Looks like no one added any tags here yet for you.
controlled experiment
scientific test done under controlled conditions (one or a few factors changed, all others constant
has independent variable, dependent variable, control groups, experimental groups, and constants (controls)
before the experiment
observation, scientific question, hypothesis, controlled experiment developed, research (typically)
independent variable
experimenter controls and changes across groups
ex) changes in water temperature for goldfish
dependent variable
changes in response to other variable
ex) change in respiration rate for goldfish
constants/controls
identical other variables across all groups
ex) goldfish species, size, tanks
experimental groups
receives some independent variable
control groups
no independent variable, baseline to see if IV has an effect
negative control groups
no additional treatment, identical to experimental groups just without the independent variable
positive control group
includes additional treatment that is known to cause the effect, identical to control groups except they have additional treatment and no independent variable
tests to see if the setup is working right
hypothesis
strong if it has a dependent variable, independent variable, is testable, and makes a specific prediction
“if-then”
bar graph
used for categorical groups
histogram
distribution of outcomes
count the occurrence of outcomes in divided ranges for continuous variables
one axis ranges, one axis frequences
BARS DO TOUCH
continuous variables
not discrete/categorical values, but along a spectrum
ex) temperature, time (usually), test scores
box and whisker plots
summarized data distribution
uses a “five number summary”
arrange smallest to largest, find minimum and maximum, median (middle number), lower quartile (left median), and upper quartile (right median)
scatter plots
overall relationship between two continuous variables
independent variable x-axis, dependent variable y-axis
draw line of best fit through average/middle of all points
line graphs
track changes over time between two continuous variables
x-axis usually time
chosen over scatter plots whenever individual changes from point to point need to be seen
pie charts
compare categorical groups proportionally (fraction or percentage of total)
graphs
have title, labeled axes with units if possible, scaled x and y axes
standard error of the mean
higher = more inaccurate, lots of deviation/variation
lower = less inaccurate, less deviation/variation
±2 of this to construct error bars (95% certainty not to chance)
chi squared test
1) null hypothesis (consistent with expected) and alternative hypothesis (doesn’t match expected
2) find expected counts
3) chi-squared value = sum of (o-e)²/2
4) degrees of freedom = possible outcomes - 1; p-value of 0.05 on chart
5) chi-squared less than critical value = NOT STATISTICALLY SIGNIFICANT
if chi-squared is greater than critical value, it is statistically significant (do not accept null).