Research Concepts Flashcards
What is a positive control?
Ensures that a change in the dependent variable occurs when expected
What is a negative control?
Ensures that no change in the dependent variable occurs when none is expected
How to calculate the standard deviation?
- Calculate the mean
- Calculate the difference between each value and the mean
- Square and add all of these differences
- Divide by n-1 (number of values)
- Take the square root of this final number
How to answer questions regarding which direction a graph is skewed toward?
depends on the tail of the graph
Can case studies determine causation?
No
What are the 2 things a researcher is ethically allowed to discriminate against?
Age
Number of people affected
What is the Hawthorne effect?
An electrical company performed an experiment to see if changing the lighting in their factory would affect productivity, but they told their employees about it, so each time they changed the lighting productivity increased → people change their behavior when they know they are being watched in the direction they think the researcher wants
When is a bar graph used?
when one of the variables is not continuous
Which is larger, the IQR or the standard deviation?
SD
What does equipoise mean?
nonmaleficence = do no harm!
Are you allowed to give a drug to people if they do not have a disease to test side effects?
Yes
What is a plausibility flaw?
The connection between variables is not scientifically plausible
How to make the distinction between random and systematic errors?
if increasing the sample size would fix the issue, then it is a random error
What does the confidence interval represent? What does it need to be?
indicates certainty with which the values are representative of the population.
Needs to be >95%.
What does coherence of results mean?
how the results relate to what is already known
What is the power of a study?
The probability that the study rejects the null hypothesis: opposite of alpha: probability that if you missed something it isn’t happening
What is a type 1 error?
False positive
What is a type 2 error?
False negative
What is a meta-analysis?
conducting research about previous research
What is external validity affected by? 4 things
subject selection
population characteristics
effect of time
effect of research environment
What is internal validity affected by? 5 things What does it measure?
instrument sensitivity sample size attrition inclusion of a non-randomized group data necessary to draw the conclusion was not collected
Measures the tendency of the same experiment to produce the same results
What is attrition?
People dropping out of a study
What are demand characteristics? What is the solution to this issue?
the tendency of a researcher to reveal, perhaps unintentionally, the point of the study to the participant, in such a way that the participant changes his/her behavior.
Solution = double-blinding
What does construct validity mean?
degree to which the study measures what you are trying to measure: how well does my MCAT grade affect my success in medical school?
How can one simply increase the accuracy of a measurement?
take multiple measurements and average the results together.
What is necessary for a causal link to be established?
the study must involve an experiment with the manipulation of an independent variable!
if the study only includes observations then it doesn’t prove a causal link
What is an error in precision?
Inability for an instrument to measure values consistently within a narrow range
What is an error in accuracy?
Inability for an instrument to measure the true value
What are the 3 types of observational studies?
- Cohort
- Cross-sectional
- Case-control
What does a cohort study do?
Records exposures throughout time and then assesses the rate of a certain outcome
What does a cross-sectional study do?
Assesses both exposure and outcome at one point in time
What does a case-control study do?
Assesses outcome status and then assesses for exposure history
What is selection bias?
Sample differs from the population
What is detection bias?
Educated professionals using their knowledge inconsistently by searching for an outcome disproportionately in certain populations
What are population data called?
Parameters
What is the difference between observational and experimental research?
The manipulation of the subjects’ environment
What populations are at increased risk of coercion?
Children, pregnant women, and prisoners
What is important when setting up a negative control?
Keep the conditions of 2 experiments as close as possible
What does unreliable data suffer from?
Random error
What does invalid data suffer from?
Systematic error
What is the mode?
The number that appears the most in the data
How do you calculate the median value?
median position = (n+1)/2
What is the difference between the mean of a sample and the mean of a population?
Mean of sample = statistic
Mean of population = parameter
What is a bimodal distribution?
A distribution with 2 peaks
In a negatively skewed distribution, which will be greater: mean, mode, or median?
Mode > Median > Mean
In a positively skewed distribution, which will be greater: mean, mode, or median?
Mean > Median > Mode
What is the IQR?
Q3 - Q1
What is the probability of one of 2 events occurring equal to?
P(A) + P(B) - P(A and B)
How to calculate the slope?
Delta y / Delta x
What is the probability of two independent events co-occurring?
P(A) x P(B)
What does alpha represent?
The significant p value, usually 0.05
A 95% confidence interval will fall within what distance of the mean?
2 sd
What value corresponds to the probability of a type 1 error?
alpha
What are the 2 ways to check if a value is an outlier?
- Take the IQR and multiply it by 1.5, then add that number to Q1 and Q3, anything below or above those 2 numbers is an outlier
- Any number that is 3 standard deviations from the mean is an outlier
How to calculate IQR?
n/4 and n.3/4 and then subtract those numbers
What is another name for the Hawthorne effect?
Observation bias
What is the difference between a case-control study and a cohort study?
Case-control: pick participants based on outcomes
Cohort: pick participants based on exposure
What does it mean to operationalize a variable?
Operationalizing a variable means finding a measurable, quantifiable, and valid index for your variable (independent and dependent variables)
