Chapter #5 - 09/27/24 Flashcards
experiments and observational studies
describe explanatory versus response variable :
An explanatory variable is what you think might cause or influence something, while the response variable is the outcome or effect you’re measuring
describe randomized experiment versus observational
studies :
randomized experiments control the setup, while observational studies watch what naturally happens
what are cofounding variables ?
an unforeseen and uncaccounted for variavle that jeopardizes reliability and validity of an experiment’s outcome
what is an explanatory variable ?
The explanatory variable is the cause
what is a response variable ?
The response variable is the effect
is the explanatory variable considered the independent or dependent ?
independent
is the response variable considered the independent or dependent ?
dependent
in the following example what is the explanatory and response variable ?
Study found that overall left-handed people die at a younger age than right-handed people.
response = age at death
explanatory = habdednenss (left or right)
what is a treatment ?
is one or a combination of categories of the explanatory variable(s) assigned by the experimenter
what is a randomized experiment ?
create differences in the explanatory variable and examine results (response variable)
what is a observational study ?
observe differences in the explanatory variable and notice whether these are related to differences in the response variable
what are two reasons we must sometimes use an observational study instead of an experiment …
1) it is unethical or impossible to assign people to recieve a specific treatment (e.g., smoking)
2) certain explanatory variables are inherent traits and cannot be randomly assigned (right or left handedness)
TRUE OR FALSE
The effect of a confounding variable on the response variable cannot be separated from the effect of the explanatory variable on the response variable
TRUE
are confounding variables bigger problems in explantory experiments or obervational studies ?
observational studies
in the following example :
“Study of the relationship between getting a college degree and having good health later in life”
what might be the explanatory variable ?
whether or not the person graduated from college
in the following example :
“Study of the relationship between getting a college degree and having good health later in life”
what might be the response variable ?
quality of health later in life
in the following example :
“Study of the relationship between getting a college degree and having good health later in life”
what might be the possible confounfing variables ?
eating and exercise habits
what are experimental units ?
smallest basic objects to which we assign different treatments in a randomized experiment
what are observational units ?
objects or people measured in any study
TRUE OR FALSE
participants are often volunteers ?
TRUE
explain “randomly assigning the type of treatment” :
- In the basic experiment, each participant is assigned to receive one treatment.
- Randomly assigning the treatments to the experimental units helps protect against hidden or unknown biases.
explain “ randomizing the order of the treatments” :
- In some experiments, all treatments are applied to each unit.
- Randomization should be used to determine the order in which the treatments are applied
what are “countrol groups” ?
- handled identically to the treatment group(s) in all respects, except that they don’t receive the actual treatment
what are “placebos” ?
- Research shows people also respond to placebos – looks like the real drug but has no active ingredients.
- Randomly assign some patients to receive the drug and others to receive a placebo, without telling which they are receiving.
what are the two kinds of blinding ?
double and single blind
define “double blind” :
neither the participant nor the researcher taking measurements know who had which treatment
define “single-blind” :
only one of the two (participant or researcher) knows which treatment the participant was assigned
between double and single blinding which could lead to possible bias ?
single blinded
what term is used to define “one not recieving the drug but will look completly similar” ?
placebo
what term is used to define “ those that did not recieve any treatment” ?
control groups
what is “matched-pair design” ?
- Use two matched individuals or the same individual to receive each of two treatments
- Randomization used to assign the order of the two treatments
what is “randomized block design” ?
- Extension of matched-pair to three or more treatments.
- Similar experimental units are assigned into groups (blocks).
- Then treatments are randomly assigned separately within each block. This is done for very block, thus making sure that all treatments are assigned at least once in each block
what is “repeated-measure design” ?
Block designs in which the same participants are measured at many times
what are “randomized block designs” ?
Similar experimental units are assigned into groups (blocks). Then treatments are randomly assigned separately within each block. This is done for every block, thus making sure that all treatments are assigned at least once in each block
FILL IN THE BLANK
Random assignment to treatments is used to reduce ____________ ___________ biases due to confounding variables, that might otherwise exist between treatment groups
unknown systematic
FILL IN THE BLANK
Matched pairs, repeated measures and blocks are used to reduce _______ _________ __ ________ _________ in the response variable, so that differences due to explanatory variable can be detected more easily
known sources of natural variability
Does regular physical activity during high school influence academic performance in college? Researchers examined data on 2500 students, for which they had information on both their physical activity levels during high school and their GPA during their first year of college. The study found that students who engaged in regular physical activity had, on average, higher GPAs than those who did not. This is an example of:
A) an observational study
B) a randomized experiment
C) a block design, with level of physical activity as the
blocks
A) an observational study
The study above also found that students who exercise regularly were more likely to be men and
having good eating habits. Gender and eating habits are examples of:
A) Blocking variables
B) Explanatory variables
C) Confounding variables
C) Confounding variables
where might we see potential complications in regards to difficulties and disasters in experiments ?
1) confounding variables
2) interacting variables
3) placebo, Hawthorne, and experimenter effects
4) ecological validity and generalizability
what is a probelm with confounding variables ?
variables connected with explanatory variable can distort results because they may be agent actually causing change in the response
what is the solution for problems with confounding variables ?
randomization => effects of confounding variables should apply equally to each treatment
what is a probelm with interacting variables ?
second variable interacts with explanatory variable but results reported without noting it
what is the solution for problems with interacting variables ?
researchers should measure/report variables that may interact with explanatory variables
what is a probelm with placebo, hawthorn, and experimenter effects ?
power of suggestion (placebo effect), just being included in a study (Hawthorne effect), and experimenter recording data erroneously, treating subjects differently – all these can bias results
what is the solution for problems with interacting variables ?
use double-blinding and control group, have data entered automatically in computer as collected
what is a probelm with ecological validity and generalizability ?
variables measured in labs or artificial setting, results do not accurately reflect impact in real world; results for volunteers may not extend to larger group
what is the solution for problems with interacting variables ecogical validity and generalizability ?
try to design experiment that can be performed in natural setting with a random sample from the population of interest; measure other variables to see if related to the response or the explanatory variables
what is a case control study ?
A case-control study is a type of observational study where two groups are compared: one group with a specific outcome or condition (the “cases”) and another group without it (the “controls”). Researchers look back to identify possible differences in exposure or risk factors between the two groups to understand what might be associated with the outcome
what are the two types of observational studies ?
restrospective and prospective
define “retrospective studies” :
participants are asked to recall past events. (e.g., examine charts of patients previously hospitalized)
define “prospective studies” :
participants followed into future, and events recorded. Better because people often do not remember past events accurately. (e.g., follow patients upon hospital admission)
what is the major advantages of case-control studies ?
- efficiency
(Efficient in terms of time, money, inclusion of enough people with disease. (e.g, study the relationship between owning a bird and getting lung cancer via a randomized experiment would take a very long time) - reducing potential confounding variables
(Controls chosen to try to reduce potential confounding variables (but must be careful not to introduce new ones)
what are 3 potential complications in regards to difficulties and disasters in obesrvational studies ?
1) Confounding variables and the implication of causation
2) Extending the results inappropriately
3) Using the past as a source of data
what is the problem regarding confounding variables and the implication of causation ?
no way to establish causation with an observational study – can’t separate out all potential confounding factors w/o randomization
what is the solution in regards to the problems related to confounding variables and the implication of causation ?
measure potential confounding variables; choosing controls as similar as possible to cases
what is the problem regarding extending the results inappropriately?
many use convenience samples, not representative of any population
what is the solution in regards to the problems related to extending the results inappropriately ?
researchers should use entire segment of population of interest
what is the problem regarding using the oast as a source of data?
retrospective studies unreliable –ask people to recall past behavior; confounding variables in past not similar current ones
what is the solution in regards to the problems related to using the past as a source of data ?
use prospective studies if possible; else use authoritative sources versus memory. (e.g., hospital charts)
random sample vs. random assignment
what is positive about random sampling ?
it extends results to a larger sample
random sample vs. random assignment
what is positive about random assignment ?
establishes a cause and effect