Now you've got 2 approvals (MH and myself). OpenPsych allows authors to pick an external reviewer with qualifications (it's not for the editor to vet that as this would introduce subjectivity and potential abuse of power so I cannot forbid you from naming anyone. I'll just trust your good sense and of course I can raise my objections if I am not satisfied and other reviewers can do the same, but as editor I cannot veto it. However, if 3 reviewers are against your external reviewer, you'll have to find another one, but hopefully this won't be necessary).

Attached is the latest .doc and .xl file, along with the document provided by Hart et al.

Now you've got 2 approvals (MH and myself). OpenPsych allows authors to pick an external reviewer with qualifications (it's not for the editor to vet that as this would introduce subjectivity and potential abuse of power so I cannot forbid you from naming anyone. I'll just trust your good sense and of course I can raise my objections if I am not satisfied and other reviewers can do the same, but as editor I cannot veto it. However, if 3 reviewers are against your external reviewer, you'll have to find another one, but hopefully this won't be necessary).

Attached is the latest .doc and .xl file, along with the document provided by Hart et al.

Thanks. I will wait for the 3rd reviewer before proceeding with publication.

Here's the newest version of the article. I revised some language and deleted the mention of the biasing influence of assortative mating in our new analysis. As MH noted, assortative mating does not appear to bias estimates in our sibling model. I also added a note on why the decision to estimate only ACE models in our CNLSY study was reasonable (for the age range investigated the three components are usually present).

Let's say rMZ is 0.8 and rDZ is 0.3 (and there is no AM). This could be because there's D variance and no C variance. But it could as well be that there's both D and C variance but more D. In the classic twin design, C and D are confounded; they cannot be simultaneously estimated because of underidentification. You need additional types of relatives to estimate both C and D.

I do not understand. What you seem to say is that we need twin + non twin data. But, in what is it a problem inherent to ADE model ?

Nothing wrong with the ADE model as such, but if you have only twin data and there is in fact both C and D variance, you cannot estimate both and your results will be off.

I have attempted to replicate your results. I failed, and I will probably give up. NlsyLinks must hate me. You may want to look at your syntax again.

The syntax should be ok. What's the specific problem you're having?

Besides, if you want confidence intervals, given that your syntax ended up like this :

ace <-AceLavaanGroup(dsClean)
ace

In general, the thing to do is :

confint(ace, level=0.95)

That should do the trick.

I rather doubt that the CIs would be correct because NlsyLinks ignores non-independence across pairs.

The syntax should be ok. What's the specific problem you're having?

Remember your blog post on race-ACE interaction on CNLSY79 ? When I use your syntax, R stops at the beginning and says the variable C0000100 can't be found. That's nonsense because saw it in your data (the one you sent me few months ago). When I use the syntax John gave me, it works better:

rm(list=ls(all=TRUE))
library(NlsyLinks)
filePathOutcomes <-'c:\\Users\\John\\Desktop\\CNLSYALL\\CNLSYALLS10312013PLUSresbyyear3.csv'
dsOutcomes <- ReadCsvNlsy79Gen2(filePathOutcomes)
dsOutcomes <- RenameNlsyColumn(dsOutcomes, "C0005300", "Race")
dsOutcomesSubset <- dsOutcomes
dsOutcomesSubset <- dsOutcomes[dsOutcomes$Race==3, ]
dsLinks <- Links79PairExpanded
dsDouble <- CreatePairLinksDoubleEntered(outcomeDataset=dsOutcomesSubset, linksPairDataset=dsLinks, outcomeNames=c("RC94res"))
ace <- AceUnivariate( method="DeFriesFulkerMethod3", dataSet=dsDouble, oName_1="RC94res_1", oName_2="RC94res_2" )
ace
GetDetails(ace)
summary(GetDetails(ace))

But I have a message when I arrive at this point :

ace <- AceUnivariate( method="DeFriesFulkerMethod3", dataSet=dsDouble, oName_1="RC94res_1", oName_2="RC94res_2" )

unused argument(s) (oName_1 = "RC94res_1", oName_2 = "RC94res_2")

And this, is curious, because the two variables exist in the data. In general, I don't use the data in SPSS format, but rather, the .csv file you and John you gave me before. This package works for him and you, but not with me.

I rather doubt that the CIs would be correct because NlsyLinks ignores non-independence across pairs.

I couldn't verify, because in fact the NlsyLinks never run properly for me.

If you are willing to email people around for review, I can suggest it's best to email people who had works closest to yours. For example, Beaver (2013) referenced in your article has reported the heritability for blacks and whites. You can also try to find the mail of Guo and Hart. Perhaps Loehlin too.

The syntax should be ok. What's the specific problem you're having?

Remember your blog post on race-ACE interaction on CNLSY79 ? When I use your syntax, R stops at the beginning and says the variable C0000100 can't be found. That's nonsense because saw it in your data (the one you sent me few months ago). When I use the syntax John gave me, it works better:

rm(list=ls(all=TRUE))
library(NlsyLinks)
filePathOutcomes <-'c:\\Users\\John\\Desktop\\CNLSYALL\\CNLSYALLS10312013PLUSresbyyear3.csv'
dsOutcomes <- ReadCsvNlsy79Gen2(filePathOutcomes)
dsOutcomes <- RenameNlsyColumn(dsOutcomes, "C0005300", "Race")
dsOutcomesSubset <- dsOutcomes
dsOutcomesSubset <- dsOutcomes[dsOutcomes$Race==3, ]
dsLinks <- Links79PairExpanded
dsDouble <- CreatePairLinksDoubleEntered(outcomeDataset=dsOutcomesSubset, linksPairDataset=dsLinks, outcomeNames=c("RC94res"))
ace <- AceUnivariate( method="DeFriesFulkerMethod3", dataSet=dsDouble, oName_1="RC94res_1", oName_2="RC94res_2" )
ace
GetDetails(ace)
summary(GetDetails(ace))

But I have a message when I arrive at this point :

ace <- AceUnivariate( method="DeFriesFulkerMethod3", dataSet=dsDouble, oName_1="RC94res_1", oName_2="RC94res_2" )

unused argument(s) (oName_1 = "RC94res_1", oName_2 = "RC94res_2")

And this, is curious, because the two variables exist in the data. In general, I don't use the data in SPSS format, but rather, the .csv file you and John you gave me before. This package works for him and you, but not with me.

I tested my syntax again, and it works without problems for me. This is what I did:

1) Used the racially disaggregated SPSS files I sent you on 3/19/2014.

2) Saved the SPSS files as .csv.

3) Used the syntax included in the Excel file in John's latest post above.

Just change the file path, and it should work. I'm using OS X, the path formatting may be a bit different in Windows. Here's the syntax:

rm(list=ls(all=TRUE))
require(NlsyLinks)
dsLinking <- subset (Links79Pair, RelationshipPath=="Gen2Siblings")
filePathOutcomes <-'/Users/d/Desktop/blacks.csv'
dsOutcomes <- ReadCsvNlsy79Gen2(filePathOutcomes)
dsSingle <- CreatePairLinksSingleEntered (outcomeDataset=dsOutcomes, linksPairDataset=dsLinking, outcomeNames= c('DSB_AGE_7'))
oName_1 <-"DSB_AGE_7_1" #Stands for Outcome1
oName_2 <-"DSB_AGE_7_2" #Stands for Outcome2
dsGroupSummary <-RGroupSummary(dsSingle, oName_1, oName_2)
dsGroupSummary
dsClean <-CleanSemAceDataset(dsDirty=dsSingle, dsGroupSummary, oName_1, oName_2)
ace <-AceLavaanGroup(dsClean)
ace

Thanks. I will wait for the 3rd reviewer before proceeding with publication.

Dr. Meisenberg kindly reviewed the paper. His permission given, for documentation, I attached his marked up version of the paper along with our email exchange.

He noted:

Great work, interesting results. You mention that the bioecological model proposes that heritability is reduced when scores are environmentally reduced. I don’t think that such a simple connection is self-evident, and you could mention this in the discussion.
...
Better specify the age range. Black-White IQ differences tend to increase with age. There is also some uncertainty about changes over time. The black-white gap in NAEP was reduced by about 40% for birth cohorts between about 1960 and 1975, but it is not certain how IQ gaps changed during that time.

I asked:

JF:

Thanks for your thoughtful commentary. Do I have permission to post your notes on the 'Open Behavioral Genetic' thread for this paper? Can I also note that you give conditional approval -- that is, on condition that my colleague and I make the noted clarifications? As for these, I agree that they will improve the paper somewhat

GM cordially agreed.

To fulfill the conditions, I made the following modifications:

(1) I added a footnote to our discussion of the group differences, which reads:

"[1] These represent crude characterizations of the magnitudes of the typically found differences. A more precise description would take into account variability by age, birth cohort, and migrant generation; to note, Dickens and Flynn (2006) found evidence of an age and birth cohort interaction with regards to the B/W differential, while Fuerst (2014) found evidence of a migrant generation interaction with regards to the H/W difference. "

(2) I added the following paragraph in the discussion section:

"Alternatively, the results may imply that the key prediction of the bio-ecological model – that environmental depression in a trait will be associated with significantly reduced heritability -- is incorrect. Another alternative is that the between group differences in question are more or less uniformly distributed – since the bio-ecological logic seems to be that environmental depression between groups will lead to increased environmental variance in the depressed group, which will lead to lower genetic influence in that group; this, of course, only holds if the between group difference is not uniformly distributed. Either way, the matter is difficult to determine, since a clear quantitative prediction has not been offered regarding the effect of environmental depression, measured in terms of standardized differences in cognitive ability, on variance components"

I attached the newest version the the paper. Can the journal's administrator confirm that we have filled the condition for GM's approval and that we now have three approvals?

Malloy's analysis seems to show that the BW gap is fully present at age 3. http://humanvarieties.org/2013/05/26/the-onset-and-development-of-b-w-ability-differences-early-infancy-to-age-3-part-1/

And my own analysis seemed to show little secular change when setting the starting point at 1920. Thus, I said "Dickens and Flynn (2006) found evidence of" not "Dickens and Flynn (2006) showed/proved/established". I worded the statement loosely.

You probably need to rewrite this passage, like this :

Dickens and Flynn (2006) found evidence of an age and birth cohort interaction with regards to the Black/White differential, while Fuerst (2014) found evidence of a migrant generation interaction with regards to the Hispanic/White difference.

Maybe it's obvious to you, but others may not guess what is H/W. It's not a problem if you have specified in the abstract or introduction that you use H/W for hispanic/white, but as I can see, you didn't mention it. Throughout your text you always write "hispanic/white".

Thanks. I will wait for the 3rd reviewer before proceeding with publication.

I attached the newest version the the paper. Can the journal's administrator confirm that we have filled the condition for GM's approval and that we now have three approvals?

Yes I confirm. If the authors give their final approval, I'll publish the paper.

Can the author please send me via email (or attach it here) a .docx file of the final manuscript, with keywords? Your updated PDF doesn't have a publication date (it's written under review instead) and I need a .docx to add it. It also doesn't have keywords.

Can the author please send me via email (or attach it here) a .docx file of the final manuscript, with keywords? Your updated PDF doesn't have a publication date (it's written under review instead) and I need a .docx to add it. It also doesn't have keywords.

I attached all of the relevant files.