analyses:konstantopoulos2011
Differences
This shows you the differences between two versions of the page.
Next revision | Previous revisionNext revisionBoth sides next revision | ||
analyses:konstantopoulos2011 [2018/12/08 12:56] – external edit 127.0.0.1 | analyses:konstantopoulos2011 [2020/05/01 14:01] – Wolfgang Viechtbauer | ||
---|---|---|---|
Line 21: | Line 21: | ||
</ | </ | ||
<code output> | <code output> | ||
- | | + | |
- | 1 11 1 1976 -0.18 0.118 | + | 1 11 |
- | 2 11 2 1976 -0.22 0.118 | + | 2 11 |
- | 3 11 3 1976 0.23 0.144 | + | 3 11 |
- | 4 11 4 1976 -0.30 0.144 | + | 4 11 |
- | 5 12 5 1989 0.13 0.014 | + | 5 12 |
- | 6 12 6 1989 -0.26 0.014 | + | 6 12 |
- | 7 12 7 1989 0.19 0.015 | + | 7 12 |
- | 8 12 8 1989 0.32 0.024 | + | 8 12 |
- | 9 18 9 1994 0.45 0.023 | + | 9 18 |
- | 10 | + | 10 |
- | 11 | + | 11 |
- | 12 | + | 12 |
- | 13 | + | 13 |
- | 14 | + | 14 |
- | 15 | + | 15 |
</ | </ | ||
So, 4 studies were conducted in district 11, 4 studies in district 12, 3 studies in district 18, and so on. Variables '' | So, 4 studies were conducted in district 11, 4 studies in district 12, 3 studies in district 18, and so on. Variables '' | ||
Line 76: | Line 76: | ||
4 | 4 | ||
</ | </ | ||
- | So, as noted in the article, the data have an unbalanced structure, with the number of studies per district ranging from 3 to 11 ('' | + | So, as noted in the article, the data have an unbalanced structure, with the number of studies/ |
To obtain the descriptives about the effect size estimates per district (Table 3 in the paper), we can use: | To obtain the descriptives about the effect size estimates per district (Table 3 in the paper), we can use: | ||
Line 99: | Line 99: | ||
==== Two-Level Model ==== | ==== Two-Level Model ==== | ||
- | First, a standard (two-level) random-effects model is fitted to the data. We can do the same with: | + | First, a standard (two-level) random-effects model is fitted to the data. Here, we treat the 56 studies as independent (which we later will see is not justified). We can fit such a model with: |
<code rsplus> | <code rsplus> | ||
res <- rma(yi, vi, data=dat) | res <- rma(yi, vi, data=dat) | ||
Line 112: | Line 112: | ||
H^2 (total variability / sampling variability): | H^2 (total variability / sampling variability): | ||
- | Test for Heterogeneity: | + | Test for Heterogeneity: |
Q(df = 55) = 578.864, p-val < .001 | Q(df = 55) = 578.864, p-val < .001 | ||
Model Results: | Model Results: | ||
- | estimate | + | estimate |
- | | + | |
--- | --- | ||
Line 148: | Line 148: | ||
R^2 (amount of heterogeneity accounted for): 0.00% | R^2 (amount of heterogeneity accounted for): 0.00% | ||
- | Test for Residual Heterogeneity: | + | Test for Residual Heterogeneity: |
QE(df = 54) = 550.260, p-val < .001 | QE(df = 54) = 550.260, p-val < .001 | ||
- | Test of Moderators (coefficient(s) 2): | + | Test of Moderators (coefficient 2): |
QM(df = 1) = 1.383, p-val = 0.240 | QM(df = 1) = 1.383, p-val = 0.240 | ||
Model Results: | Model Results: | ||
- | estimate | + | estimate |
- | intrcpt | + | intrcpt |
- | I(year - mean(year)) | + | I(year - mean(year)) |
--- | --- | ||
Line 173: | Line 173: | ||
Multivariate Meta-Analysis Model (k = 56; method: REML) | Multivariate Meta-Analysis Model (k = 56; method: REML) | ||
- | Variance Components: | + | Variance Components: |
- | | + | |
- | sigma^2 | + | sigma^2 |
- | Test for Heterogeneity: | + | Test for Heterogeneity: |
Q(df = 55) = 578.864, p-val < .001 | Q(df = 55) = 578.864, p-val < .001 | ||
Model Results: | Model Results: | ||
- | estimate | + | estimate |
- | | + | |
--- | --- | ||
Signif. codes: | Signif. codes: | ||
</ | </ | ||
- | The '' | + | The '' |
==== Three-Level Model ==== | ==== Three-Level Model ==== | ||
Line 201: | Line 201: | ||
Multivariate Meta-Analysis Model (k = 56; method: REML) | Multivariate Meta-Analysis Model (k = 56; method: REML) | ||
- | Variance Components: | + | Variance Components: |
- | | + | |
- | sigma^2.1 | + | sigma^2.1 |
- | sigma^2.2 | + | sigma^2.2 |
- | Test for Heterogeneity: | + | Test for Heterogeneity: |
Q(df = 55) = 578.864, p-val < .001 | Q(df = 55) = 578.864, p-val < .001 | ||
Model Results: | Model Results: | ||
- | estimate | + | estimate |
- | | + | |
--- | --- | ||
Line 219: | Line 219: | ||
</ | </ | ||
These results correspond to those given on the left-hand side of Table 5 in the paper. Somewhat confusingly, | These results correspond to those given on the left-hand side of Table 5 in the paper. Somewhat confusingly, | ||
+ | |||
+ | **Note**: We would obtain the same results when using '' | ||
==== Profile Likelihood Plots ==== | ==== Profile Likelihood Plots ==== | ||
- | Whenever we start fitting more complicated models with the '' | + | Whenever we start fitting more complicated models with the '' |
<code rsplus> | <code rsplus> | ||
par(mfrow=c(2, | par(mfrow=c(2, | ||
Line 244: | Line 246: | ||
[1] 0.665 | [1] 0.665 | ||
</ | </ | ||
- | Therefore, the underlying true effects within districts are estimated to correlate quite strongly. | + | Therefore, the underlying true effects within districts are estimated to correlate quite strongly |
Also, it is worth noting that the sum of the two variance components can be interpreted as the total amount of heterogeneity in the true effects: | Also, it is worth noting that the sum of the two variance components can be interpreted as the total amount of heterogeneity in the true effects: | ||
Line 264: | Line 266: | ||
Multivariate Meta-Analysis Model (k = 56; method: REML) | Multivariate Meta-Analysis Model (k = 56; method: REML) | ||
- | Variance Components: | + | Variance Components: |
outer factor: district | outer factor: district | ||
- | inner factor: factor(study) (nlvls = 11) | + | inner factor: factor(study) (nlvls = 56) |
- | | + | |
- | tau^2 0.098 0.313 no | + | tau^2 0.098 0.313 no |
- | rho 0.665 no | + | rho 0.665 no |
- | Test for Heterogeneity: | + | Test for Heterogeneity: |
Q(df = 55) = 578.864, p-val < .001 | Q(df = 55) = 578.864, p-val < .001 | ||
Model Results: | Model Results: | ||
- | estimate | + | estimate |
- | | + | |
--- | --- | ||
Signif. codes: | Signif. codes: | ||
</ | </ | ||
- | The '' | + | The '' |
As long as $\rho$ is estimated to be positive, the multilevel and multivariate parametrizations are in essence identical. In fact, the log likelihoods of the two models should be identical, which we can confirm with: | As long as $\rho$ is estimated to be positive, the multilevel and multivariate parametrizations are in essence identical. In fact, the log likelihoods of the two models should be identical, which we can confirm with: | ||
Line 309: | Line 311: | ||
Again, both plots indicate that the estimates obtained in fact correspond to the peaks of the respective likelihood profiles, with decreasing log likelihoods as the values of parameters are moved away from the actual estimates. | Again, both plots indicate that the estimates obtained in fact correspond to the peaks of the respective likelihood profiles, with decreasing log likelihoods as the values of parameters are moved away from the actual estimates. | ||
- | Since the log likelihood drops of quite dramatically when $\rho$ is set equal to a value very close to 1, the left-hand side of the profile gets ' | + | Since the log likelihood drops of quite dramatically when $\rho$ is set equal to a value very close to 1, the left-hand side of the profile gets ' |
==== Uncorrelated Sampling Errors ==== | ==== Uncorrelated Sampling Errors ==== | ||
Line 324: | Line 326: | ||
Multivariate Meta-Analysis Model (k = 56; method: REML) | Multivariate Meta-Analysis Model (k = 56; method: REML) | ||
- | Variance Components: | + | Variance Components: |
- | | + | |
- | sigma^2 | + | sigma^2 |
- | Test for Heterogeneity: | + | Test for Heterogeneity: |
Q(df = 55) = 578.864, p-val < .001 | Q(df = 55) = 578.864, p-val < .001 | ||
Model Results: | Model Results: | ||
- | estimate | + | estimate |
- | | + | |
--- | --- | ||
Line 352: | Line 354: | ||
Multivariate Meta-Analysis Model (k = 56; method: REML) | Multivariate Meta-Analysis Model (k = 56; method: REML) | ||
- | Variance Components: | + | Variance Components: |
- | | + | |
- | sigma^2.1 | + | sigma^2.1 |
- | sigma^2.2 | + | sigma^2.2 |
outer factor: district | outer factor: district | ||
inner factor: factor(study) (nlvls = 56) | inner factor: factor(study) (nlvls = 56) | ||
- | | + | |
- | tau^2 0.030 0.174 no | + | tau^2 0.030 0.174 no |
- | rho 0.784 no | + | rho 0.784 no |
- | Test for Heterogeneity: | + | Test for Heterogeneity: |
Q(df = 55) = 578.864, p-val < .001 | Q(df = 55) = 578.864, p-val < .001 | ||
Model Results: | Model Results: | ||
- | estimate | + | estimate |
- | | + | |
--- | --- |
analyses/konstantopoulos2011.txt · Last modified: 2022/08/22 16:00 by Wolfgang Viechtbauer