If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

AP®︎/College Statistics

Course: AP®︎/College Statistics>Unit 6

Lesson 4: Introduction to experimental design

Introduction to experiment design

Explanatory and response variables. Random sampling and random assignment (including block design). Placebos and blind/double-blind experiments.

Want to join the conversation?

• At , Sal mentions that you can use "block design" to randomly divide patients with common properties, such as sex, evenly throughout the control and experiment groups. Isn't that the same as using stratified sampling in a sample study?
• I think it is the same idea, it just has a different name. Experimental surveys and observational studies are pretty different in form, so I think it would be confusing to use the same term for both.
Hope this helps!
• how to control error using blocking
• You use blocking to minimize the potential variables(also known as extraneous variables) from influencing your experimental result.

Let's use the experiment example that Mr.Khan used in the video. To verify the effect of the pill, we need to make sure that the person's gender, health, or other personal traits don't affect the result. We want to test if the pill would be effective for everyone in general. This is where blocking method comes in; we need to divide the samples into two groups in a way that each group is similar to each other in terms of aforementioned factors(gender, etc.). From this, you can ensure that the result will be affected minimum by the unexpected factors, hence controlling the error.
• Is there a reason the KA video about correlation of A1C with blood sugar can't be found on this site anymore?
(It is still present on YouTube though)
• At I start getting confused
• How could you confirm your experiment is replicable?, do you have to do another experiment with different populations or just mention on the potential paper that it is necessary to do more experiments considering other samples?
• it is usual that other researchers who have no direct interests in and are independent from one study would do the similar experiments, following the procedures and parameters published in the original paper.

then they check if their results would be compatible with the proposed ones.

and as you said, it's almost impossible to do exactly the same experiments with the same subjects again, espeically when it comes to the clinical cases like this. thus the purpose of replicating an experiment is not in duplicating the experiment and getting the exact same data itself, but in confirming general trends with similar conditions not by the authors of the paper but by the others (usually peer researchers).