• Shuffle
Toggle On
Toggle Off
• Alphabetize
Toggle On
Toggle Off
• Front First
Toggle On
Toggle Off
• Both Sides
Toggle On
Toggle Off
Toggle On
Toggle Off
Front

### How to study your flashcards.

Right/Left arrow keys: Navigate between flashcards.right arrow keyleft arrow key

Up/Down arrow keys: Flip the card between the front and back.down keyup key

H key: Show hint (3rd side).h key

A key: Read text to speech.a key

Play button

Play button

Progress

1/12

Click to flip

### 12 Cards in this Set

• Front
• Back
 What is observational experiment? Choose 2 groups of objects with uncontrolled difference. e.g. one group is smoker, and another group is non-smoker, since smoking or not controlled by the object => it is uncontrollable What is experimental experiment? Choose 2 groups of objects with controlled difference. The difference is assigned by the tester. e.g. one group of objects is assigned to eat food A and another group is assigned to eat food B. => since we are assigning A or B, the difference is controllable. What is advantages of observational experiment? - cheap - easy since we just observed and recorded What is disadvantages of observational experiment? - Most of the time, the results are affected by the object's preference. e.g. the result shows that people who eat more veggies are healthier; but the result may be affected by the people's choices since people who eat more veggies are tended to care about their health more for example. What is disadvantages of experimental experiment? - expensive - more difficult to implement - most of the time, experiment is difficult to control e.g. want to find if people who have engineering degree tend to be better prepared for MBA. If for experimental experiment, we need to assign 2 groups of people to take engineering degree and other degrees. Then, assign them to take MBA in order to get the result. However, in the real world, we can't do it since we can't force people to take the chosen choice. What is advantages of experimental experiment? - More accurate since result may not affect by the object's preference What is matched pairs experiment? Sample A & B are not independent. There is a strong or weak connection between each pair of data in these two sample. For example, data 1 in sample A is somewhat related to data 1 in sample B. So, each pair is matched up in these two samples. How to evaluate the difference in means of the 2 samples in the experiment? we will form the difference between each pair as a new sample. Use t-statistics to figure out the test statistics with df = n-1, where n = size of the new sample. Why we want to perform matched experiment? b/c we want to reduce the variation btw two samples so that we can observe the true difference of the means. Since if the variance is too large it will reduce test statistics greatly. So, we can't reject H0 even though we really should. What is the difference btw df of matched experiment & unmatched experiment? matched => df is half of the unmatched. How to distinguish matched experiment? if there is connection between sample 1 and 2. If there is very small variance in the samples, then which design should we pick? We should choose usual non-matched experiment instead of matched since df of matched is smaller than df of unmatched. the test statistics's denominator will be smaller for the unmatched experiment, so we will have higher chance to reject H0 in this case.