stat212 exam 3

experiments

designed to identify casual relationships

can assign participants to an intervention

observational studies

identify associations that may signal causation

can only observe what participants chose to do

reasons to complete an observational study

may be impossible or extremely difficult to assign participants to an intervention

may not be ethical

studies may be rare and difficult to produce without gathering a very large sample size or waiting a long time

gernally more expensive and may require time and extensive planning

scientific model

identify casual mechanisms trying to represent how vaiables relate to one another in some kind of system or process

cofounder

2 outcomes may be correlated but not causally linked

must truly cause direct changes in the response variable and be linked to the explanatory variable, but not necessarily in a known casual way

mediating variable

two outcomes may be casually connected but if the mediating varibale is disrupted the causality chain breaks

satisfaction

analytical process of breaking down our comparison groups into smaller subgroups based on a potential cofounder to see if differences still show up in the more targeted comparisons

cross sectional studies

collect both explanatory and response outcome data for a single point in time

ex: survey on vaping status and current known health conditions

cohort study

has explanatory outcomes and then at a later point in time we see if any differences emerge with their health

typically the response variable data is not available until a later time when we collect it (prospective-looking forward)

ex: might identify vapers and non vapers first then wait several years to see if any differences emerge with their health

case control studies

identify people based on having or not having certain response outcomes, then data for one or more explanatory variables are collected seperately perhaps by asking the participants or through records

typically retrospective(looking back) meaning the explanatory varibale data is not available unitl we collect it later

ex: might indentify poeple with pneumonia and compare them to similar people without pneumonia, then ask if they have a history of vaping

cohort studies allow for extended observation

often provide more data in real time and sometimes help researcheres better construct causaility arguments

however they can take a long tikme and may need lots of participants in rare incidence cases

case control studies allow for efficiency data collection

can directly identify people with this rare outcome rather than wait for it to happen

however we may have to rely on participants memory of previous activity or exposure or find records and we typically dont have the same detailed accounts of timing as cohort studies

single group pre-post designs

all units complete the same interventions in the same order, we then compare the pre and post measures to see if there is a systematic difference on average

muti group designs

can now seperate the treatment/control factors into seperate groups and potentially avoid other confounding differeneces, such as timing differences, test familiarity, or reactance/placebo effects

randomized controlled expiriments

use random assignment to sort units- it may be pure random assignment or random assignment with blocking by some relavant factors

random assignment

random chance process to sort units

random asssignment with blocking

first involves blocking units by possible confounding factors then randomly assigning from each block

non randomized controlled expiriment

ises a non random assingment method to sort groups

random controlled expiriment with repeated measures

by taking multiple measures at different time points we can better track how individuals report

may not always be adventageous

randomized controlled experiment with crossover trials

crossover trials are a special case of repeated measures, each group completes both experimental conditions and produce response measures from both

group selection

if we are comparing 2 or more groups in an expiriment the groups should be similar

Random assignment with large groups or random assignment with blocking for smaller groups are the best way to guard against systematic differences between groups

drop out differences

attrition- when participants do not continye the study

mortality- when participants pass away during the study

drop out differences can affect the causality argument

test familiarity

exacerbated ina single group pre post design with no control

timing effects

systematic differences between the groups intervention times that could lead to timing related cofounders

setting effects

do any other setting or experimental differences affect the response

placebo effect

improving because they know they are receiving something

researcher effects

researchers interacting with the participants know who is in which group they may act different around each group

use double blinding when this is a significant threat

environment condition differences

are environmental conditions different between the treatment and control conditions beyond the treatment factor you wish to study

independence

are the units in each group providing independent reponse outcomes

grorup dynamics may impact

generalizability

whether findings in our study extend to a broader population setting and time

ex: is this medication effective at increasing serotonin levels for adults of all ages

sampling and group sorting are not the same

simple random sample

ideal method of sampling in large sample studies two things should be true

every member of the population has an equal chance to be chosen

sampling remains independent: the possability of one person being chosen does not affect the chances that somoene else is chosen

undercoverage bais

some people in the population dont have an equal chance to be selected or contacted

volunteer bias

once being selected those who chose to participate may be systematically different than those who chose not to participate

survivorship

sampling process results in an inordinate number of successful cases that doesnt represent the population

convenience sampling

non random sampling plan

quota sampling

identifies key features that we wish to have proper representation of and attempts to sample a target number from each subgroup

weighted sampling

attempts to use a mathmatical model to weight responses if that person comes from a subgroup that is over or under represented in my sample

participant selection

does this group of participants represent the population

setting limitations

is the setting representative of all settings we wish to generalize to

physical environment

might the physical space that this study took place in affect any outcomes we found

social environment

how might any social interactions or specific people involved have affected the outcomes

contexual features

what other contextual features did this study take place within