Combine means and SDs

Enter the sample size, mean and standard deviation of each group. Groups are combined in order with Cochrane's formula; rows left empty are skipped.

Groups

GroupSample size (n)MeanSDRemove

Combined

N=N1+N2 M=N1M1+N2M2N1+N2 SD=(N1−1)SD12+(N2−1)SD22+N1N2N1+N2(M12+M22−2M1M2)N1+N2−1

Fill in at least one group.

Source: Cochrane Handbook 6.5, section 6.5.2.10, table 6.5.a (opens in a new tab). With three or more groups, the formula is applied to each pair in turn.

Questions

What does it combine?

Two or more groups into the single group they would have formed if they had never been split: a pooled sample size, mean and standard deviation. The standard deviation accounts for the spread within each group and for the distance between their means, which is why it is not an average of the input SDs.

When is that the right thing to do?

When a trial reports arms your analysis has to treat as one. The usual cases are a multi-arm trial with two doses of the same drug against one control, subgroups reported separately with no overall row, or centres reported one by one. Combining first avoids counting a shared control group twice.

Which formula?

Cochrane Handbook 6.5, section 6.5.2.10, table 6.5.a (table 7.7.a in the older version 5.1), applied to one pair at a time. The running table shows each intermediate result, so the arithmetic can be checked rather than taken on trust. Rows left empty are skipped.

What it will not do

It does not weight by study precision, pool across studies or produce an effect estimate. This is one step of data preparation; the pooling itself belongs in your meta-analysis software.