↓ Skip to main content

The heterogeneity statistic I2 can be biased in small meta-analyses

Overview of attention for article published in BMC Medical Research Methodology, April 2015
Altmetric Badge

About this Attention Score

  • Above-average Attention Score compared to outputs of the same age (57th percentile)

Mentioned by

twitter
4 tweeters

Citations

dimensions_citation
139 Dimensions

Readers on

mendeley
122 Mendeley
You are seeing a free-to-access but limited selection of the activity Altmetric has collected about this research output. Click here to find out more.
Title
The heterogeneity statistic I2 can be biased in small meta-analyses
Published in
BMC Medical Research Methodology, April 2015
DOI 10.1186/s12874-015-0024-z
Pubmed ID
Authors

Paul T von Hippel

Abstract

Estimated effects vary across studies, partly because of random sampling error and partly because of heterogeneity. In meta-analysis, the fraction of variance that is due to heterogeneity is estimated by the statistic I (2). We calculate the bias of I (2), focusing on the situation where the number of studies in the meta-analysis is small. Small meta-analyses are common; in the Cochrane Library, the median number of studies per meta-analysis is 7 or fewer. We use Mathematica software to calculate the expectation and bias of I (2). I (2) has a substantial bias when the number of studies is small. The bias is positive when the true fraction of heterogeneity is small, but the bias is typically negative when the true fraction of heterogeneity is large. For example, with 7 studies and no true heterogeneity, I (2) will overestimate heterogeneity by an average of 12 percentage points, but with 7 studies and 80 percent true heterogeneity, I (2) can underestimate heterogeneity by an average of 28 percentage points. Biases of 12-28 percentage points are not trivial when one considers that, in the Cochrane Library, the median I (2) estimate is 21 percent. The point estimate I (2) should be interpreted cautiously when a meta-analysis has few studies. In small meta-analyses, confidence intervals should supplement or replace the biased point estimate I (2).

Twitter Demographics

The data shown below were collected from the profiles of 4 tweeters who shared this research output. Click here to find out more about how the information was compiled.

Mendeley readers

The data shown below were compiled from readership statistics for 122 Mendeley readers of this research output. Click here to see the associated Mendeley record.

Geographical breakdown

Country Count As %
United States 1 <1%
Germany 1 <1%
Brazil 1 <1%
Unknown 119 98%

Demographic breakdown

Readers by professional status Count As %
Student > Ph. D. Student 29 24%
Student > Bachelor 19 16%
Student > Master 18 15%
Researcher 15 12%
Other 8 7%
Other 22 18%
Unknown 11 9%
Readers by discipline Count As %
Medicine and Dentistry 33 27%
Psychology 17 14%
Social Sciences 9 7%
Agricultural and Biological Sciences 8 7%
Biochemistry, Genetics and Molecular Biology 5 4%
Other 26 21%
Unknown 24 20%

Attention Score in Context

This research output has an Altmetric Attention Score of 2. This is our high-level measure of the quality and quantity of online attention that it has received. This Attention Score, as well as the ranking and number of research outputs shown below, was calculated when the research output was last mentioned on 30 April 2015.
All research outputs
#6,959,980
of 12,365,836 outputs
Outputs from BMC Medical Research Methodology
#668
of 1,086 outputs
Outputs of similar age
#90,120
of 218,884 outputs
Outputs of similar age from BMC Medical Research Methodology
#12
of 15 outputs
Altmetric has tracked 12,365,836 research outputs across all sources so far. This one is in the 42nd percentile – i.e., 42% of other outputs scored the same or lower than it.
So far Altmetric has tracked 1,086 research outputs from this source. They typically receive more attention than average, with a mean Attention Score of 8.5. This one is in the 36th percentile – i.e., 36% of its peers scored the same or lower than it.
Older research outputs will score higher simply because they've had more time to accumulate mentions. To account for age we can compare this Altmetric Attention Score to the 218,884 tracked outputs that were published within six weeks on either side of this one in any source. This one has gotten more attention than average, scoring higher than 57% of its contemporaries.
We're also able to compare this research output to 15 others from the same source and published within six weeks on either side of this one. This one is in the 20th percentile – i.e., 20% of its contemporaries scored the same or lower than it.