Experimental Design

hypotheses

proposed explanations for observations

scientific hypotheses

testable (possible to evaluate), falsifiable (potentially false) hypotheses

scientific theory

powerful, broad explanation for related observations based on well-supported hypotheses and supported by different, independent lines of research

deductive

reasoning that proceeds from the general case to the specific case

inductive

reasoning that proceeds from the specific case to the general case

syllogism

name for the set of statements of deductive reasoning

major/minor premises

names for statements 1 and 2 of deductive reasoning

conclusion

name for statement 3 of deductive reasoning

certain inference

other word for deductive reasoning

probable inference

other word for inductive reasoning

independent/explanatory/predictor

variable manipulated by the researcher

dependent/outcome/response

variable not changed by the researcher; may change during the experiment (what is being measured)

bias

influence on results

blind studies

studies that limit participant knowledge to reduce bias

double-blind studies

studies that limit both participant knowledge and researcher knowledge of participant groups to reduce bias

correlation

used to describe a possible relationship/association between two factors

statistically significant

meaning there is a low probability that differences between groups is due to random chance

sampling error

the expected average of a sample is the same as the true average of the population.

null hypothesis

hypothesis that tries to explain patterns in data in the simplest way possible

confounding

an unmeasured variable that could interfere with the study

analysis of covariance (ANOVA)

statistical technique used to control for the influence of a confounding variable. lets us look at a confounding variable as a covariate

covariate

a continuous variable that potentially affects the response (dependent) variable, but is not necessarily controlled or manipulated by the investigator (measurable confounding variables)

p-value

the probability of making a Type I error

Type I error

statistical error where we reject the null hypothesis when it is true (false positive/false alarm)

Type II error

statistical error where we fail to reject the null hypothesis when it is false (false negative/missed it)

power

concept related to the probability of committing a Type II error. equals the probability of correctly rejecting the null hypothesis when it is false

press studies

studies where manipulation remains constant over time

pulse studies

studies where there is a one time application at the start of the study

independence

observations collected in one replicate do not have an influence on another replicate

replication

there are multiple plots within the same treatment or comparison group

randomization

random assignment of treatments or selection of samples

grain

the size of the smallest unit of study, individual replicate or plot

extent

the total area encompassed by all the sampling units in the study