Caregiver Experiences of Children with Cerebral Palsy
Conducted a baseline equivalence analysis for a two-arm study with 77 participants: a control group (n = 37) and an intervention group (n = 40). The groups were compared on six sociodemographic characteristics: age, marital status, educational level, employment status, occupation and residence. Chi-square tests and an independent samples t-test showed no statistically significant differences between the groups on any variable (all p > 0.05). This confirms the groups were comparable before the intervention and supports valid comparison of post-intervention outcomes.
Challenge
In experimental and quasi-experimental studies, any outcome difference between groups can only be attributed to the intervention if the groups were similar at the start. Pre-existing differences in age, education, income source or location could confound the results and weaken the study’s conclusions. The challenge was to test whether the control and intervention groups were balanced at baseline, using a modest sample with several sparse subcategories and some missing responses.
Objectives
Describe the sociodemographic profile of participants in both groups
Test for statistically significant baseline differences between the control and intervention groups
Identify variables that could act as confounders in later outcome analyses
Flag descriptive imbalances worth monitoring, even where they are not statistically significant
Present findings in a clear report suitable for researchers and non-technical stakeholders
Methods
Frequency distributions (n, %) for categorical variables and mean ± SD for age
Independent-samples t-test to compare mean age between groups
Chi-square (χ²) tests of independence for marital status, educational level, employment status, occupation, and residence
Significance level set at α = 0.05
Variable-by-variable interpretation covering statistical significance and practical relevance
Review of sample size limitations and their effect on detecting differences in small subcategories
Data Sources
Confidential
Key Findings
Confidential (the work has not been made public)
Impact
Confidential (the work has not been made public)