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.

## Integrated math 3

### Course: Integrated math 3>Unit 10

Lesson 2: Observational studies and experiments

# Types of statistical studies

Created by Sal Khan.

## Want to join the conversation?

• Please correct me if I am wrong, I would love to know what other people think. This video isn't about types of statistical studies as it is the process from hypothesis to experiment for a single idea. Each of the "types" seem to be more "stages" in that progression as opposed to different ways to conduct statistical research.

Kind Regards!
• In AP Statistics we teach about Observational Studies (includes Sample Studies), which must have random Selection, and Experiments, which must have random assignment of treatments. Often we have to rely on volunteers for experiments, which may affect our scope of inference, but otherwise does not invalidate our results. We can see a correlation from an observational study and we can establish cause and effect from an experiment.
• if for example some one asks me to define Statistics!! than how should i define??
• This is not necessarily the most intelligent answer, but according to Sal in the "basic probability" section, he states that Statistics is the analysis of events governed by probability. Wikipedia states that statistics is the study of the collection, organization, analysis, interpretation and presentation of data! Hope this helps!
• Whats the difference between Sample study and Observational study as both of them uses sample from the original (or large) pool of data ? moreover both techniques work on these sampled data...
• I had the same question after the video. From my brief gatherings it seems the first type of study mentioned at () is a Survey type study with data gathered via polling questionnaires and relying on honesty. The observational study ( ) could be measuring taking actual data samples (blood samples for sugar intake). You could sample millions with a survey but not so much with an observational study.
• What does the word STATISTICAL mean ? (I looked in both of my dictionaries and could not find an answer.
• I'm sorry that plan didn't work out, because it was a very good idea! My dictionary is a little better than yours apparently, and it says "of, relating to, based on, or employing the principles of statistics." Zzzzzz.

In real-person words, statistics is the branch of mathematics that deals with collecting and analyzing data to solve a problem, and something like a study is statistical if it is using statistics in the right way to answer the question it was designed to ask.
• I don't get how this describes Types of Statistical Studies...
(1 vote)
• At => Sample Study
At => Observational Study
At => Experiment
• What about instrumental variables, can't they be used to establish causality when experiments aren't possible?
• In theory, although coming up with an appropriate instrument that will withstand criticism might impact the perceived validity of the results. At any rate, instrumental variables are an advanced topic that is beyond what Sal's going to teach in the statistics playlist.
• I'm having trouble discerning between the different types of statistical studies. What are some bullet points for each study type?
• An experiment must have random assignment of treatment(s). Often an experiment will incorporate control groups and/or blocking. We don't necessarily have random selection of participants in an experiment because we often have to rely on volunteers. This does not affect our ability to determine cause and effect, but could affect the inference of those results if the volunteers are not representative of the bigger population.
Observational studies require random selection and generally can only show correlation (not cause and effect!) between variables. We have to be careful of biases and confounding variables. Sample studies, surveys are a type of observational study.
• In regards to a sample study, shouldn't the data be somewhat controlled so most variables are eliminated? For example, in the sugar leads to heart disease study shouldn't you sample people of a certain age, height, weight, ethnicity, etc? Then, if you see any fluctuations or discrepancies within the data you can point to sugar being a cause because all the other variables were controlled whereas if you sampled randomly you don't know if the cause is sugar or age.
• Sal was talking about studying the ENTIRE population of the US by gathering data from a very broad, varied selection of people. However, in the Observational Study part of the video, he did restrict the people to 60 year olds. I think controlling the data like you described would be the next step if you were doing a serious study on sugar influencing heart disease and wanted to go further.

Hope this helped and didn't make you too much more confused!