Week 3: Discussion questions
Lewis and Oppenheimer (2000)
1. What are the main differences among the three perspectives (sex ratio, structural theory - concentration and marriage search theory)? The first two could be thought as measures of flow and stock, respectively. What about the third one?
2. Which age range do you think that it is more adequate to match potential partners (Lichter et al (1992) used 10-year age groups and Lewis and Oppenheimer, 5-year)?
3. The authors explain on the footnote 6 that 11% of all person-years had imputed values. What are the advantages and the disadvantages in using procedures to impute values for missing data?
4. The authors found that "women's chance of marrying down rather than homogamously or up increases more with age if they life in educationally sparse marriage markets". However, they did not find this evidence for men. Why the marriage market is more "cruel" to women's aging? Do you believe that this is an indication that men value attractiveness more than women? Do you have another explanation for this finding?
5. The authors include education in their analysis but excludes race. Do you believe that the results could be different for Blacks? In which direction?
6. What is your opinion regarding the use of group-standardized measures?
7. What are the differences between logistic regression models and discrete-time hazard models with cause-specific risk?
Lichter, LeClere, and McLaughlin (1991)
1. What are the assumptions that Lichter, LeClere, & McLaughlin use to establish causal ordering?
2. What are some possible mechanisms (besides "traditional family values") that could be driving the urbanization hypothesis? Should we be differentiating between urban and rural marriage markets?
3. This paper doesn't show much support for the "marriage squeeze" phenomenon among blacks. What's going on here?
Lichter et al (1992)
1. The authors utilize four measures of sex ratio, one is the sex ratio per se and the other three are measures of the availability of economically attractive men (pool of currently employed men, poof of full-time employed men, pool of men with adequate earnings). They argue in the footnote 8 that the sex ratio measure per se is not redundant with the other three given the fact that the correlations between the first and the other three were respectively: .896, .713 and .709. What is the threshold to accept that a new measure is not redundant?
2. Do you understand why the sex ratio per se (Table 3) is so different for U.S. white and black females? Even though the mortality is higher among Black men than among White ones, they are not so different to generate this substantial difference. What is your guess?
3. The idea of local marriage markets permeates the discussion and its concept defined on the basis of residential propinquity. The population analyzed, women aged 18 to 28, is probably the most mobile age group. In this context, do you agree with the use of the concept of local marriage markets? If so, do you understand how authors handled this issue of mobility on their analysis?
4. The sex ratios are calculated based on the pool of never married men and women. However, in the analysis of the transition to the first marriage for women, it is not clear if their mates were also doing the transition to the first marriage. Therefore, it is possible that sex ratios are controlling for one thing and the model is measuring another one (if observed marriages are not restricted to first marriages for both partners). What is your feeling about this?
5. How good is the information about observed marriages as an indicator of preferences in the marriage market? In reality, what we really know about preferences is in fact a result of the preferences constrained by the marriage markets. What do we really know about preferences?
6. My impression about part of the literature is that many times models are built in order to "control for" all possible dimensions. Do you have the same feeling? Do you believe that there is a need for a more formal/theoretical modeling?
7. Do you think that it is adequate to describe the results in terms of delaying marriage when the data used has only 6 years of follow up? If yes, how about non-marriage?
8. Given the previous readings, do you agree with the focus on the traditional marriages instead of unions in general?
Raley (1996)
1. When it comes to aggregate-level data, we are usually pretty limited as to the data available (especially when combining aggregate with micro-level data). How can we best define and quantify "marriage markets." What would be the ideal way to identify a marriage market geographically and/or statistically (counties, LMAs, something else)? (Apply to Lichter, LeClere, McLaughlin, 1991 also)
2. Raley emphasizes that cohabiting unions cannot be ignored in the examination of Black marriage markets. In light of this, to what extent can we apply Oppenheimer's job search model to Black marriage markets?
3. Is employment the best way to measure the quality of mates in a marriage market? How can we be sure to capture mate quality and "availability" in aggregate data?