and demand characteristics

Depression questionnaires rely on self-report. Participants may misjudge their symptoms, answer in a socially desirable way, or respond to what they think the researchers expect. This can reduce validity.

Mixed evidence for CBM-I

CBM-I is promising, but its effects in clinical depression are not always as strong or consistent as traditional CBT. Later research on cognitive bias modification has sometimes found smaller effects, so Williams et al. should be treated as supportive evidence rather than a final answer.

Ethics

Williams et al. studied people with depression, so ethical safeguards were especially important under the BPS Code of Ethics and Conduct (2009).

Researchers needed informed consent, confidentiality for online data, the right to withdraw, and debriefing. Protection from harm was particularly important because some participants may have experienced severe low mood or suicidal thoughts. Screening, monitoring and referral procedures would be needed to ensure online treatment did not leave vulnerable participants unsupported.

How to use Williams et al. in an essay

For AO1, describe it as a randomised controlled trial testing internet-CBT with positive imagery-based CBM-I for depression.

For AO2, apply it to a person with depression who negatively interprets ambiguous events. Explain that CBM-I trains more positive interpretations, while CBT challenges negative thoughts through reasoning.

For AO3, focus on RCT control, standardisation and real-world usefulness, balanced against sample limitations, attrition, self-report data and ethical issues in online clinical treatment.