🚨 Hurry, space in our FREE summer bootcamps is running out. 🚨Claim your spot here.

# Elementary schools in rural India are usually small, with a single teacher. The teachers often fail to show up for work. Here is an idea for improving attendance: give the teacher a digital camera with a tamper proof time and date stamp and ask a student to take a photo of the teacher and class at the beginning and end of the day. Offer the teacher better pay for good attendance, verified by the photos. Will this work? Researchers obtained permission to use 120 rural schools in Rajasthan for an experiment to find out.$^{37}$(a) Explain why it would not be a good idea to offer better pay for good attendance to the teachers in all 120 schools and then to compare this year’s attendance with last year’s.(b) Write a few sentences describing how you would design this experiment.

## a) To design a good experiment, we need to apply a treatment to part of the group and NOT apply a treatment to a similar group (a control group) so we can compare results. The proposed method would not allow for comparison.b) 60 randomly selected schools will receive the offer, while the other schools won't receive the offer and compare the results.

#### Topics

Sampling and Data

### Discussion

You must be signed in to discuss.

### Video Transcript

okay for this problem. We're looking at the context of doing some kind of experiment. Thio identify a solution to this problem we have in rural India of teachers not coming, not showing up. We're not having good attendance for work, so I'm gonna type out the problem, and then we have a proposed solution, and then we're gonna talk about how how ballot it is. Okay, So the problem is Ah, yeah, the teachers aren't coming to work consistently. And they in the problem statement they have this proposed solution. Yeah, that they're going thio plan to pay teachers more and then have the students take pictures as video evidence to kind of for accountability, we can time stamped. Well, doing this on all 120 teachers work. Okay, More specifically. So they say, do it on 120 teachers one year and then the next year and just compare it to the previous year. So that's their initial plan that we're going to analyze. Okay, so I'll start answering this question. So the solution is so for good experimental design. So they quit. The quick answer is no, it's not a good plan. to test all 120 teachers. Eso what you really want to do with good experimental design has a control group? Yeah, What you need to do is not do it to all 120 then wait to see what happens the next year. What you need to do is is, um do it like half the group in this case or a part of the group. Do the experiment, which is called the treatment to that. And then do it thio other group don't do the treatment, which is the control group, and then you can compare the results picture. Yeah. Okay, so that is the answer for part A for party. I'm gonna kind of right. It kind of brought this write a few sentences, but these experimental design are often best shown with, like, kind of a diagram. So let's show this diagram here so we can say for party. We have 120 students, 100 teachers. Sorry. And we're going to see, uh, what do you do for this experiment is we're gonna split the group and actually specifically, you want to do it randomly because write it out in detail. Gonna randomly pick 60 of them. Okay, so we're gonna take 60 of the teachers, You're gonna apply the treatment. In this case, the treatment is the better pay and the video evidence. And then for the other group, you're gonna take 60 teachers and not do that Just kind of do what you normally doing. Then after a certain time period, which you can dio as you can compare the results. So good experiment has that if you want to verify this is working or not, compare the results. And so we'll say, Well, after a certain amount of period, don't have to be in a year. You and it could be like six months or something. So then you say, Well, is the attendance better for what I draw? I've drawn out of the Blue group or the Red Group, but in general, we're just gonna compare the results. So that's the diagram way of showing the answer. But you could put that into a sentence. If you want to type your answer or teacher, probably love it. If you just drew that picture, if you're doing it, um by hand, or have a ways to submit it because this kind of explains what happens. But generally we take 120. We do a random sample. 60. Do this new proposed solution. 60 don't. And that way we can compare the results. All right, so I hope that helps you out with some good ideas about good experimental design.

City College of New York

#### Topics

Sampling and Data