confint.margstd_boot

Confidence intervals for models fit using marginal standardization
based on parametric bootstrapping.

Risk ratios and risk differences are estimated using regression
models that allow for binary, categorical, and continuous exposures and
confounders. Implemented are marginal standardization after fitting logistic
models (g-computation) with delta-method and bootstrap standard errors,
Miettinen's case-duplication approach (Schouten et al. 1993,
<doi:10.1002/sim.4780121808>), log-binomial (Poisson) models with empirical
variance (Zou 2004, <doi:10.1093/aje/kwh090>), binomial models with starting
values from Poisson models (Spiegelman and Hertzmark 2005,
<doi:10.1093/aje/kwi188>), and others.

Konrad Stopsack

risks

Estimate Risk Ratios and Risk Differences using Regression

Travis Gerke

confint.margstd_boot function

<dl><dt>object</dt>
<dd>Model fitted through marginal standardization</dd>
<dt>parm</dt>
<dd>Not used, for compatibility</dd>
<dt>level</dt>
<dd>Confidence level, defaults to 0.95.</dd>
<dt>bootrepeats</dt>
<dd>Bootstrap repeats. Defaults to 1000. Consider increasing.</dd>
<dt>bootci</dt>
<dd>Type of bootstrap confidence interval:<ul>
<li><code>"bca"</code> Default. Parametric BCa (bias-corrected accelerated)
confidence intervals.</li>
<li><code>"normal"</code> Parametric normality-based confidence intervals,
which require lower repeat numbers but are less accurate and
may result in invalid results for ratios.</li>
<li><code>"nonpar"</code> Non-parametric BCa confidence intervals,
which should be used with caution because of the risk
of sparse-data bias with non-parametric bootstrapping.</li>
</ul></dd>
<dt>jacksd</dt>
<dd>Return jackknife Monte-Carlo error for the confidence limits?
Only functional with BCa confidence intervals. Defaults to <code>FALSE</code>.</dd>
<dt>...</dt>
<dd>Not used</dd></dl>

Arguments

Bootstrap confidence intervals — confint.margstd_boot

<dl>

<dt>object</dt>
<dd>Model fitted through marginal standardization</dd>


<dt>parm</dt>
<dd>Not used, for compatibility</dd>


<dt>level</dt>
<dd>Confidence level, defaults to 0.95.</dd>


<dt>bootrepeats</dt>
<dd>Bootstrap repeats. Defaults to 1000. Consider increasing.</dd>


<dt>bootci</dt>
<dd>Type of bootstrap confidence interval:<ul>
<li><code>"bca"</code> Default. Parametric BCa (bias-corrected accelerated)
confidence intervals.</li>
<li><code>"normal"</code> Parametric normality-based confidence intervals,
which require lower repeat numbers but are less accurate and
may result in invalid results for ratios.</li>
<li><code>"nonpar"</code> Non-parametric BCa confidence intervals,
which should be used with caution because of the risk
of sparse-data bias with non-parametric bootstrapping.</li>
</ul></dd>


<dt>jacksd</dt>
<dd>Return jackknife Monte-Carlo error for the confidence limits?
Only functional with BCa confidence intervals. Defaults to <code>FALSE</code>.</dd>


<dt>...</dt>
<dd>Not used</dd>

</dl>

Data engineering and BI courses are free!

confint.margstd_boot: Bootstrap confidence intervals

Description

Usage

Value

Arguments