The Psychology Internal Assessment requires you to design and carry out a simple experimental study, typically replicating (or modifying) an existing study from the syllabus, and analyze the results using inferential statistics. Getting the experimental design right at the planning stage is essential — a flawed design can’t be fixed after data collection, no matter how good your write-up is.
Step 1: Choose a Study to Replicate or Modify
The strongest starting point is a well-established study you’ve already covered in class, adapted with a small, clearly justified change — a different sample population, a modified stimulus, or an additional condition. Full-scale, completely original designs are risky, because psychology experiments require careful controls that established studies have already tested and refined.
Step 2: Define Your Variables Precisely
Every Psychology IA needs an operationalized independent variable (IV) and dependent variable (DV):
- Independent Variable (IV): what you manipulate between conditions — must have exactly two conditions/levels for a typical IA design.
- Dependent Variable (DV): what you measure — must be operationalized clearly (not «memory,» but «the number of words correctly recalled out of a 15-word list after a 2-minute delay»).
Step 3: Choose an Appropriate Experimental Design
There are three standard designs, each with trade-offs:
- Independent Samples Design — different participants in each condition; avoids order effects but requires more participants and is vulnerable to participant variables.
- Repeated Measures Design — same participants complete both conditions; controls participant variables but introduces order effects (usually addressed through counterbalancing).
- Matched Pairs Design — participants are paired on a relevant variable (e.g., prior test score) and then split between conditions; balances the strengths of the other two designs but is harder to arrange.
For most school-based IAs, independent samples or repeated measures (with counterbalancing) are the most practical choices.
Step 4: Sampling Method
Your sampling method needs to be explicitly named and justified, not just described. The most common methods used in Psychology IAs are:
- Opportunity sampling — easiest to arrange (classmates, school community), but introduces potential bias that must be acknowledged in your evaluation.
- Volunteer sampling — participants self-select, often via a sign-up sheet; introduces volunteer bias, which should also be discussed critically.
Whichever method you choose, be explicit about your target population and how your actual sample may or may not represent it — this feeds directly into your evaluation and discussion of generalizability.
Step 5: Ethical Considerations You Must Address
The IB requires explicit ethical planning, not just an afterthought paragraph. Make sure you address:
- Informed consent (and parental/guardian consent if participants are minors, which they typically will be)
- The right to withdraw at any point without penalty
- Debriefing after the study, explaining the true aim if any deception was used
- Confidentiality and anonymization of data
- Minimizing psychological harm or discomfort
Step 6: Choosing the Right Statistical Test
Once you have your data, choose the correct inferential test based on your design and data type:
- Mann-Whitney U test — independent samples, ordinal/non-normal data
- Wilcoxon signed-rank test — repeated measures, ordinal/non-normal data
- Chi-squared test — categorical/nominal data
State your null and alternative hypotheses clearly before analysis, and interpret your result in the context of the original study you replicated — did your findings support, contradict, or partially replicate the original conclusion?
Common Mistakes That Lower the Score
- Vague operationalization of the DV, making it unclear exactly what was measured
- No named, justified sampling method
- Missing or superficial treatment of ethical considerations
- Choosing a statistical test that doesn’t match the data type or design
- An evaluation that doesn’t meaningfully connect back to the original study being replicated
Final Tip
Pilot your procedure on 2-3 people before running the full study. Small design flaws — confusing instructions, a DV that’s hard to measure consistently — are far easier to fix before data collection than after.