Inferential Statistics

test, effects, treatments, causality, explore relationships, variables, association, generalize behavior, sample, population

inferential statistics allow researchers to _____ the _____ of _____ (determining _____), _____ among _____ (determining _____), and _____ of _____ to a _____

estimate unknown population, sample, differences, association, assumption, sample represents population

inferential statistics uses decision making process to _____ characteristics from _____ data, include both tests of _____ and _____, requires an _____ about how well _____

probability, sampling

inferential statistics are based on _____ and _____

likelihood, occur, outcomes, proportion, time 

probability is _____ that any event will _____ given all possible _____, refers to the _____ of _____ we can expect

prediction, should, will, ratio, decimal

probability is used as means of _____ of what _____ happen over long run NOT what _____ happen, expressed as _____ or _____

guideline, decisions, sample, estimates, population, effects, chance

research probability is used as _____ for making _____ about how well _____ data _____ characteristics of _____ and determining if observed _____ are likely due to _____

no bias

probability is based on assumption of _____

error, differ, population 

sampling _____ is tendency for sample values to _____ from _____ values

unpredictable, chance, selection, random, valid representation, population

sampling error is _____ (occurs strictly by _____ due to _____), when estimating population characteristics assuming sampling is _____ and _____ of _____

sample mean - population mean

sampling error = _____

repeated, varying, sample means, normal, predictable, inferences, parametric

_____ infinite random samples of same population will exhibit _____ degrees of sampling error, plot of _____ will create _____ distribution, allows use of _____ properties to make _____ (using _____ statistics)

standard error of mean 

estimate of population standard deviation (variance) 

sample size, decreases, better estimate, population

as _____ increases, SEM _____, considered _____ of _____

sample, sampling distribution, sampling error, representative, population

larger _____ = less variable _____ = smaller _____ = more _____ of _____

range, population parameter, confidence 

confidence interval is _____ of scores that is likely to contain the _____ given a specified level of _____

meaningful, accurate, point estimate, value, relative accuracy 

confidence interval is more _____ (and _____) than _____ because it takes into consideration _____ and _____ of sample statistic

repeat, 95%, confidence interval, true population mean

CI states: if I were to _____ sampling, _____of the time _____ would contain the _____

constant, unknown, varies

population mean remains _____ (but _____) and confidence interval _____

precision, wider

increase confidence by sacrificing _____, interval is _____

no significant difference, chance, reject, disprove 

null hypothesis (H₀) states _____ between groups that cannot be explained by _____, either _____ or do not, aim to _____

significant difference, direction, focused question 

alternative hypothesis (H₁) states there is _____, with or without _____, should include elements of _____

significantly more effective

superiority trials show that treatment is _____ than placebo or standard care

no less effective, other benefits

noninferiority trials show that new treatment is _____, intent to show _____ for selection (safety, convenience, cost)

noninferiority margin, largest difference, functionally equivalent

noninferiority trials require _____ = _____ between two treatments that would still be considered _____

standard, better, new, new, not inferior 

noninferiority null hypothesis states _____ treatment is _____ than _____ treatment and alternative hypothesis states _____ treatment is _____

finding difference, no difference exists, reject null hypothesis, positive

type I error is _____ when _____, incorrectly _____, false _____

finding no difference, difference exists, not rejecting null hypothesis, negative

type II error is _____ when _____, incorrectly _____, false _____

one

type _____ error is more risky and is tightly controlled

alpha, 0.05, predetermined, significance, researchers 

probability of making type I error = _____, typical value is _____ (% chance of incorrectly rejecting null), _____ level of _____ set by _____

confidence interval

_____ = 1-alpha

p value

probability that statistical finding occurred by chance = _____

reject, accept alternative hypothesis, significant difference, chance

if p ≤ a _____ null hypothesis / _____, conclude there is _____ between groups or relationships that did not occur by _____

do not reject, no significant difference, chance 

if p > a _____ null hypothesis, conclude there is _____ between groups or relationship that can be explained by _____

effect, sample, null hypothesis, null hypothesis is true, prove, difference, large, chance

p value is probability of finding _____ as big as one observed in _____ when _____ is true, based on assumption that _____ although can not be used to _____, only tells how rarely to expect _____ this _____ in population by _____

small, more significant

not accurate to imply that _____ p value means observed result is _____

beta, 0.2, lower, small, important

probability of making type II error = _____, often use _____, use _____ (% risk of missing difference of expected effect size if it exists) if want to show _____ but _____ effect

statistical power

_____ = 1-B

rejection, statistical significance, sensitivity, important difference, truly exists

statistical power is probability that a test will lead to _____ of null hypothesis and attain _____, think of in terms of _____, more power = more likely to detect _____ that _____

power, alpha, sample size, effect size

power analysis involves (4)

differences, precision, significance, manipulated, importance, effect magnitude, sample size, statistical significance, increasing sample size

when comparing 2+ groups, _____ will always exist at some level of _____, statistical _____ can be _____, _____ and _____ are not accurately reflected due to role of _____ in probability, getting _____ by _____ means nothing

effect size, effect magnitude, sample size

_____ is used to indicate _____ and is less influenced by _____

magnitude, clinically relevant, together, statistical significance, independent, sample size, same, p value 

effect size quantifies true _____ of measured intervention, provides more _____ criterion than statistical significance alone but should be presented _____, not impacted by _____, _____ of _____, can stay _____ even if _____ changes

unitless, comparison, statistical test

effect size index is _____ standardized value to allow _____ across samples and studies, many types depending on _____ used

ratio, effect size, variance, 0.2-0.49, 0.5-0.79, 0.8+

effect size index is _____ of _____ to _____ within data, minimal effect = _____, moderate effect = _____, large effect = _____

variables, measurement scale, clinically meaningful, method, directionality, negative

effect size index judgment based on types of _____ used, understanding of _____, _____ change, and research _____, _____ is important consideration to understand _____ values

sample size, level of significance, power, sample size, significant difference, effect size 

a priori power analysis = estimating _____, specifying _____ and desired _____ can estimate _____ needed to detect _____ for an expected _____

educated guess, previous research, pilot data, estimate, opinion, meaningful difference, large, clinically important

during a priori analysis researchers make _____ based on _____/_____ OR base _____ on researchers _____ of _____ or how _____ an effect would be considered _____

smaller, larger, sample

_____ the expected effect size = _____ the required _____

power 

post hoc power analysis = determining achieved _____

large effect, sample, not significant, type II error, not effective, variance, recruitment, drop out

predict _____ and choose an appropriate _____ but scores are _____, not necessarily _____, possible treatment is _____, _____ in scores is greater than anticipated, _____/_____ issues with participants

nonsignificant trend toward difference, larger sample, power, further investigation

should be cautious with statement of _____, possible that _____ needed if _____ is an issue (warrants _____)

power, effect size, small, clinically meaningless

_____ can be high even when _____ is _____ and _____

effect size, confidence intervals, statistical significance, clinical expertise, underlying, patient 

should consider _____ and _____ in addition to _____, use _____ to understand _____ mechanisms of interventions and relevant _____ changes

patient values, clinical expertise, comprehensive

never underestimate power of _____ and _____ for _____ EBP