Scientists may also be reluctant to publish their data because readers may be less likely to believe the authors' conclusions if they can see the underlying data that the conclusions were based on. Data are often messy: outliers, missing values, unexpected clusters of points, non-normal distributions, unequal variances between groups or conditions, etc., all of which can be hidden and ignored when presenting the mean +/- SEM in graphs and only p-values in the text.
In addition, publishing data makes it easy for anyone to reanalyse it (with more modern or appropriate methods) and perhaps come to a different conclusion. The potential for embarrassing corrections or retractions is thus greatly increased.
The down-side of publishing data
3 August 2009
Scientists may also be reluctant to publish their data because readers may be less likely to believe the authors' conclusions if they can see the underlying data that the conclusions were based on. Data are often messy: outliers, missing values, unexpected clusters of points, non-normal distributions, unequal variances between groups or conditions, etc., all of which can be hidden and ignored when presenting the mean +/- SEM in graphs and only p-values in the text.
In addition, publishing data makes it easy for anyone to reanalyse it (with more modern or appropriate methods) and perhaps come to a different conclusion. The potential for embarrassing corrections or retractions is thus greatly increased.
Competing interests
No competing interests.