Standards for statistical reporting for the Journal of Studies on Alcohol and Drugs
We at the Journal of Studies on Alcohol and Drugs want to be certain that the Journal demonstrates as much consistency as possible regarding the manner in which statistical issues are presented to our readers. Therefore, we are routinely sending to all authors of potential revisions the following suggested abbreviations and standards. We would be grateful if you would incorporate these suggestions into your revisions.
Regression
| b |
The
unstandardized regression coefficient (weight); b has a
standard error (SE). |
| ß |
The standardized
regression coefficient (weight). ß does not have
a standard error. It cannot be calculated. |
| r |
The Pearson
zero-order correlation coefficient. |
| r2 |
The proportion
of variance shared by two variables that have a
correlation of r. |
| R2 |
The proportion
of variance accounted for by all predictors in a
regression model. |
| ΔR2 |
The increase
in R, or change in R. R has an F test. |
| adj R |
The estimate
of R in the population. adj R does not have an F test. |
| s r2 |
The squared
semi-partial correlation; unique variance. |
| OR |
Odds ratio.
OR has an asymmetric confidence interval. |
| adj OR |
Odds ratio.
OR has an asymmetric confidence interval. |
Structural Equation Modeling
| SEM |
Structural equation
modeling. |
| RMSEA |
Root means square
error of approximation. |
Other
| t |
t test |
| F |
F test |
| χ2 |
Chi square |
| p |
as in p < .05 |
| z |
as in
z = (x - μ) / σ |
| df |
Degrees of freedom |
| t, χ2,
F |
t and χ2 require df;
F requires df for numerator and denominator. |
Confidence Intervals
| mean (SD) |
SD (standard deviation).
Do not place ± between mean and SD. |
| parameter
(SE) |
Parameter indicate
mean, b, OR, etc. Do not place ± between
parameter and SE (standard error).
|
