[r0cinante] Mathematics and Research Methods

knockout genes

-/- +/+ have both alleles (WT)

correlation

A measure of the extent to which two factors vary together, and thus of how well either factor predicts the other R = 0 means no relationship R = 1 means perfect relationship R = -1 means perfect inverse relationship

reliability

Ability of a test to yield very similar results over repeated testings

odds ratio

Ratio of the odds occurring in one group vs. the odds occuring in another group

calculating units

always write units on your numbers, check the units in the question and the answers for area and volume, conversion values are ^2 or ^3 1 m = 10^2 cm 1 m^3 = 10^6 cm^3

surveys

are cheap and easy

observational study

attempts to determine relationships between variables, but the researcher imposes no conditions as in an experiment

flaws in research

both internal and external validity are flaws its it generalizable is it proving causation

validity

construct validity- whether a measurement actually measures the variable/construct being studied criterion validity- whether a test correlates with other existing tests or measurements external validity- generalizing results to other people and situations internal validity- extent of causation proved ecological validity- whether experiment replicates real-world conditions

control

controls are important for comparing things you can be your own control if tested before and after a treatment

long term studies

cross-sectional study- one point in time cohort study- longitudinal study follows group over time

best experiment to prove a hypothesis

do an experiment with and without a certain variable to establish the strongest evidence for cause-effect all other variables are kept the same

controlling

eliminate influence of confounding variables, isolate the variable of interest

precision and accuracy

get a larger sample size

temporal

independent variable needs to be measured before the dependent variable

autoclave

kills everything you can autoclave a biohazard bag

variables

mediating variable- explains the relationship between independent and dependent variable confounding variable- influences both the independent and dependent variable

mortality/morbidity

mortality- deaths by disease during a time interval morbidity- either incidence or prevalence incidence- new cases of disease during a time interval prevalence- how many total people have a disease

necessary and sufficient

necessary- without it, no effect sufficient- with it by itself, has effect

causation

need a control without the variable to prove causation usually unethical in psychology and sociology experiments

ethnography

observing social interactions in real social settings

ethics

participants must volunteer, can withdraw at any time shocking people is ethical as long as you have consent

sampling

pick sample that represents population you want to study

which result does not validate results

pick the answer that directly contradicts the data in the passage, no just that one that isn't supported

best treatment

proactive treatment is better than reactive treatment even better if its specific, targeted with no side effects

median

reduces influence of outliers when compared to the mean

methodological limitations

sampling biases confounding variables

biographical sketch

study environmental influences of behavior

survey biases

subjective bias- self-reported social desirability bias- participants deal with sensitive/ethical subjects selection bias- participants are usually self-selected attrition bias- participants drop out of study over time

supporting vs. elaborating

supporting- discussing details and evidence to support a point elaborating- explaining in more depth

operationalization

the process of turning an abstract variable into a measured or manipulated variable, even if cannot be quantified directly

rearranging equations

to calculate a variable, rearrange equation so all other variables are on the other side

ratios control for variation

using a ratio measurement or a baseline comparison helps control for differences in cell count, conditions, etc.

counterbalancing

using subjects as their own control group by having half of them be experimental first and half of them be control first controls for the order of tests presented fatigue and boredom can affect test performance over time