package baseCode.math.metaanalysis;

 import baseCode.math.Constants;

 import cern.colt.list.DoubleArrayList;

 import cern.jet.stat.Descriptive;

 import cern.jet.stat.Probability;

 /**

  * Statistics for meta-analysis. Methods from Cooper and Hedges (CH); Hunter and Schmidt (HS).

  * <p>

  * In this class "conditional variance" means the variance for one data set. Unconditional means "between data set", or

  * across data set.

  * <p>

  * Copyright (c) 2004

  * </p>

  * <p>

  * Institution: Columbia University

  * </p>

  * 

  * @author Paul Pavlidis

  * @version $Id: MetaAnalysis.java,v 1.5 2005/03/21 18:01:03 pavlidis Exp $

  */

    /**

     * Fisher's method for combining p values. (Cooper and Hedges 15-8)

     * 

     * @param pvals DoubleArrayList

     * @return double

     */

       }

    }

    /**

     * Fisher's method for combining p values (Cooper and Hedges 15-8)

     * <p>

     * Use for p values that have already been log transformed.

     * 

     * @param pvals DoubleArrayList

     * @return double

     */

       }

    }

    /**

     * The "Q" statistic used to test homogeneity of effect sizes. (Cooper and Hedges 18-6)

     * 

     * @param effectSizes DoubleArrayList

     * @param variances DoubleArrayList

     * @param globalMean double

     * @return double

     */

       }

    }

    /**

     * Test for statistical significance of Q.

     * 

     * @param Q - computed using qStatistic

     * @param N - number of studies.

     * @see qStatistic

     * @return The upper tail chi-square probability for Q with N - degrees of freedom.

     */

    }

    /**

     * General formula for weighted mean of effect sizes. Cooper and Hedges 18-1, or HS pg. 100.

     * <p>

     * In HS, the weights are simply the sample sizes. For CH, the weights are 1/v for a fixed effect model. Under a

     * random effects model, we would use 1/(v + v_bs) where v_bs is the between-studies variance.

     * 

     * @param effectSizes

     * @param sampleSizes

     * @return

     */

       }

    }

    /**

     * General formula for weighted mean of effect sizes including quality index scores for each value. Cooper and Hedges

     * 18-1, or HS pg. 100.

     * 

     * @param effectSizes

     * @param sampleSizes

     * @param qualityIndices

     * @return

     */

       }

    }

    /**

     * CH 18-3. Can be used for fixed or random effects model, the variances just have to computed differently.

     * 

     * <pre>      

     *          v_dot = 1/sum_i=1ˆk ( 1/v_i) 

     * </pre>

     * 

     * @param variances

     * @return

     */

       }

          //   throw new IllegalStateException( "Variance of zero" );

       }

    }

    /**

     * CH 18-3 version 2 for quality weighted. ( page 266 ) in Fixed effects model.

     * 

     * <pre>

     * 

     *  

     *   

     *    

     *     

     *      

     *       

     *          v_dot = [ sum_i=1ˆk ( q_i ˆ 2 * w_i) ]/[ sum_i=1ˆk  q_i * w_i ]ˆ2

     *        

     *       

     *      

     *     

     *    

     *   

     *  

     * </pre>

     * 

     * @param variances

     * @return

     */

       }

       }

    }

    /**

     * Test statistic for H0: effectSize == 0. CH 18-5. For fixed effects model.

     * 

     * @param metaEffectSize

     * @param metaVariance

     * @return

     */

    }

    /**

     * CH sample variance under random effects model, equation 18-20

     * 

     * @param

     * @return

     */

    }

    /**

     * CH estimate of between-studies variance, equation 18-22, for random effects model.

     * 

     * @param effectSizes

     * @param variances

     * @return

     */

    //   protected double metaREVariance( DoubleArrayList effectSizes,

    //         DoubleArrayList variances ) {

    //      return Math.max( metaRESampleVariance( effectSizes )

    //            - Descriptive.mean( variances ), 0.0 );

    //   }

    /**

     * CH equation 18-23. Another estimator of the between-studies variance s<sup>2 </sup> for random effects model.

     * This is non-zero only if Q is larger than expected under the null hypothesis that the variance is zero.

     * 

     * <pre>

     * 

     *  

     *   

     *    sˆ2 = [Q - ( k - 1 ) ] / c

     *    

     *   

     *  

     * </pre>

     * 

     * where

     * 

     * <pre>

     * 

     *  

     *   

     *    c = Max(sum_i=1ˆk w_i - [ sum_iˆk w_iˆ2 / sum_iˆk w_i ], 0)

     *    

     *   

     *  

     * </pre>

     * 

     * @param effectSizes

     * @param variances

     * @param weights

     * @return

     */

       // the weighted unconditional variance.

    }

    /**

     * Under a random effects model, CH eqn. 18-24, we replace the conditional variance with the sum of the

     * between-sample variance and the conditional variance.

     * 

     * <pre>

     * 

     *  

     *     

     *    v_iˆ* = sigma-hat_thetaˆ2 + v_i.

     *    

     *   

     *  

     * </pre>

     * 

     * @param variances Conditional variances

     * @param sampleVariance estimated...somehow.

     * @return

     */

          }

       }

    }

    /**

     * Weights under a fixed effects model. Simply w_i = 1/v_i. CH eqn 18-2.

     * @param variances 

     * @return

     */

          }

       }

    }

 }