ocmath_anova_mean_comparison

1 Description
2 Syntax
3 Parameters
4 Return
5 Examples
6 Remark
7 See Also
8 header to Included
9 Reference

Description

Function to compute One-Way ANOVA Mean Comparisons. Computes simultaneous confidence intervals for the differences between means.

Syntax

int ocmath_anova_mean_comparison( const double * pMean, uint nSizeMean, const int * pNPTS, uint nSizeNPTs, const double dMSE, const int nDOF, const double dAlpha, const int nGroupSize, int * pIndexComparison, uint nSizeComparison, double * pMeanDiffs, uint nSizeMeanDiff, double * pLCLs, uint nSizeLCL, double * pUCLs, uint nSizeUCL, double * pSe, uint nSizeSe, double * pStat, uint nSizeStat, double * pProb, uint nSizeProb, long * pSig, uint nSizeSig, double * pAlpha, uint nSizeAlpha, int nIntervalType )

Parameters

pMean

Input pointer to Mean of each group

nSizeMean

Input size of pointer pMean, which should not be less than nGroupSize

pNPTS

Input pointer to size of each group

nSizeNPTs

Input size of pointer pNPTs, which should not be less than nGroupSize

dMSE

Input Mean Square for Error term of the model

nDOF

Input Degree of Freedom for Error term of the model

dAlpha

Input number indicates the significant level of the model

nGroupSize

Input size indicates group number.

pIndexComparison

Output pointer to order of means comparison

nSizeComparison

Input size of pointer pIndexComparison, which should not be less than (nGroupSize*(nGroupSize-1)/2)

pMeanDiffs

Output pointer to difference of means of two Samples

nSizeMeanDiff

Input size of pointer pMeanDiffs, which should not be less than (nGroupSize*(nGroupSize-1)/2)

pLCLs

Output pointer to lower confidence limits for difference of means

nSizeLCL

Input size of pointer pLCLs, which should not be less than (nGroupSize*(nGroupSize-1)/2)

pUCLs

Output pointer to upper confidence limits for difference of means

nSizeUCL

Input size of pointer pUCLs, which should not be less than (nGroupSize*(nGroupSize-1)/2)

pSe

Output pointer to Standard Errors of difference between means

nSizeSe

Input size of pointer pSe, which should not be less than (nGroupSize*(nGroupSize-1)/2)

pStat

Output pointer to Statistic used to test difference between means

nSizeStat

Input size of pointer pStat, which should not be less than (nGroupSize*(nGroupSize-1)/2)

pProb

Output pointer to Probability of the test for difference between means

nSizeProb

Input size of pointer pProb, which should not be less than (nGroupSize*(nGroupSize-1)/2)

pSig

Output pointer to indicator whether the difference between means is significant or not.

pSig[i] =1: the difference is significant

pSig[i] =0: the difference is NOT significant i=0, 1, ..., nSizeSig-1

nSizeSig

Input size of pointer pSig, which should not be less than (nGroupSize*(nGroupSize-1)/2)

pAlpha

Output pointer to significance level that depends upon the test

nSizeAlpha

Input size of pointer pAlpha, which should not be less than (nGroupSize*(nGroupSize-1)/2)

nIntervalType

Input interger that indicates which method to be used. Optional choice included as below

ANOVA_BONFERRON = 1: the Bonferroni method is used
ANOVA_TUKEY = 2: the Tukey-Kramer method is used
ANOVA_SCHEFFE = 3: the Scheffe method is used
ANOVA_SIDAK = 4: the Dunn-Sidak method is used
ANOVA_FISHER = 5: the Fisher's LSD method is used
ANOVA_BONHOLM = 6: the Bonferroni-Holm method is used
ANOVA_SIDAKHOLM = 7: the Sidak-Holm method is used
ANOVA_DUNNETT = 8: the Dunnett method is used

Return

Returns STATS_NO_ERROR on successful exit and a non-zero STATS error code on failure.

Examples

EX1

void ocmath_anova_mean_comparison_ex1()
{
    vector pData ={77.7111,88.07729,92.15046,77.70871,72.25362,  //Group 1
                     83.52885,75.17097,87.96739,91.59955,81.93563, //Group 2
                     78.15473,79.42937,83.52293,71.4648,72.90409 }; //Group 3
    int nSize = pData.GetSize();
    vector<int> pIndex = { 5,5,5 };
    int iLevels = pIndex.GetSize();
 
    ANOVADescStats DescStats[3];
    ANOVAStats TotalStats;
    RegANOVA ANOVATable;
 
    int nRet = ocmath_anova_one_way(pData, nSize, iLevels, pIndex, DescStats, 3, &TotalStats, &ANOVATable);
 
    vector vMeans(3), vMeanDiffs(3), vLCLs(3), vUCLs(3), vSE(3), vStats(3), vProbs(3), vAlphas(3);
    vector<int> vnIndexComp(3), vnSigs(3);
    vMeans[0] = DescStats[0].Mean;
    vMeans[1] = DescStats[1].Mean;
    vMeans[2] = DescStats[2].Mean;
    
    double dMSE = ANOVATable.AnovaTable[1].MeanSq;
    int DF = ANOVATable.AnovaTable[1].DOF;
    
    nRet = ocmath_anova_mean_comparison(vMeans, 3, pIndex, iLevels, dMSE, DF, 0.05, 3, vnIndexComp, 3, vMeanDiffs, 3, vLCLs, 3, vUCLs, 3, vSE, 3, vStats, 3, vProbs, 3, vnSigs, 3, vAlphas, 3, ANOVA_TUKEY);
    //print result if no error
    if (0 == nRet)
    {
        vector<string> vstrMeanComp = {"2 v.s.1", "3 v.s. 1", "3 v.s. 2"};
            for (int ii = 0; ii < 3; ii++)
            {
                printf("Mean Comparison Results (%s):\n", vstrMeanComp[ii]);
                out_double("MeanDiff = ", vMeanDiffs[ii]);
                    out_double("SEM = ", vSE[ii]);
                    out_double("t-val = ", vStats[ii]);
                    out_double("LCL = ", vLCLs[ii]);
                    out_double("UCL = ", vUCLs[ii]);
                    printf("\n");
        }
    }
}