Related Experiment Videos
Researchers misunderstand confidence intervals and standard error bars
Sarah Belia1, Fiona Fidler, Jennifer Williams
1School of Psychological Science, La Trobe University, Melbourne, VIC, Australia.
Psychological Methods
|January 6, 2006
Summary
Many researchers misunderstand confidence intervals (CIs) and standard error (SE) bars, impacting statistical significance interpretation. Clearer guidelines and graphical conventions are needed for effective research communication.
Area of Science:
- Psychology
- Behavioral Neuroscience
- Medicine
Background:
- Understanding of confidence intervals (CIs) and standard error (SE) bars among researchers is not well-documented.
- Error bars are frequently used in scientific publications to denote uncertainty and variability.
Purpose of the Study:
- To assess researchers' understanding of how error bars (CIs and SE bars) relate to statistical significance.
- To identify potential misconceptions in interpreting graphical representations of statistical data.
Main Methods:
- Researchers from psychology, behavioral neuroscience, and medicine were surveyed.
- Participants adjusted figures with error bars until means appeared statistically significant (p < .05).
Main Results:
- A significant portion of researchers demonstrated misconceptions regarding error bars and statistical significance.
- Distinguishing between CIs and SE bars was challenging for many.
- The influence of study design (independent vs. repeated measures) on interpretation was often overlooked.
Conclusions:
- Leading researchers exhibit significant misunderstandings of CIs and SE bars.
- Current graphical conventions may be ambiguous, hindering accurate data interpretation.
- Improved guidelines and clearer visualization methods are essential for enhancing research communication.