Research Method Exam 2 Study Guide
Parenthetical citations (2 ways)
(Brown, 2019) Uses &
Sensitization effect
After repeated exposures to an IV, participants realize (become sensitive to) what hypothesis is
References section of paper
Author1LastName, FirstInitial. MiddleInitial., & Author2LastName, FirstInitial. MiddleInitial. (Year#). Title with no uppercases: Except after colon. Capitalized Journal Name, volume#(issue#), articlefirstpage# - articlelastpage#. doi = Grady, J. S., Her, M., Moreno, G., Perez, C., & Yelinek, J. (2019). Emotions in storybooks: A comparison of storybooks that represent ethnic and racial groups in the United States. Psychology of Popular Media Culture, 8(3), 207-217. https://doi.org/10.1037/ppm0000185
____________ if each variable is correlated with ITSELF at MULTIPLE time points
Autocorrelations
__________ research studies association between exactly 2 variables ___________ research studies association between more than two variables
Bivariate correlational & Multivariate correlational
Non-parenthetical citation
Brown (2019) found ... Uses AND
3 things establish internal validity to conclude causal relationship
Cause leads to effect Cause 1st & effect 2nd Only causal variable is causing effect
Solution for minimizing regression to the mean threat
Comparison groups & experimental group are equally extreme at pretest! well be able see any regression effect in results
College student participants & adult participants recruited via online platforms (M-Turk) is example of what sampling?
Convenience sampling
___________ testing association between two variables (range from -1 to +1, vary in direction & magnitude)
Correlation statistics
___________ ensures all possible sequences of conditions are presented to participants reduce order effects (ex. Alone, Stranger & Stranger, Alone)
Counterbalancing
_________if earlier measures of 1 variable are correlated with later measures of other variable
Cross-lag correlations
_________if 2 variables, measured at SAME point in time, are correlated
Cross-sectional correlations
Fatigue effect
Decreased performance on task due to being more tired or less enthusiastic as study progresses
Why attrition threat is a threat to internal validity
Different rates of attrition between treatment & control groups can skew results. make variables appear correlated when they are not.
____________ occurs when attrition or dropout rates differ systematically between intervention & control group (proportion of control group & randomly assigned are missing).
Differential attrition
In study, theres association between deep talk & well being, both being measured short period, we cannot be sure whether people engaged in deep tlak first, followed by increase in well being or if people were happy first & later engaged in more meaningful conversation. Example of what problem?
Directionality
Regression to the Mean
Extreme measurements get less extreme with time (level next time its measured with or without experimental treatment or intervention)
Dr. Geerling is grading our final replication study papers, I graded them all in one sitting, in which she was tired/bored from performing tasks over and over again. Example of what effect?
Fatigue
_________ may indicate problems with IV or DV
Floor or ceiling effects
Posttest-Only Between-Subjects Experiment CONS
High risk of attrition bias Effect of intervention is not always clear Requires large sample size & Less generalizble
Dr. Geerling is grading our study papers. I find out shegraded your group's paper after a group that made her very frustrated, how do you feel?
I would feel bad because she will most likely to carryover her frustration on our exam, which can influence our grade
How can we use longitudinal designs to minimize the directionality problem in correlational research?
Longitudinal study help us see an effect over a long period of time, which allows us to see if there is a directionality problem in research (which same first & second).
Sleep quality (IV) affects quality of your work (DV) through alertness is ______ variable. Mental health status can moderate relationship between sleep & work quality, relationship is stronger for people without mental health conditions than for their counterparts is ______ variable.
Mediator & moderator
Posttest-Only Between-Subjects Experiment PROS
Meets all 3 criteria for causation Detecting differences in DV Establish internal validty & IV comes first in time
Treatment carryover effect solutions
Minimize any factors may influence other aspects of study OR counter balancing
How we can use moderators and mediators to minimize the third-variable problem in correlational research
Moderators & mediators give an important part of "story" of association between two variables (even though moderators & mediators are technically "third variables" )
2 author citations
Name both authors each time you cite the work. (Research by Brown and Smith (2019) supports... & Research supports a link between exercise and cognition (Brown & Smith, 2019).)
PRO of using correlational design
Natural/real world = more realistic (no manipulation) Can only use correlation in certain situations where things cant be manipulated (like age) Typically easier, faster, cheaper for researchers
Are all large samples representative of their respective population?
No, sample size is not a guarantee of ability to accurately represent target population
Why we do not use the one group, pretest-post test design in psychology?
Only one group, no comparison condition (ALL THREATS TO INTERNAL VALIDTY TO HAPPEN!), NO control group makes it difficult to see if attribute gains lead to results
Posttest-Only Between-Subjects Experiment
Participants are randomly assigned to IV groups & are tested on DV group ONCE (experiment using an independent group design)
Pretest-Posttest Between-Subjects Experiment
Participants are randomly assigned to at least two groups & are tested on DV twice, once before & once after exposure to IV (experiment using an independent group design)
Subject variables
Participants characteristics arent of primary interest but may influence study results (gender identity, ethnicity, race, income, education, etc.)
Interest & motivation is a _______ correlation Interest & boredom is a _______ correlation
Positive & negative
Participants are random assigned to a laptop or handwriting condition , and they were tested only once on the video they watched. Example of what experiment?
Posttest-Only Between-Subjects
Confounds
Potential threats to internal validity (in form of possible alternative explanations)
Participants completed a verbal GRE score, then are random assigned to two groups, mindfulness classes or nutrition classes, then after completion of group, participants take same verbal GRE score again This is an example of what experiment?
Pretest-Posttest-Between-Subjects
Importance of using double blind study procedures?
Prevents bias from both increases validity & reliability of the results
Random assignment to condition helps establish internal validity why?
Random assignment ensures cause is only reason for change, differ between two groups is IV exposure
Consequences of WEIRD Samples using American college students
Represent single stage in lifespan "sharp" cognitive abilities Less formed sense of self & attitudes Specific emotional experiences
One group, pretest-posttest design
Researcher recruits one group of participants, measures them on pretest by exposing them to treatment/intervention & then measures them again on posttest to determine change/difference between initial (pre-) & second (post-) measurement.
______________ is an error arising from taking a sample, rather than full population (equally likely to go in each direction)
Sampling error
Randoms assignment decreases __________ if we address DV at study beginning
Selection effect
Solution for minimizing instrumentaiton threat
Standardized & control of measurements, administration or data collection procedures, study personnel training
Addressing the Third-Variable Problem may use ________ variables
Statistical control
Effect size
Strength of relationship between variables ( .1 - small, .3 - moderate, .5 - large)
Longitudinal Designs
Studies measuring same variables in same people at several points in time
Correlational research
Studies where all variables are MEASURED, examine naturally-occurring relationships, without manipulating (bivariate & multivariate)
Double-blind study
Study in which assignment to condition is unknown to both participant & researcher (ex. both researcher & participants unaware which people are receiving actual drug or placebo)
_____________ is an error in data values obtained from a sample deviate by a fixed amount from true values within population (MORE PROBLEMATIC SINCE BIAS SAMPLES)
Systematic error
Pretest-Posttest Between-Subjects Experiment PROS
Testing of DV before & after intervention with IV (directionally of research) Enable researchers to track peoples change in performance over time no selection effects
Within-subjuect studies arent perfect because?
Threats to internal validity still exist! Reduces establishing direct relationship between variables effects can influence outcomes Carryover effects meaning specific order of different treatments affect outcomes
Large Within-Groups Differences
Too many individual differences within each condition & Too much situational noise (causes are measurement error, individual differences, situation noise)
Treatment carryover effect
Treatment administered earlier in study, some form of contamination carries over from one condition to next (affects participants)
Remember a sample coming from population is different from that sample being representative of population True or False
True
Solution for minimizing testing threat
Use different but equivalent forms of a test for pre-test & post-test (or no pretest) Use control group
3+ authors citations
Use only the first author's last name, followed by "et al." (Brown et al., 2019) or Brown et al. (2019) found ...)
Solution for minimxing maturation & history threat
Using comparison groups! using control group from same background to compare Shorter duration of experiment
External validity
Validity from ability to generalize study's results outside of a particular sample & context (generalizability is key!)
Statistical control variables
Variables that are considered to have an effect on DV in study but arent of particular interest to researcher
Small Between-Groups Differences
Weak manipulation & Insensitive DV (& ceiling or floor effects)
What does it mean that our participants in psych tend to be WEIRD?
Western, Educated, Industrialized, Rich, Democratic & convenient/easy to use as sample
How to interpret a correlation statistic
When 1 variable increases as other increases = postive correlation (strong & high) when 1 decreases as other increases = negative (weak & low)
Selection Effect
When differences between groups exist before manipulation even occurs (really bad study idea!)
When is maturation & history a problem?
When it happens for everyone!
Why maturation threat is a threat to internal validity
When its internal to individual participant, mental or physical changes occur within participants themselves that could affect results
When is regression to the mean a problem?
When we selected our sample based on extreme scores at pretest, can expect them to regress towards mean for post test
Directionality problem is threat for internal validity of correlational studies?
Which variable came first & which came second if we cannot tell which came first, we cannot infer causation!
Statistical power
ability to detect a statistically significant effect, if one exists
Within-subjects experimental design
all participants are exposed to all treatment conditons (same partipant tests all conditons, repeated measures)
On dating show, participants match with potential date, only some participants choose to go to love camp with their date. Only certain people, those in love, are left in the study and finish the study leaving as couples from the camp. Only couples are left. This is an example of what threat?
attruton
Selection effects affect our results at the ________
beginning
Imagine sipping OJ right after brushing your teeth, first taste is yuck ruins experience of the second sip. Example of what effect?
carryover
Correlation, DOES NOT EQUAL _________
causation!
________ effect scores cluster together at HIGH end of scale & ________ effect scores cluster together at LOW end of scale
ceiling & floor
Systematic attrition is when only a ?
certain kind of participant drops out
Low difficultly, high-volume studies?
conducting is easier by cheap online data collection platforms only takes days & larger sample
How third variable problems is a threat for the internal validity of correlational studies.
correlation between two variables can be misleading & even confusing (if third vairbale is not taken into account)
Solutions for practice effect?
counterbalancing or Give participants some time to warm up with each tasks before experiment begins
___________ is solution for order effects (important for within-subjects designs!)
counterbalaning
In longitudinal studies, 3 types of correlations to minimize directionality problem
cross sectional autocorelation crosslag
The sign of a correlation indicates the ________ of association, positive or negative.
direction
Random sampling
each member of population has an equal chance of being selected for sample or being assigned to condition (ex. Hunger Games, the reaping everyone 50/50 shot)
Random assignment to condition
each participant has equal & independent chance of being assigned to each condition in experiment (50/50 chance to go in condition)
Between-subjects experimental design
each participant is exposed to only one treatment condition (everything else/variables are same expect IV in which participants are assigned)
Between-Subjects Experiments & Random assignment to condition helps with internal validity because?
establish only cause for changes in DV is the IV
Why history is a threat to internal validity
events occuring outside of experiment can influence participants in a study (such as historical events)
Design confound
experimenter's mistake in designing study, when second variable happens to vary systematically along with IV
Mediators variables Asks why/how
explains process through which 2 variables are related (connection between two variables)
Consequences of random sampling for external validity
external validity is compromised since random assignment represents "ideal" conditions, which different from really happens
Study is __________ if results can be generalized to other groups in population, situations, settings, measures, apart from sample studied. (generalizability is key!)
externally valid
Dr. Geerling is grading our final replication study papers. If you find out that your group's paper is the last one I am grading, how do you feel?
good because of practice effect, she will probably grade it good
Moderator variables ask when/for whom
influencing strength & direction of relationship between 2 variables
When researcher rate behavior, they often improve over times. They become more accurate with pracice, but instrument used to collect data (i.e., trained observers) has changed from beginning to end of study. Example of what threat?
instrumentation
Random assignment to condition help us establish ______________
internal validity
History threat
its unclear wheather change in treatment group is caused by treatment or by external or historical factors effecting most members of group
When something changes, but it's not manipulation is _______ &___________
maturation & history
Threats to Internal Validity include
maturation, history, regression to the mean, attrition, testing, instrumentation
Consequences of conducting low-difficult, high-volume studies in pscyh?
may not always be replicable or yield accurate results internet samples have even crowded out college student samples! (short online studys increased frequency over time)
Consequences of using WEIRD Samples in psych?
may not be demographically or psychologically representative of humans as whole
3 possible reqasons null effects may occur
maybe no difference between groups in study we are just wrong between-groups differences = too small OR Maybe within-group differences = too large
When we get a null effect, its sad because?
means our hypothesis is wrong
CON of using correlational design
no cause & effect statements (NO CAUSATION)
Nonprobability Sampling Results biased (unrepresentative) sample Poor External Validity! (Hunger Games, katiness volunteering as tribute for prim) What kind of sampling?
non random
Maturation threat
oberved change in experiemental group have emgered more or less spontaneously over time
Sampling
obtaining our sample from our population (decide upon population of interest & obtain your sample)
Curriculum or teacher assessments, as teachers frequently use the _________ research design to assess their own effectiveness as teachers or effectiveness of the curriculum.
one-group pretest-posttest
Solution for minimizing attrition threat
only look at scores of participants who completed all of study (remove those who left) Recruit large participant group
Order effects
order in which IV is administered has an impact on DV
Practice effect
participants performance improves over time from previous practice with test (not because of manipualiton or treatment) (ex. taking ACT multiple times, will get better with practice)
Testing threat affect our conclusions?
partipants may get different/better scores if they have had repeated exposer to test, gives you practice
According to APA style, what tense should the Methods section be written in?
past tense!! intended to describe completed actions
___________ is set of all individuals in particular area of research & ___________ is subset of population from whom we collect data in our study
population & sample
Order effects types
practice effects, fatigue effects, sensitization effects, treatment carryover effects
Probability sampling unbiased & representative sample Great External Validity! generalizability to population of interest What kind of sampling?
random
Randomized within-subjects design
randomly determines, for each individual, sequence of treatment
Groups average is extremely depressed at pretest, since participants volunteered for therapy when they were feeling much more depressed than usual. Example of what threat?
regression to the mean
Consequences of non random sampling for external validity?
result in biased samples, creates poor generalizability to population (unrepresenative)
Null effect
results indicate there is NO difference between conditions in study
Participant in within-subjects design may behave very differently during last trial of experiment (participant knows what experiment is about) than participant did in arly trials (when participant unware what manipulation was). Example of what effect?
sensitization
Solutions for fatigue effect?
shortening procedures & time, making task less intense & more interesting/fun to keep participants engaged
By reducing random error due to differences between participants, we increase _____ power
statistical
The numerical value of a correlation indicates____________
strength of association
An attrition threat becomes a problem for intenal validty when the attition is _______
systematic
Testing threat
taking test twice changes one's scores, scores change over time as participants have taken test more than once (practice effects) (ex. participants take same survey at start & end, show improvement since shown test multiple times)
Experimental Studies
test cause-&-effect relationships MANIPULATE IV & measure its effect on DV, other variables controlled so they cant impact study results. (internal validity is proirtized)
Pretest-Posttest Between-Subjects Experiment CONS
testing threat, threat to internal vality, due to interaction between pretest & treatment
"Icecream sales increase crime", however summer is causing crime & ice cream sales, both are assocated with third variable which is summer season (icecream & crime is correlated but summer is causation of crime & icecream sales. Example of what problem?
third variable
The third-variable problem
third variable may be causing changes in both of our variables (in correlational studys)
Solutions for small between group differences that result in null effects
use manipulation check to fix ineffective manipulation, check for IV (for ceiling & floor effects)
Solutions for large within group differences that result in null effects
use precise measure & measure more often (for measurement error) measure more people use a controlled environment (for situation variability/noise)
The directionality problem
we cannot determine which variable came "first" & which came "second" we cannot tell which came first, we cannot infer causation!
Why Conduct a Within-Subjects Study?
we take care of individual differences problem & get more data out of our participants (only needs small sample)
Instrumentation threat
when measuring itself or measuring instrument changes over time causes misleading conclusions about intervention effect
Attrition threat
when only certain types of people make it to end of study (some participant drops out of study before it ends)