1 / 18

11. Controlling for a 3 rd Variable

11. Controlling for a 3 rd Variable. +. a. b. c. +. +. a. b. 0. Explicating a bivariate relationship with a third variable. Identifying a misspecified relationship : A spurious relationship is one type of misspecification Observe But.

malikh
Télécharger la présentation

11. Controlling for a 3 rd Variable

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. 11. Controlling for a 3rd Variable

  2. + a b c + + a b 0 Explicating a bivariate relationship with a third variable Identifying a misspecified relationship: A spurious relationship is one type of misspecification Observe But

  3. Explicating a bivariate relationship with a third variable (continued) More generally a misspecified relationship is when the magnitude or direction of the relationship you observe between a and b is not due to a causing b, but to c partly or wholly causing both a and b. When you control for c the relationship between a and b changes in magnitude or direction.

  4. # of Fire trucks sent to fire Severity of Damage Initial Report Severity of damage + + # Fire trucks Severity of damage - • E.g. Should we keep the fire trucks home?

  5. 1 truck > 1 truck 1 truck > 1 truck 70% 80% 30% 50% Not severe damage Not severe damage 30% 20% 70% 50% Severe damage Severe damage n= 100 n= 100 n= 10 n= 10 Taub < 0 Taub < 0 1 truck > 1 truck Not severe damage Taub > 0 Severe damage n= 110 n= 110 Not Serious Serious Initial Report

  6. 30% 30% 70% 70% a 60% 40% Taub > 0 Spurious b 40% 60% 100% 100% c (3 values) low high medium a a a 70% 70% 50% 50% b b b 30% 30% 50% 50% Taub = 0 Taub = 0 Taub = 0

  7. ++ a b c + + a b + Or you might find: Taub = 0.6 0 < Taub < 0.6 for each value of c Here we would have overestimated the impact of a on b. A does cause b, but controlling for c we realize the effect is less than we initially thought.

  8. Controlling for a third variable thus allows us to test alternative explanations for a hypothesis. • When you cannot do a proper true experimental design that eliminates alternative explanations, you need to do statistical controls. Here we have just looked at how you do a statistical control.

  9. a b 83% 70% 17% 30% Conditional Relationships: Specification is another reason to control for a third variable c Low Ed. High Ed. No Worked for Political Candidate Yes

  10. 75% 90% 70% 70% 25% 10% 30% 30% Men Women Low Ed. High Ed. Low Ed. High Ed. No No Worked Worked Yes Yes Small + Taub Large + Taub Relationship between education and working for a candidate is positive for both men and women, but is stronger for women than men.

  11. + a b c 40% 30% 0 60% 70% a b + Multiple Causes (Enhancement): Two variables may be causes of a third variable, while the two are unrelated to each other. a b n = 200 n = 200 c

  12. 50% 30% 40% 20% 50% 70% 60% 80% a a b b n = 100 n = 100 n = 100 n = 100 Our estimate of the impact of a on b is unchanged, but by also looking at c we can better predict b. Both a and c are causes of b.

  13. a a b c Race Race Income Education Using a third variable to find an intervening relationship: b A causes b. All or some of the way a causes b is through c. Income First, we observe minorities earn lower incomes than non-minorities. Then we ask, to what extent is that because they achieve lower levels of education and lower levels of education result in less income?

  14. 63% 70% 50% 47% 60% 40% 37% 30% 50% 60% 53% 40% Race Min. Non-Min. Low Taub ++ Income High n = 150 n = 300 Control for Education Low Education High Education Min. Non-Min. Min. Non-Min. Low Low Income High High n = 100 n = 100 n = 50 n = 200 Taub + Taub +

  15. Some, but not all, of the impact of race on income is due to education. Education partly explains the way in which race affects income. • Remember race is still the cause, we are looking at the mechanism. • If Taub = 0 with control, then all the effect of race would have worked through education.

  16. a a c b a b + Using a third variable to find an antecedent cause: b + Acauses b, but we can learn more by finding a is caused by c. Here we start with: a b We ascertain: c a c With… a Then we identify a as intervening by predicting b with c and controlling for a. To the extent the relationship is attenuated by the control, c is antecedent and works through a.

  17. a a b b c c a a b b Theory is key in drawing the causal arrows. , then the simple If will be misspecified. But if then c is intervening. is correct in estimating the magnitude of the effect of a on b.C become a mechanism of how a causes b. The researcher must draw the arrow correctly. Statistics can’t solve this problem.

  18. a b Hint: Typically (though not always) Demographic Attitude Behavior Avoid reciprocal relationships: a b But if you think: a b You can mention that b may have a small impact on a, but the overwhelming effect is of a causing b. You can then just consider:

More Related