hsls: High School Longitudinal Study

In mvmeta: Multivariate and Univariate Meta-Analysis and Meta-Regression

Description

This is a nationally representative, longitudinal study of more than 21,000 9th graders in 944 schools who will be followed through their secondary and postsecondary years. The data are used for testing whether sex, socioeconomic status and sex by socio-economic status interaction are predictive of the mathematics standardized score in each of the eight race groups.

Usage

hsls

Format

A data frame with 8 observations on the following 10 variables:

• race: race group.

• b1, b2, b3: estimated regression coefficients for sex, socio-economic status and sex by socio-economic status interaction, respectively, on the mathematics standardized score.

• V11, V22, V33: variances of the estimated coefficients.

• V12, V13, V23: covariances of the estimated coefficients.

Source

Chen H, Manning AK, Dupuis J (2012). A method of moments estimator for random effect multivariate meta-analysis. Biometrics. 68(4):1278-1284.

Examples

### REPRODUCE THE RESULTS IN CHEN ET AL. (2012)

# INSPECT THE DATA
hsls

# FIXED-EFFECTS MODEL
S <- as.matrix(hsls[5:10])
model <- mvmeta(cbind(b1,b2,b3),S,data=hsls,method="fixed")
summary(model)

# MM MODEL
model <- mvmeta(cbind(b1,b2,b3),S,data=hsls,method="mm")
summary(model)
model$Psi

Example output

This is mvmeta 0.4.11. For an overview type: help('mvmeta-package').
  race      b1     b2      b3    V11     V12     V13     V22      V23    V33
1    1  0.3161 7.4015  0.4278 2.3568 -1.2105  0.8524  9.7029  -6.1753 4.4114
2    2 -0.3201 6.9426 -0.9816 0.2529  0.1498 -0.1019  0.7016  -0.4167 0.2743
3    3  0.6983 4.6680 -0.2415 0.1444 -0.0652  0.0433  0.6481  -0.3899 0.2608
4    4  3.2736 4.3080  0.2052 3.8428 -4.5587  3.2892 10.3517  -6.6684 4.8268
5    5 -0.1599 5.6398 -0.6782 0.1161 -0.0992  0.0645  0.4363  -0.2610 0.1733
6    6 -0.6989 6.3158 -0.7918 0.1603  0.0242 -0.0129  0.7697  -0.4686 0.3180
7    7 -3.6094 9.3429 -2.8711 3.2054 -1.1984  0.8437 17.8889 -10.7697 7.2101
8    8  0.2172 6.4078 -0.6093 0.0278  0.0136 -0.0091  0.1184  -0.0716 0.0482
Call:  mvmeta(formula = cbind(b1, b2, b3) ~ 1, S = S, data = hsls, method = "fixed")

Multivariate fixed-effects meta-analysis
Dimension: 3

Fixed-effects coefficients
    Estimate  Std. Error        z  Pr(>|z|)  95%ci.lb  95%ci.ub     
b1    0.0799      0.1208   0.6615    0.5083   -0.1568    0.3166     
b2    6.2031      0.2448  25.3398    0.0000    5.7234    6.6829  ***
b3   -0.6591      0.1550  -4.2527    0.0000   -0.9629   -0.3554  ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 

Multivariate Cochran Q-test for heterogeneity:
Q = 54.6278 (df = 21), p-value = 0.0001
I-square statistic = 61.6%

8 studies, 24 observations, 3 fixed and 0 random-effects parameters
  logLik       AIC       BIC  
-35.7329   77.4659   81.0000  

Call:  mvmeta(formula = cbind(b1, b2, b3) ~ 1, S = S, data = hsls, method = "mm")

Multivariate random-effects meta-analysis
Dimension: 3
Estimation method: Method of moments

Fixed-effects coefficients
    Estimate  Std. Error        z  Pr(>|z|)  95%ci.lb  95%ci.ub     
b1   -0.0604      0.2684  -0.2250    0.8220   -0.5864    0.4656     
b2    6.1821      0.2887  21.4109    0.0000    5.6162    6.7480  ***
b3   -0.7009      0.1894  -3.6996    0.0002   -1.0722   -0.3296  ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 

Between-study random-effects (co)variance components
Structure: General positive-definite
    Std. Dev      Corr          
b1    0.5296        b1        b2
b2    0.3201  -0.55870          
b3    0.2308   0.02538   0.81492
(Note: Truncated estimate - 1 negative eigenvalues set to 0)

Multivariate Cochran Q-test for heterogeneity:
Q = 54.6278 (df = 21), p-value = 0.0001
I-square statistic = 61.6%

8 studies, 24 observations, 3 fixed and 1 random-effects parameters

             b1          b2          b3
b1  0.280483923 -0.09472445 0.003102559
b2 -0.094724450  0.10248384 0.060208554
b3  0.003102559  0.06020855 0.053263783

mvmeta documentation built on Dec. 10, 2019, 5:07 p.m.