Meta-Analysis Binary: Bias

The Bias dialog provides settings for enabling the publication bias by conducting regression-based tests for meta-analysis with binary outcomes on raw data that are provided in the active dataset for the estimation of the effect size.

Regression-based Tests
Provides options for specifying the regression based tests. More than one test can be selected.
Egger's test
When selected, conducts Egger's test.
Harbord's test
When selected, conducts Harbord's test. The test is available when the Effect Size is specified as Logs Odd Ratio or Log Risk Ratio.
Peters' test
When selected, conducts Peters' test. The test is available when the Effect Size is specified as Logs Odd Ratio.
Variables
The list provides all available dataset variables.
Covariate(s)
Variables selected from the Variables list are treated as covariates. Multiple covariates are allowed.
Factor(s)
Variables selected from the Variables list are treated as factors. Multiple factors are allowed.
Include intercept in regression
Controls the intercept term in the regression-based test.
Include dispersion parameter in fixed-effects model
Controls the multiplicative model setting and introduces the multiplicative dispersion parameter to the analysis. The setting is available only when a fixed-effects model is selected.
Estimate statistics based on t-distribution
Controls the distribution used in the regression-based tests. The setting is enabled by default, which estimates the statistics based on the t-distribution. When the setting is not selected, the statistics are estimated based on the normal distribution.

Defining Meta-Analysis Binary bias settings

  1. From the menus choose:

    Analyze > Meta Analysis > Binary Outcomes > Raw Data...

  2. In the Meta-Analysis Binary dialog, click Bias.
  3. Select and define the appropriate bias settings.
  4. Click Continue.