Better than a dog biscuit!
Subscribe to:
Post Comments (Atom)
Read it -- it's good for you!
About Me
- John Reynolds
- I'm an associate professor of sociology at Florida State, where I teach Social Inequalities, Social Problems, and Social Methods and Statistical Analysis.
8 comments:
Since I'm the first to comment, I get leniency if I say something stupid or don't do this correctly -- I've never blogged so I don't know the etiquette...
RE: Editor letter, Meta Review, Corey’s comments
1) I think we should shorten the front section as suggested by reviewer B and Corey Fields. However, since I don't know psychiatry's history with mastery, I'm not sure of the need to add this (suggested by editor). Could we instead just delete the reference to psychiatry?
2) I agree that we need to expound upon the limitations and null finding of unemployment. Remember how surprised we all were when we initially saw these results? I'm sure that's how these folks are reacting and we could talk about supplemental analyses we did with NLSY that found similar results? Maybe as Reviewer B suggested, we could talk about (if we know) movement of respondents between counties and theoretical reasons we predicted unemployment rates would affect attainment later.
3) Related to this, I think we could add depth to the lifecourse argument by better situating the analyses historically (per the last bullet in the meta review comments).
4) I also like the metareview suggestion to identify race and/or gender differences as important directions for future research.
5) Regarding the metareview comment about Table 2 not controlling on SES--how difficult would it be to run a logistic regression (outcome = professional/mgr/teacher) and control for SES and other variables we included? If mastery is significant, wouldn't that back up the statement about ambition?
RE: Reviewer B Comments Not Addressed Above
1) I agree with comment “On page 5, top, it is too soon to tell us what will be controlled; that is for a later section conveying the Plan of analysis.”
2) I agree with the comment “Essentially, this paper is about young people meeting their expectations for occupational status, not for occupations. This is not a trivial distinction and is one that I think deserves more play throughout the rest of the paper.”
3) Add footnotes acknowledging diversity of black and Hispanic populations and handling of missing data.
4) As mentioned in section above, we could talk about supplemental analyses with NLSY and mention the limitation of using it (reinforce earlier statement in paper about NLS being the best choice).
RE: Reviewer C Comments Not Addressed Above
1) Address the second suggestion (“how mastery is linked to homemaker expectations and pursuits….Young adults who choose to be homemakers and then pursue this role in adulthood would also be displaying planful competence.”) in discussion.
2) We should do all three minor comments.
Sounds like a lot of work and I’m not sure how much time I can commit in the next few months but I will stay involved and volunteer to do whatever pieces I know that I can do.
FYI--I'm still working and I’m defending next month and must have committee approval (i.e., make all changes) by beginning of April (job condition), trying to get two houses ready to sell, and we’re coordinating a wedding at our house for my stepdaughter.
Cheryl
I agree with Cheryl’s comments. A few other points I would like to add:
-I think we got a few requests to explain the unemployment measure (why we chose the measure we did, strengths/weaknesses of measure), and I agree that we should give attention to that. Theorize/discuss how moving out of the local area in adulthood might relate unemployment (does it show mastery?).
-Reviewer C (point 1) asks if we can find a way to account for changing occupational expectations over time—that those who adjust their expectations might be displaying mastery. I think this is something for us to look into.
-Regarding the literature review, I took the following from the reviews: We need to tighten and shorten the review, add more of a lifecourse perspective to the literature, include lit on the structure vs. agency debate and the economic/historical context of the time period. Like Cheryl, I wonder if a discussion of mastery and psychiatry is necessary.
I’m working towards defending in April/May, and looking for a job…so of course I am busy, but I can devote some time to edits.
Analysis:
1. Labor market variable needs further justification, as does NLS data set.
2. Figure out who stays in labor market of origin. (I'm not sure we can do this.)
3. What does the distribution of mastery look like for group that met their occupational expectations in 1973?
4. Does mastery predict higher occupational expectations, net of other factors?
5. How do results change if expectations are allowed to change over time? (How would we do this? Use yearly dummies to indicate if their expectations increased or decreased relative to staying the same? Alter the DV?)
6. How does mastery work for those expecting to be homemakers? (I'm not sure how important this is, especially if we're linking to stratification lit).
7. How are respondents coded if they achieve better than their early expectations? Is this justifiable?
8. Does effect of mastery vary by race? (Can’t we just use an interaction rather than having separate models?)
9. Does effect of mastery vary by gender? (Didn’t we do this already and note it?)
10. Specify the effect of time in the model (dummies or some other functional form)
11. How does analysis account for the ‘structural constraints’ in which respondents find themselves (e.g. marriage, kids, etc.). I’m confused by this. Don’t we take care of this by having annual measures of kids, but measures of mastery in only 73, 74, 76, 79, & 86? I mean, at many points in the analysis isn’t mastery actually preceding childbearing? I’m probably just misunderstanding this.
Intro/Lit Review:
1. Tighten and make the contribution clearer.
2. Get rid of debate between two sides of expectations (motivation vs. indicator of structural constraints).
3. Add psychiatry literature (if fruitful)?
4. Link to broader occupational stratification literature.
5. Change language to reflect that DV is meeting expectations for occupational status.
6. Add historical context for early 1970s.
Timeline:
I've got the psychiatry works that Fine recommends. I'm looking at them to see how useful they are. My schedule is (fairly) open for working on this. So, just let me know which parts of this you want me to take on and I'm happy to do it.
-steph
p.s. Congrats Emily and Cheryl on defending! That's so exciting.
Okay, let's proceed and Brandy can chime in when she's got the chance. I want to split up the tasks by type. There are probably three major types: (a) consider this, (b) add/remove that, and (c) look at this.
What, that's not obvious? I mean that several comments ask us to speculate on the implications of various issues. Other comments ask us to add material (psychiatrists' work on mastery) or delete stuff (shortening the literature review). Another chunk of comments can only be addressed by looking at the data.
I'll take on the data analysis tasks first. You three (Cheryl, Emily, and Steph) can take on pieces of the other tasks. For example, Steph has started looking at the psychiatric literature.
Steph, would you also like to add a paragraph on the "true life course" stuff--how does the broad economic context of the 1970s affect the results? You might want to look at the Statistical Abstract of the US to get numbers on unemp, job growth and decline, etc.
Emily, can you take a crack at the comment "I still think the discussion could do a better job of linking the findings about mastery to the findings about the effects of race, class and gender?" Also think how we could better nuance the front end to be more clearly about the tension between structure and agency from a general inequality standpoint (not just a life course thing).
Cheryl, can you try to address the concerns of Reviewer B as stated in the first paragraph of B's comments?
That should get us started.
John (and others),
Do we have a working time table for these tasks? I have a job interview(!) Feb 15 so I'm working pretty hard on my job talk/teaching demonstration...would it be throwing a major wrench in things to work on this when I return?
Em
...would it be throwing a major wrench in things to work on this when I return?
Congratulations on the job interview! That definitely should be your priority right now, and I'd rather wait until you can focus on addressing this important comment. Good luck! -John
Editors’ and Reviewers’ Comments
Editor
1. Discuss how “the historical moment in which this research took place … had some considerable effect on beliefs about mastery and occupational expectations.” Relate this to the findings of Shanahan, Elder, and Miech 1997.
2. Link the concept of mastery to its earlier treatment in psychiatry and social psychology.
3. Shorten the literature by focusing on the references that we truly build upon. “Look forward, not backward.”
4. Replace Appendix A with a reference to an NLS study to provides the same information.
Deputy Editor
1. Either incorporate analyses of group differences related to race/ethnicity and gender, or point to these as important directions for future research.
2. Further discuss limitations of the NLS measure of unemployment.
3. Explain why the level of unemployment in the county in which you live during high school would affect later occupational attainment.
4. Clarify the argument vis-à-vis Reviewer B’s comments; integrate material on pp. 2-4 into the argument on pp. 8-9.
5. Address Reviewer C’s concern that many teenagers’ occupational expectations likely changed between 1973 and 1974. Can this be incorporated in the DV? If not, what are the implications for estimates of the effects of mastery on achieving occup. expectations?
6. Discuss the levels of mastery for those who already met the occupational achievements in 1973, better highlighting the information provided in footnote 2. Perhaps this group gives further evidence of the finding that “mastery does not predict the likelihood of achieving one’s occupational expectations if those expectations are low.”
7. Since the analyses in Table 2 don’t control for SES, and part of the association between mastery and expectations is due to their common associations with past grades, family SES, etc., how certain can we be that mastery increases ambition, as stated on p. 18?
8. How do the analyses take into account the way that mastery selects people into or out of “contexts of constraint and opportunity”? If these endogenous effects are not accounted for, what does this imply for interpreting the regression results?
9. Discuss the significance of the historical context in which these youth are pursuing their occupational expectations. Would the macroeconomic context predict a lesser or greater impact of mastery, given past work (Shanahan, Elder, and Miech 1997)?
10. How was the dependent variable coded if respondents exceeded their expectations?
Reviewer A
1. Is there some way to address the weakness of the measure of unemployment? Perhaps analyze those who did not move since 1970, or look to alternative data sets to better assess whether the null findings related to the unemployment rate are valid.
2. Determine if the effects of mastery differ for blacks and whites and report these results.
Reviewer B
1. Rework the material on pages 2-4 and begin the paper with the life course perspective.
2. Move the text on page 5 describing control measures to later in the paper.
3. The issue of what meeting one’s occupational expectation really means need to be discussed in more detail on page 13 and should receive more attention throughout the paper. “Essentially, this paper is about young people meeting their expectations for occupational status, not for occupations.”
4. Better address the limitations of the measure of the unemployment rate: note many people leave the county in which they attended high school, explain why those rates should affect later occupational attainment.
5. Note the limitation of treating all Hispanics as a single group.
6. Describe how missing data were handled. Consider using regression imputation or other current technique to estimate missing data.
7. On page 27, the paper notes two other data sets that may be tapped to better estimate the effects of local labor markets, which is not necessarily consistent with the claim that the NLS is the best data set for the study. Discuss this issue earlier, clarify why NLS is best.
Reviewer C
1. Concerned with the analysis strategy that does not allow expectations to change over time. “The authors need to address how changes in occupational expectations over time affect their pattern of findings.” Discuss the possible effects of mastery on change in expectations. Changing expectations means “they are no longer at risk for achieving their early occupational goals.”
2. Discuss how mastery is linked to homemaker expectations and pursuits.
3. Be clearer in the discussion that the null findings for the unemployment rate may be due to the fact that the measure is time invariant. “Young people with high self-efficacy may move to labor markets with lower rates of unemployment to make the most of their abilities and expectations.”
4. Provide more discussion of the timing of event occurrence in the event history analysis. “Why does the odds of achieving early occupational goals increase and then decrease during the transition to adulthood?” Also, confirm that a quadratic function is the best time specification.
5. Assess whether the effects of mastery vary by gender. If so, it might be useful to estimate the event history models for men and women separately. The effect of mastery may also vary by race/ethnicity or family SES (though may detract from paper’s focus).
6. Prefers the term “statistically non-significant” over “insignificant.”
7. Tables 3 to 5 should distinguish time invariant from time varying predictors.
8. Helpful to see the odds ratios in Tables 3 to 5.
Comments from Corey Fields
1. Be more specific about the aims of the paper—statement about the agency/structure debate, an expansion of the concept of mastery, or exploration of the antecedents of occupational stratification.
2. Tighten the literature review, and the front of the paper “could do a better job of situating this study in a larger research tradition.”
3. Expound on why county level unemployment is a reasonable proxy for economic structural conditions.
4. The effects of race, gender, and SES on achievement of occupational goals gets lost behind the focus on mastery. Discuss the effects of both “structure” and “agency” in this regard. The discussion should foreground the finding that “the relationship between agency and social constraint is nuanced and contextual, not direct and unvarying.”
The NLS has measures of geographic mobility. In the first through 4th followups they ask whether they've moved and how far. We could explore whether high mastery folks are likely to move out of high unemployment areas, and whether high unemployment rates hurt ambition-achievement for those who don't move. What do you think?
Post a Comment