Emily
Nice example. For reliability assessment of the timing measure, it would have been ideal to report a test-retest assessment. Cronbach's alpha measure of internal consistency reliability is just one way to assess reliability. Test-retest is very useful and intuitive, and often inter-rater reliability is also worth calculating. In many cases, the ideal would be to calculate all three measures of reliability.
The part 2 paper you describe is very interesting. They do make a lot of disparities comparisons but almost never directly quantify - they just report the numbers in each group and say which one is larger or smaller. This is somewhat appealing because it leaves the reader able to calculate whatever they like (either relative measures like ratios or absolute differences), but often to communicate it is nice to be able to summarize the magnitude of the disparities.
They do say: "Low-income women had much higher rates of unintended pregnancy than did wealthier women; this disparity increased between 1994 and 2001, manifesting as growing disparities in the rates of both abortion and unintended birth.." which implies that they have a particular metric for the magnitude of the disparities in mind. Because the trends for low and high income women were actually in opposite directions though (increasing unintended pregnancy for low income women and decreasing unintended pregnancy for high income women) you could draw the same conclusion using either absolute or relative measures.
Maria
Low-income women had much higher rates of unintended pregnancy than did wealthier women; this disparity increased between 1994 and 2001, manifesting as growing disparities in the rates of both abortion and unintended birth