1. Chapter 13- Experiments and Observational Studies
Presented By: The Danielles of Mr.Petrosky’s AP Statistics Class

2. Observational Studies?
It is a study based on data in which no manipulation of factors have been employed
researchers dont assign choices, the are just observed.

3. Retrospective?
It is an observational study in which subjects are selected and determined based on their past behaviors
They may contain errors because they are based on historical data
They have a restricted view of the world because they are usually restricted to a small part of an entire population

4. Prospective
It is an observational study in which subjects are followed to observe future outcomes
-collecting data as events unfold

5. Observational Studies..What are they used for?
discovering trends and possible relationships.
discovering variables related to rare outcomes..Ex: specific diseases
identify important variable related to the outcome we want ,but there is no guarantee that we have found the right/most important related variables
demonstrate a causal relationship

6. Randomized, Comparative Experiments
Experiments manipulates factor levels to create treatments, randomly assigns subjects to these treatment levels, then compares the responses of the subject groups across treatment levels
Random Assignment- to be valid, an experiment must assign experimental units to treatment groups at random

7. continues...
experiments must study relationships between two or more variables…
an experimenter must identify at least one explanatory variable or FACTOR…this is needed to manipulate
…and at least one response variable (variable whose values are compared across different treatments) is needed to measure.

8. still continuing... Experimental Units
humans that are experimented on are called subjects or participants
others are referred to as the general term: Experimental Units, when experimented on
- ( rats, days, petri dish of bacteria)

9. sadly continues...never ending...
The specific values that the experimenter chooses for a factor are usually called the levels of the factor
ex: participants sleep for 4, 6, or 8 hours
some exercise regularly, other are couch potatoes

10. How would we assign them you ask?
To ensure that the conclusion is fair, the experimenter must assign the participants their treatment at random

11. The 4 Principles of Experimental Design
1.) Control. We control sources of variation other than the factors we are testing by making conditions as similar as possible for all treatment groups.
we control a factor by assigning experimental units to different factor levels because we want to see how the response will change at the different levels.
we control other sources of variation to prevent them from changing and affecting the response variable

12. 2.) Randomize. This allows experimenters to equalize the effects of unknown or uncontrollable sources of variation. If the experimental units aren’t assigned to treatments at random, you will not be able to use the powerful methods of Statistics(YAYYY!) to draw conclusions from your study.
assigning at random reduces bias due to uncontrolled sources of variation.
randomization protects us even from effects we didn’t know about
“control what you can, randomize the rest”

13. 3.) Replicate. Two kinds of replication show up in comparative experiments.
1st: Repeat the experiment, applying the treatments to a number of subjects. you can estimate the variability of the response. you can’t find the variation, you experiment is not complete
2nd: experimental units are not a representative sample from the population of interest. Ex: If one tests that if one person in a household gets sick then everyone will get sick, and he expects this to be true for all humans, he has to replicate this same experiment for people in another country of different ages and different times of the year.
the outcome of an experiment on a single subject is an anecdote, not data
replication of an entire experiment with the controlled sources of variation at different levels is an essential step in science

14. 4.) Block. Experimenters use blocking to reduce the effects of identifiable attributes of the subjects that cannot be controlled.
Sometimes, attributes of the experimental units that we are not studying and that we can’t control may affect the outcomes of an experiment. If we group the similar experimental units together, then randomize between each of the blocks, we can eliminate most of the variability due to the difference among the blocks.

15. Need Help? A diagram can help in thinking about experiments.
The diagram emphasizes the random allocation of subjects to treatment groups, the separate treatments applied to these groups, and the ultimate comparison of results. Boxplots are a good way to start comparing.

17. Designing an Experiment
Step 1: Plan- state what you want to know
“I want to know…”
Step 2: Response- Specify the response variable
“I’ll evaluate…”
Step 3: Treatments- Specify the factor levels and the treatments
“The factor is ____. The different levels are____. These are the treatments.”
Step 4: Experimental Units- Specify the Experimental Units
“I’ll use ____ of the same variety…”

18. Step 5: Experimental Design- Observe the principles of design
Control, Randomly Assign, Replicate, Make a Picture
Be sure to specify how to measure the response
Step 6: Once the data is collected, display them and compare the results for the treatment groups
Step 7: Answer the initial question.

19. Does the Difference Make a Difference?
Does the differences we observed are about as big as we’d get just from the randomization alone, or whether they’re bigger than that. If we decided that they’re bigger,we’ll attribute the differences to the treatments. This means that we will describe the differences as Statistically significant. (too large to believe it has occurred naturally)

20. Experiments and Samples
Experiment and sample surveys use randomization to get unbiased data, but in different ways, for different purposes.
Sample Surveys try to estimate population parameters, so the sample needs to be as representative of the population as possible.
Experiments try to assess the effects of treatments.Experimental units are not always drawn randomly from population.
when subjects are assigned randomly to treatment groups, all the groups are still biased, but in the same way
when considering the differences in their responses, these biased cancel out, allowing us to see the differences due to treatment effects more clearly.

21. Control Treatments
Control Treatment is another level of the factor in order to compare the treatment results to the situation in which “nothing happens”. The experimental unit to whom this is applied to is the control group.
ex: tomato plant without fertilizer, a plant with half the specified amount of fertilizer, and a plant will the full does.
-Control Treatment- Plant without fertilizer

22. Blinded by what?
Any individual associated with an experiment who is not aware of how subjects have been allocated to treatment groups is considered “blinded”
There are 2 groups that can affect the outcome of the study…
those who could influence the results (subjects, treatment administrators, technicians
those who evaluate the results (judges, treating physicians, etc.)
Within the some treatments some are given a placebo or “fake” treatment. This is considered the best way to blind subjects.
The best experiments usually are:
randomized, double blind, comparative, and placebo-controlled

23. Sorry, I BLOCKED you.
But forgive me it was only because our experimental units were similar. By blocking we isolate the variability attributable to the differences between the blocks so that we can see the differences caused by the treatments more clearly. But it’s not the only way to reduce variation, you could match subjects. matching can be used to group subjects by in similar ways that aren’t under study.
See we have randomized blocks where the randomization occurs within our block

24. Factors of?
With complete randomization design all experimental units have an equal chance of receiving any treatment.

25. Lurking or cofounding?
Cofounding is when the levels of one factor are associated with the levels of another factor so their effects cannot be separated, we say these 2 factors are cofounded.
Cofounding is very similar to lurking because lurking they interfere w/ our ability to interepret our analyses.

26. What can go wrong?
Don’t give up just because you can’t run an experiment.
Beware of cofounding.
Bad things can happen even to good experiments.
Don’t spend your entire budget on the first run.