Finding Standard Deviations

Download as xls, pdf, or txt
Download as xls, pdf, or txt
You are on page 1of 4

Finding Standard Deviations

When extracting data for outcomes of continuous variables (e.g. pain, anxiety, blood pressure, heart rate), you need to find the mean score
for each group, the associated standard deviation of the mean for each group, and the number of people in each group.

Some times the standard deviation is not presented in the article. It is possible to derive standard deviations from other related information,
such as standard errors, confidence intervals, p-values, and t-values.

There are formulas to make these calculations, and these are presented in the Cochrane Handbook. This tool has been created to help make
the calculations for you. If you are uncertain, you should still consult a statistician for advice.

When using this tool, be aware that it may alert you if your data looks odd, or if you need to seek further help from a statistician.
Some times these alerts pop-up prematurely, so enter all your data before paying attention to them!

There are two tabs below, called "Within Groups" and "Between Groups". Within Groups is to be used to derive standard deviations where the
information you have is either a standard error for each study group, or a confidence interval for each study group. Between Groups is to be used
where the information you have, has been derived from a t-test to compare the mean scores between the two study groups, or a standard error
or confidence interval that is associated with the mean difference of scores between groups.

To complete the calculations you need to enter the relevant data into the white boxes and the answers will appear in the yellow boxes.

If you want to derive standard deviations from information related to an F-value (from an ANOVA), please consult a statistician.
Finding Standard Deviations by using numbers calculated from with-in groups

IF YOU ARE UNCERTAIN OF ANYTHING PLEASE: CONSULT THE HANDBOOK

Finding the Standard Deviation using the Standard Error:


When making this transformation, standard errors must be of means calculated from within an intervention group and
not standard errors of the difference in means computed between study groups.

GROUP 1 GROUP 2
Standard Error = Standard Error =
Number of participants = Number of participants =

Standard Deviation = 0 Standard Deviation = 0

Calculation used: SD = SE x √N

Finding the Standard Deviation using Confidence Intervals (CI):


This transformation is for Confidence Intervals for mean values calculated within a study group and not for estimates of differences between study groups.
It is important to check that the confidence interval is symmetrical about the mean (the distance between the lower limit and the mean is the same as the distance
between the mean and the upper limit). If this is not the case, the confidence interval may have been calculated on transformed values.

GROUP 1 GROUP 2
What % was the CI? (for 95% CI type 0.95): What % was the CI? (for 95% CI type 0.95):
Number of participants = Number of participants =
Upper Confidence Interval limit = Upper Confidence Interval limit =
Lower Confidence Interval limit = Lower Confidence Interval limit =
Mean = Mean =

The Confidence Interval is Err:502 standard errors wide. The Confidence Interval is Err:502 standard errors wide.

Standard Deviation = Err:502 Standard Deviation = Err:502


Finding Standard Deviations by using numbers calculated from between groups
Standard deviations can be obtained from standard errors, confidence intervals, t values or P values that relate to the
The difference in means itself (MD) is required in the calculations from the t value or the P value.
These calculations are based on an assumption that the standard deviations of outcome measurements are the same
so the one standard deviation derived would then be entered into RevMan for both study groups.

Using the p-value of a t-test to find the SD. Using the t-value to find the SD

First you must derive the corresponding t-value: First you must derive the corresponding Standard E

p-value = 0.008 t-value = 2.78

No. of participants in Group 1 = 25 Difference in means (between groups) =

No. of participants in Group 2 = 22


Standard Error =
t-value = 2.775545
Now go to the calculation box for t-values. Now go to the calculation box for Standard Errors.

Using the Confidence Intervals to find the SD Using the Standard Error to find the SD

First you must derive the Standard Error:

Upper CI limit = 60 Standard Error of the difference in means =


Lower CI limit = 40
% CI used (enter 0.95 for 95%) = 0.95 No. of participants in Group 1 =
No. of participants in Group 1 = 60
No. of participants in Group 2 = 75 No. of participants in Group 2 =

(divisor = 3.92 )
Estimated SD for each group =
Standard Error = 5.102041
Now go to the calculation box for Standard Errors.
ed from between groups
es or P values that relate to the differences between means in two groups.

me measurements are the same in both groups,

o find the SD

e the corresponding Standard Error:

(between groups) = 3.8

1.3669065 If unsure, please:


CONSULT THE HANDBOOK
tion box for Standard Errors.

d Error to find the SD

e difference in means = 1.37

25

22

4.686547

You might also like