Class 9- Do Teenagers Respond to HIV Risk information? (Dupas,2011)

Question 1

Who wrote “Do teenagers respond to HIV Risk Information?”

Answer 1

Dupas in 2011

Question 2

What was the main aim of the paper?

Answer 2

To find out the response of teenagers unprotected sex levels when given (HIV) risk reduction information compared to risk avoidance.

Question 3

What are risk reduction measures

Answer 3

To teach teenagers how to reduce their risk of HIV by giving them information about protection and that on average older partners are more likely to have HIV

Question 4

What are risk avoidance measures

Answer 4

To advise teenagers to abstain from sex until marriage in order to avoid HIV

Question 5

What are intensive margins?

Answer 5

Advising on what type of sex to have and with whom (age). Risk reduction programs use intensive margins

Question 6

What are extensive margins?

Answer 6

Advise on whether to have sex or not. Risk avoidance programs use these margins.

Question 7

Why might giving teenage girls information about the level of HIV by age reduce their risk of contracting HIV?

Answer 7

Risk avoidance programs just tell girls the overall HIv rate rather than dissagregating by age. If girls are made aware that older men are more likely to have HIV, they may substitute away from them to boys the same age as them and thus reduce their risk of contracting HIV.

Question 8

What are the keypoints of the data set ?

Answer 8

328 schools took part.
This was a randomized control trial. 71 schools were randomly chosen to have the “Relative Risk Information Campaign” at their schools and the others had the “Teacher Training Program” where information about differing HIV rates by age were not given.

Question 9

What does previous literature say about the impact on teenagers in Africa’s sexual behaviour when sexual education is given ?

Answer 9

Gallant in 2004 looked at the effect of HIV education programs in Africa. The results of this review is likely biased as peoples riskiness habits (i.e condom use) was self-reported and people are likely to underestimate their risk levels.

Question 10

What are the predictions (hypothesese) of the study from the Teacher Training side?

Answer 10

That for girls who over estimated the risk of HIV before the program their demand for unprotected sex will rise following.

For girls who underestimated the risk of HIV prior they may demand a higher transfer from their sugar daddy to have unprotected sex due to the risk. If the man can meet this demand, the level of unprotected sex will stay the same but if they can’t it’ll fall.

Girls already aware of the level of risk of HIV won’t change their behaviour

Question 11

What were the predictions (hypothesese) of the study from the Relative Risk Information side?

Answer 11

If all men no matter the age are willing to give the same transfer to these girls no matter there age, then girls will change their behaviour to have less risky sex (so boys the same age as them).

If girls that were previously abstaining from sex upon hearing the Relative Risk info realise that teenagers are less risky than the population as a whole so may begin having sex.

If teenage boys however have a lower reservation transfer the effect can be ambiguous:
The girls may remain with the older man but be able to get a larger transfer as she is now more aware of the higher risk she is taking by being with an older man/

Or, if the girl is in desperate need for the high transfer that is gained with older men but wants to reduce her risk, she may sleep with several boys her age so she can remain on the same level of income as she had with the older man.
This would lead to a rise in unprotected sex.

Question 12

When did each of the programs take place?

Answer 12

The Teacher Training program on average rates of HIV and abstinence was rolled out in 2003 to students of all grades
The Relative Risk Information campaign was implemented in 71 schools in 2004 for only students in the 8th grade.

Question 13

How was the degree of unprotected sex of teenage girls measured?

Answer 13

The change in pregnancy levels was used as an objective measure. Pregnant girls would be interviewed to ask the age of the babies father. This was done to compare the level of unprotected sex with boys their age compared to older men.

Question 14

Are pregnancies a perfect measure of the degree of unprotected sex?

Answer 14

No as anal sex could occur and this is highly risky for HIV transmission.

Question 15

What was the simple regression used to estimate the impact of the interventions?

Answer 15

Yis1=a1+B1RRs+X1TTs+I’U+Eis

Yis1 is a dummy variable that tells us if the girl I, from school s was pregnant a year after the intervention.

B1 is the average effect of coming from a RR treatment school compared to a RR comparison school. (Where RR didn’t take place )Given that dummy RR=1
We assume there is no OVB
I is a vector of controls for background characteristics.

Question 16

Why did Dupas also use a difference-in-difference estimation?

Answer 16

It was used in case that the randomization was not perfect and there was OVB.
The dif-in-dif also controlled for school fixed effects, where there may be differences across schools in their attitudes to unprotected sex outside of the intervention.

Question 17

What were the main results of the Relative Risk info being provided ?

Answer 17

This reduced the amount of girls that were childbearing a year after the intervention.

Also those from the RR treatment group such girls that were pregnant on average had a baby father that was 1.7 years younger than those that did not receive the RR intervention.

Also the likelihood of having a babyfather that was more than 5 years older was 46% smaller.

Question 18

What were the main results of the Teacher Training (Abstinence) info being provided ?

Answer 18

There was no effect of statistical significance on the number of girls that began child bearing a year after the intervention.