Assessment objectives common mistakes

Common mistakes

• Confusing random assignment with random sampling

  Students often think random assignment means selecting a random sample from the population.

  Fix itRandom assignment refers to how participants are allocated to conditions within the study, whereas random sampling concerns how participants are chosen from the population.

• Confusing randomisation with random sampling

  Assuming randomisation means selecting participants randomly from the population.

  Fix itRandomisation refers to the random assignment of participants to conditions within the study, not the selection of participants from the population.

• Confusing correlation with causation

  Assuming that a correlation between two variables proves a causal relationship.

  Fix itExplain that correlation indicates association but not causation; experimental designs with randomisation are required to infer causality.

• Confusing qualitative and quantitative data

  Treating qualitative data as if it were numeric and applying statistical tests to it.

  Fix itQualitative data should be analysed through coding, thematic or content analysis, not through statistical procedures. Use appropriate qualitative methods to preserve the richness of the data.

• Confusing Standard Deviation with Standard Error

  Treating the standard deviation of a sample as if it were the standard error of the mean.

  Fix itThe standard deviation reflects variability within the sample, while the standard error is the standard deviation divided by the square root of the sample size and indicates how precisely the sample mean estimates the population mean.

• Confusing Correlation with Causation

  Assuming that a statistical association between two variables proves a causal link.

  Fix itExplain that correlation indicates a relationship but does not confirm causality; to infer causation, experimental manipulation and control of confounding variables are required.

• Confusing correlation with causation

  Assuming that a significant correlation indicates that one variable causes the other.

  Fix itClarify that correlation only shows association; causation requires experimental manipulation and control of confounds.

• Confusing internal and external validity

  Students often state that a study has high internal validity because it uses random assignment, without recognising that external validity (generalisability) may still be limited.

  Fix itClarify that internal validity concerns causal inference within the study, while external validity concerns the extent to which findings apply to other populations, settings, or times.