Please enable JavaScript.
Coggle requires JavaScript to display documents.
Producing Data (Necessary Vocabulary (Three Principals of Experimental…
Producing Data
Necessary Vocabulary
Experimental Units
A single entity that is being studied,
such as monkeys or plants or cars, as
long as they are not humans, which are
known as subjects
Placebo Effect
A treatment that has no real effect,
but subjects believe their is because
that is what they are told
Lurking Variable
Neither of the variables under investigation
that can change the outcome of an experiment
or observation. This can be confounding because
it is unexpected and not measured
-
-
Treatments
A combination of factors and their
levels that are introduced to the
groups in an experiment
Response Variables
The dependent variable of an experiment
that is changed by the explanatory or
independent variable
-
Survey Examples
SRS
Asking 100 random students in the hallway what their age is to find the average age of DHS students
Stratified SRS
Finding the average GPA of high school students by taking subgroups of freshmen, sophomores, juniors, and seniors
Cluster Sampling
Finding Massachusetts people's favorite holiday by making each town a cluster, and sampling each cluster
Sampling
DEFINITION
the process of choosing a representative sample from a target population and collecting data from them in order to understand something about the population as a whole
Errors
Non-sampling error
Wording Bias
DEFINITION:
Wording Bias is the way the question is written in a survey in a way to manipulate or influence the person filling out the survey.
EXAMPLE:
The questions "Should our cruel and unfair boss be fired?" and "Should our kind and well-meaning boss be fired?" are asking the same thing, but are not impartial and the answers will not be accurate
Response Bias
DEFINITION:
Response Bias: a response bias is someone who is taking the survey who would answer a question untruthfully that could mislead the data and influence the response.
EXAMPLE:
A person is filling out a survey and comes across a question asking if the person participates in illegal activities. The person does, but answers no because they do not want to get in trouble
Non-response
-
EXAMPLE:
A survey is mailed to someone, but the person decides to not fill it out, and throws it in the trash
Bias Definition
•Bias is how much favor one is for
or against one thing. bias is a type
of error that occurs while sampling
or testing and encourages one outcome
over another.
Bias Example: We go into the Yacht Club and ask the first 200 people where their favorite place to go out to dinner is.
This shows bias because it is a convenient sample. This sample is easy for the surveyor, but shows a bias towards the upperclass. The Yacht Club usually contains residents who belong to the higher class, this would mean the responses would be bias towards more expensive restaurants. The responses would be more spread out say somewhere like a grocery store where people with different tastes of all kinds go to shop.
Sampling Error
-
Undercoverage
-
EXAMPLE:
Asking everyone In Duxbury what they think about the ocean, and using that data to represent Massachusetts people's views on the ocean
-
Good Techniques
SRS
Selecting a smaller group
from the population, and
using them to to represent
said population
-
Disadvantages
•Difficult to contact people who are far away
•Can lead to sample size errors
•Prone to convenience sampling
Stratified SRS
Dividing a population
into subgroups, and
applying sampling
methods to each group
Advantages
•All subgroups get proper representation
•Provides better coverage
•Allows control over subgroups
Disadvantages
•Not possible in every study
•Several conditions must be met
•Sometimes, people fit into multiple subgroups
-
RANDOM DESIGN EXAMPLE
Conclusion -
This is a useful experiment because it combats the placebo effect since the subjects don't know who has the experimental treatment. It is also a randomized way to observe whether or not there is a difference in grades for the subjects with the experimental treatment. It can also be replicated.
Process -
100 subjects are randomly selected from different math class levels. Then they are randomly split into two groups. One group is drinking the experimental water, and one group is drinking regular water. Neither group knows which is which, and after two weeks, the math grades of all students are observed.
Question -
A new water is introduced that increases a students ability in math class. An experiment is conducted to see if this is true
SURVEY EXAMPLE
Process -
Create a list of numbers from 0001 from 1500 and assign students that are seen in the hallway a random number. Then randomly select numbers and record answers from the corresponding students. Then see which holiday got more votes.
Conclusion-
The use of this experiment is that a small portion of the DHS population can be surveyed on a question, but it is a large enough group to represent all DHS students. It can be assumed that a similar answer would be observed from interviewing every student
-
Observational studies
In an observational study,
the researcher has no
control over the variables
-
Experimental studies
In an experimental study,
the researcher does have
control over the variables
-
-