# Independence of observations

## Independence of observations means each participant is only counted as one observation

The statistical assumption of independence of observations stipulates that

**all participants in a sample are only counted once**. If a participant was to appear multiple times in a sample, each time as an independent observation, then the statistics would be artificially skewed in their favor and not be representative of a true sample of independent participants. Independence of observations makes sure that each participant's variance affects the overall analysis**just once**.### Statistical designs and independence of observations

**Between-subjects**designs can be grossly affected by a violation of this assumption. Comparing groups that are artificially weighted towards ONE PARTICIPANT counted multiple times in one or both groups yields false inferences.

In

**within-subjects**designs, the participants are independent of each other and there are several observations of the outcome "within" each person or across time.

**Multivariate**analyses that account for demographic, etiological, prognostic, confounding, or clinical variables cannot meet the statistical assumptions of normality and linearity if there are multiple observations of the same participant in a sample. This is especially true if the aforementioned variables are measured at a categorical level.

**There is no statistical test associated with testing independence of observations**. It is up to

**you**, as the researcher, to meet this assumption by counting each participant in your study as one independent observation.

## Statistician For Hire

**DO YOU NEED TO HIRE A STATISTICIAN?**

Eric Heidel, Ph.D. will provide statistical consulting for your research study at $100/hour. Secure checkout is available with PayPal, Stripe, Venmo, and Zelle.

- Statistical Analysis
- Sample Size Calculations
- Diagnostic Testing and Epidemiological Calculations
- Psychometrics