Research Methods Test 3

what are an independent variables levels called?

conditions (the van kleef study had 2 levels, the serving size of the bowl)

How are design confounds and control variables related? (CYU)

control variables are used to eliminate potential confounds

what does it mean to "control for" a variable in regression?

controlling for a variable is the attempt to reduce the effect of confounding variables on an observational study. -ex: looking at income levels when looking at recess time and behavior

be able to recognize if an experiment meets criteria for establishing causation**

covariance, temporal precedence, internal validity

demand characteristics

cues in an experiment that tell the participant what behavior is expected (did the participants simply guess what the researchers expected them to do, and act accordingly? or did the manipulation actually work?)

how can regression look at multiple variables at once?**

does not establish causation because researchers may not be able to establish temporal precedence and researchers cannot control for variables they do not measure (there could be an important third variable -- one they did not consider -- that accounts for the association). However, multiple regression allows researchers to control for potential third variables, but only for variables they choose to measure

Besides generalization to other people what other aspect of generalization does external validity address? (CYU)

generalization to other situations

systematic variability *CHAPTER 11*

generally denotes an anomaly or inaccuracy in observations which are the result of factors which are not under statistical control.

how does using matched groups prevent selection effects? (CYU)

instead of the participants choosing choosing what group they will be in the researchers test them on some variable and the people with closest score are placed together.

How does random assignment prevent selection effects? (CYU)

instead the of the participants choosing what they want to be in, the researchers assign them randomly.

-*nonequivalent control group interrupted time-series design*:

it combines 2 previous designs (the non equivalent control group design and the interrupted time-series design) researchers dont have control over the manipulation of the IV or the assignment of the participants conditions

What does it mean when an effect size is large in an experiment? (CYU)

it means that the causal effect is stronger or more important.

How do manipulation checks provide evidence for the construct validity of an experiment? Why does theory matter in evaluating construct validity? (CYU)

manipulation checks is an additional DV that researchers can insert into an experiment to help them quantify how well an experimental manipulation worked. Theory matter in evaluating construct validity to see if the measures you used will be testing the right things so you do not get results that don not answer your question

how are interrupted tine-series designs and non equivalent control group interrupted time-series designs different from true within groups designs? (CYU)

only a true within-groups experiment can control the order of presentation of the levels of the independent variable

confound

possible alternative explanations or potential threats to internal validity

What are three reasons a researcher might conduct a quasi-experiment, rather than a true experiment, to study a research question? (CYU)

real world opportunities, external validity, and ethics

longitudinal designs help satisfy the what criterion? / what does multiple regression help establish?

temporal precedence / internal validity

how do you use a factorial design to test a theory

the example about alcohol and aggression

describe how you can tell whether a study has a three way connection (CYU)

the number of independent variables

*non equivalent control group pretest/posttest design*:

the participants are not randomly assigned to groups, and were tested both before and after some intervention

when researchers test an independent variable in more than 1 group at once, what are they doing?

they are testing whether the effect generalizes (external validity)

how does generalizability relate to problems associated with external validity? *CHAPTER 11*

when researchers test an independent variable in more than one group at once, they are testing whether the effect generalizes

what is a null effect? *CHAPTER 11*

when the independent variable doesn't make an impact at all

how can you graph an interaction with the independent variables, separatley

you can put either independent variable on the x-axis

Give at least 3 phrases indicating that a study used a multiple regression analysis (CYU)

"Controlled for" "Taking into account" "Correcting for"/ "Adjusting for"

when researchers want to test for interactions what do they use?

-*factorial designs* -the independent variables are called "factors" -in the most common factorial design, researchers cross the two independent variables; that is, they study each possible combination of the independent variables -ex: looking at whether the effect of driving while talking on a cell phone depended on the drivers age -each condition is called a cell (ex: older people driving while using cell phones, older people driving while not using a cell phone etc)

What are the threats to internal validity? *CHAPTER 11*

-*history*: result from a "historical* or external event that affects most members of the treatment group at the same time of treatment (not knowing if the dorms used less electricity because it was cold outside) -*maturation*: a change in behavior that emerges more or less spontaneously over time (ex: sugar diet at the camp and the boys just "maturing into" the camp) -*regression threats*: regression to the mean: when a performance is extreme at Time 1, the next time that performance is measured (Time 2), it is likely to be less extreme and closer to average (ex: world cup team scoring 7 points and people not expecting them to score as high the next time); they only occur in a pre-test/post-test design -*attrition threats*: a reduction in participant numbers that occurs when people drop before the end (ex: level of depression might have decreased because three of the most depressed patients had symptoms so severe they had to drop out of the study); it becomes a problem for internal validity when attrition is systematic (when only a certain kind of participant drops out) -*instrumentation*: occurs when a measuring instrument changes over time (ex: in observational research, the people who are coding behaviors are the measuring instrument, and over time they might change their standards for judging behavior) -*testing*: a specific kind of order effect, refers to a change in the participants as a result of taking a test more than once (ex: people might have become for practiced at taking the test, leading to improved scores) -*COMBINED THREATS*: -*selection-history threat*: an outside event or factor systematicaly affects people in the study - but only those at one level of the independent variable (ex: if the dorms that got the go green campaign has construction on their building that day and the workers used electric tools then that could have contributed to the electricity used) -*selesction-attrition threat*: only one of the experimental groups experiences attrition (ex: people dropping out of just the treatment group and not the control group) -*observer bias*: when researchers expectations influence their interpretation of the results -*demand characteristics*: are a problem when participants guess what the study is supposed to be about and change their behavior in the expected direction (ex: the patients know they are getting therapy so if they know the doctor expects them to get better, they might change their self-reports of symptoms in the expected direction); these can be controlled by using a double blind study or a masked design -*placebo effects*: occurs when people receive a treatment and really improve - but only because the recipients believe they are receiving a valid treatment -design confounds -selection effects: a confound exists because the different independent variable groups have different types of participants -order effects

how do you interpret main effects and interactions?

-*main effects* - researchers look at marginal means to inspect the main effects in a factorial design (the observed difference in marginal means is about what you would expect to see by chance if there were no difference in the population) -*interactions* -in a factorial design w/ 2 independent variables, the first two results obtained are main affects, the third result is the interaction effect - you can use a table, line graph, and bar graph -if the lines are NOT parallel there probably is an interaction

• For a specific design (e.g., 3 x 2) o How many IVs/Factors? o How many levels of each? o How many conditions? o How many main effects are tested? o How many interactions are tested?

-2 IV -3 for 1 and 2 for the other -6 conditions (cells) -2 main effects -2

what are the different types of within groups design?

-concurrent measure designs: participants are exposed to all levels of an independent variable at roughly the same time (ex: infants were shown two faces at the same time; an experimenter looked at which face they stared at the longest) -repeated measures design: participants are measured on a dependent variable more than once, that is, after exposure to each level of the independent variable (ex: the womens oxytocin levels were monitored as they interacted closely with their own toddlers and then monitored them again a couple days later with other toddlers)

what are potential ways to deal with carryover effects?

-counterbalancing: they present the levels of the IV to participants in different orders (ex: half of the women interacted with their own toddler before the other toddler and some did the opposite) -full counterbalancing: all possible conditions are represented (ex: if a repeated measure design has two conditions, there are 2 orders; A --> B and B --> A) disadv: if experimenters want to put at least a few participants in each order, the need for participants can quickly increase -partial counterbalancing: only some of the possible condition orders are represented (ex: present the conditions in a randomized order for each subject)

how does factorial design differ from what we've discussed before?

-factoral designs are more complicated because we are manipulating more than 1 IV in an experiment. Researchers cross 2 (or more) IV's to study each possible combination of IV's -used to test for interactions and theories -each IV is called a factor, each factor has a number of levels (treatments), and each condition is a specific combination of levels.

what is a matched group design?

-helps to eliminate selection effects -to create a matched group from a set of 30 people, the researchers first would measure the participants on a particular variable that might matter to the dependent variable, IQ, for instance, might matter to anagram ability. they would next match participants up set by set; that is, they would take the 3 participants with the highest IQ and randomly assign them to the different groups etc

even though you wont be asked to interpret one, in what situations would you need to consider a 3-way (or an n-way interaction?

-if strayer and drews decided to study not only the cell-phone and age independent variable, but also two kinds of traffic conditions

manipulation check / pilot study

-in an experiment, an extra dependent variable researchers can include to determine how well an experimental manipulation worked. (ex: For instance, in the humor example if no one in the 'funny story' condition rated the stories as humorous, then a null hypothesis result might mean that the stories were not funny rather than that humor has no effect on learning.) -a small study carried out to test the feasibility of a larger one, may use it to confirm the effectiveness of their manipulations

how does increasing the number of factors or levels in a factorial design affect the way you interpret the data?

-it increases the number of cells

if quasi-experimental design cannot establish causation, what is their value?

-it may be important to know that the relationship exists -Let researchers take advantages of real-world opportunities External validity Ethics (A lot of these topics would be unethical to force participants into groups. e.g plastic surgery)

How is a nonequivalent control group design different from a true independent-groups experiment? (CYU)

-only a true independent groups experiment randomly assigns participants to groups

what are the two types of independent/between groups design?

-pretest/posttest: participants are randomly assigned to at least 2 groups and are tested on the key independent variable twice, once before and once after exposure to the independent variable (ex: if elliot had included a pretest, the experiment would have been a pretest/posttest design) -posttest: participants are randomly assigned to independent variable groups and are tested on the dependent variable once; they satisfy all three criteria for causation (ex: participants were randomly assigned to a red ink, black ink, or green ink group)

what does significant beta mean? what does a nonsignificant beta mean? (CYU)

-significant beta is when p is less than 0.05 -nonsignificant beta is when p is greater than. .05 meaning we cannot conclude that beta is different from zero

systematic variability vs unsystematic variability

-systematic: did the generous assistants work only with the large-bowl group? then it would be a design confound -unsystematic: random, not a confound; can make it difficult to detect differences in the dependent variable; should not be called a design confound -ex:if the students in the red ink group all happened to be bad at anagrams and those in the green and black ink groups were really bad at them, that would vary systematically with the ink color conditions, and would be a confound / but if some participants in each condition were good at anagrams and some were not, that would be unsystematic variability and would not be a confound)

how can you detect an interaction from a table of means? from a line graph? (CYU)

-table: row differences, being with one level of the first independent variable (the alcohol photos), the difference between the aggressive and neutral words conditions for the alcohol photos is 551-562= -11 -line graph: if it is not parallel, there is an interaction

What is the beta coefficient?

-the beta that is associated with a predictor variable represents the relationship between the predictor variable and the criterion variable, when the other predictor (independent) variables are controlled for -ex: as recess minutes goes up, behavior problems go down, even while we statistically control for the other predictor variable in this table, the free lunch -basically like r but it takes into account third variables that are being controlled for -helps researchers get closer to making a causal claim; and can help them get a sense of which factors most strongly affect classroom behavior problems

what is a main effect?

-the overall effect of one independent variable on the dependent variable -ex: there is not much of a difference between reaction times to aggressive words and nuetral word. in technical terms, there appears to be no main effect of word type -basically separating the results to analyze the independent variables effects separatley -marginal means: the means for each level of the IV -the overall effect of one independent variable at a time

What is a quasi-experiment? and what are the different designs within it?

-the researchers don't have full control -they might not be able to randomly assign participants to one level or the other --*non-equivalent control group design*: a study that has at least one treatment group and one comparison group, but participants have not been randomly assigned to groups; participants in one group are exposed to a treatment, a nonequivalent group is not exposed to the treatment, and then the two groups are compared. (the head start study) --*non equivalent control group pretest/posttest design*: the participants are not randomly assigned to groups, and were tested both before and after some intervention (the plastic surgery study) --*interrupted time-series design*: a study that measures the participants repeatedly on a dependent variable (in this ex: parole decision making) before, during, and after the "interreption" caused by some event (food break) --*nonequivalent control group interrupted time-series design*: it combines 2 previous designs (the non equivalent control group design and the interrupted time-series design) researchers dont have control over the manipulation of the IV or the assignment of the participants conditions (larceny rates; they didnt have control over the manipulation of the independent variable or the assignment of participants to conditions)

Why are quasi-experimental results often misinterpreted?

-they look just like an independent variable, they're analyzed just like one, but they are NOT! -you cannot draw causal conclusions from Quasi-IVs -you could only say that the effectiveness of the intervention differs by the quasi-IV (eg: men and women respond differently to an interventions) -even though you cant make causal claims, it may be important to know that a relationship exists

Describe which patterns of temporal precedence are indicated by different cross-lag correlational results (CYU)

-time 1 TV predicts aggression at time 2 / each variable is measured at different points in time so we can definitively tell which variable had an effect on the other by comparing the relative strength (which came first), establishing temporal precedence. -Either variable 1 at time 1 is correlated to variable 2 at time 2, or variable 2 at time 1 is correlated to variable 1 at time 2, or there might be correlations between both variables at both times when cross lagged with each other (i.e., all of the above mentioned).

what are the different ways you can interpret a null effect? *CHAPTER 11*

-weak manipulations: 4 shakes vs 3 shakes of hot sauce, not enough to make a difference; you need to ask how the researcher operationalized the independent variable aka construct validity -insensitive measures: sometimes a study finds a null result because the researchers have not used an operationalization of the dependent variable with enough sensitivity (ex: telling ur friend to try 2 spices when they don't even like spicy food so they are gonna thin both are really spicy) -maybe the explanation is that the independent variable really does affect the dependent variable -maybe the study wasn't designed well enough -ceiling and floor effect: all the scores are squeezed together at the top or bottom end

order effects & what do they include

-when exposure to one level of the independent variable influences responses to the next level of the independent variable (an order effect in a within groups design is a confound) THEY INCLUDE -*practice effects*: a long sequence might lead participants to get better at a task, or to get tired or bored toward the end (ex: red ink task) -*carryover effects*: some form of contamination carries over from one condition to the next (ex: imagine sipping orange juice right after brushing your teeth, the first taste contaminates your experience of the second one)

what is an interaction?

-when the effect of one IV depends on the level of another IV -ex: does the effect of cell phones depend on age

what is the difference between a within subject design and between subjects design?

-within: only one group of participants, and each person is presented with all levels of the independent variable -between(independent groups design): different groups of participants are placed into different levels of the independent variable

how can longitudinal data be combined with correlational data to help examine potential causal relationships?

-you can investigate how people change over time, and therefore establish temporal precedence (directionability problem) -cross-sectional correlations -auto-correlations -cross-lag correlations

why do researchers use quasi experiments if they can be vulnerable to internal validity?

-you cant randomly assign people when youre trying to figure out real life events -they can enhance external validity -ethical concerns: many questions of interest to researchers would be unethical to study in a true experiment (ex: wouldnt be ethical to assign people to have cosmetic surgery) -construct validity: you know you are successfully manipulated or measuring your variables

how many criterion variables are there in a multiple-regression analysis? how many predictor variables? (CYU)

1 criterion variable and at least 2 or more predictor variables

What are the different types of quasi-experimental designs?

1. *find naturally occurring groups and see if they differ*, Quasi because you don't manipulate it (ex: gender, before and after something happens/ smokers vs non-smokers) this is similar to a *interrupted time series design* 2. *compare an experimental group to a control group WITHOUT randomly assigning*: you still manipulate the IV but no random assignment to conditions this is also called a *nonequivalent control group design* (ex: pre-existing classrooms, cannot claim that groups were the same to begin with!)

Think of a possible mediator for the relationship between having substantive conversations and being happy. Sketch a diagram of the mediator you propose (CYU)

Education level ↗️ ↘️ S. Conver. ↘️ Happiness

in the texting and driving example, what are the factors and levels and cells?

FACTORS: (independent variables) cell phone use and age LEVELS: driving while using a cell phone or not; older or younger driver there are 4 cells in this design because 2 levels of one independent variable are crossed with two levels of another independent variable (2 x 2 = 4)

describe why bartholow and Heinz's word association study on alcohol and thoughts of aggression was a factorial design (CYU)

It was a factorial design because it consisted of 2 independent variables, photo type and word type.

Why is a longitudinal design called a multivariate design? (CYU)

Longitudinal designs are multivariate because even if they measure the minimum two variables to check for correlation, they measure them at two different points in time, which actually means they are measuring four variables all together.

Describe at least two cues indicating that a popular press article is probably describing a factorial design. (CYU)

Look for "it depends" and "Participants variables"

In an empirical journal article, in what section will you find the independent and dependent variables of the design? In what section will you find whether the main effects and interactions are statistically significant? (CYU)

The methods section outlines the studies independent and dependent variables where as the MANOVA shows the main effects and significance

What are 2 reasons that multiple regression analyses cannot completely establish causation? (CYU)

They cannot establish temporal precedence and the researching cannot control for every single external variable that may have an effect.

unsystematic variability *CHAPTER 11*

Unsystematic variance is variability within individuals and/or groups of individuals. This variability is essentially random; some individuals change in one direction, others in an opposite direction, and some do not change at all. For example, some individuals may feel better than they did yesterday, others feel worse than they did yesterday, and some feel the same as they did yesterday. Another example would be in taking tests.

Give an example of a question you would ask to interrogate each of the four validities for a multivariate study? (CYU)

Was there random sampling and from what kind of population? Ask about effect size and statistical significance? Look for subgroups, outliers, and curvilinear relationships

describe how counterbalancing improves the internal validity of a within-groups design (CYU)

With counterbalancing, any order effects should cancel each other out when all the data are collected.

how do within groups designs minimize unsystematic variability? (CYU) *CHAPTER 11*

Within-groups designs compares each participant with himself or herself & controls for individual differences.

what are the minimum requirements for a study to be an experiment? (CYU)

a manipulated variable and a measured variable

*non-equivalent control group design*:

a study that has at least one treatment group and one comparison group, but participants have not been randomly assigned to groups (the head start study and the plastic surgery study)

*interrupted time-series design*:

a study that measures the participants repeatedly on a dependent variable (in this ex: parole decision making) before, during, and after the "interreption" caused by some event (food break)

Participant variables

a variable whos levels are selected, not manipulated. because the variables are not manipulated, variables such as age, gender, and ethnicity are not truly "indpendent variables" but researchers call them that for simplicity reasons

describe how the same 2x2 design might be conducted as an independent-groups factorial, a within-groups factorial, or a mixed factorial design (CYU)

an independent groups factorial requires the most participants, and a within-groups factorial requires the fewest `

measurement error *CHAPTER 11*

any factor that can inflate or deflate a person's true score on a dependent measure

why do experiments usually satisfy the three causal criteria? (CYU)

*covariance* is covered through groups that are compared to each other, treatment groups and control groups. *temporal precedence* is covered because it manipulates causal variable then tests the effect that variable had on the other. and *internal validity* is covered because the proposed causal variable and not other factors is responsible for the change in the effect variable.

what is the difference between cross-sectional, auto-correlations, and cross-lag correlations? (which are the most critical to establishing temporal precedence and causation?)

-*cross-sectional correlations*: test whether to see if 2 variables measured at the same time are correlated (ex: study reports that the correlation between a preference for TV violence in 3rd grade and aggression in 3rd grade was r=0.21) -*auto-correlations*: evaluate the associations of each with itself across time (ex: the eron team asked whether preference for TV violence in 3rd grade is associated with preference for tv violence in 13th grade) -*cross-lag correlations*: shows whether the earlier measure of one variable is associated with the later measure of the other variable (ex: was tv violence in the 3rd grade associated with aggression in the 13th grade) -cross-lag correlations are the most critical

graphically, what does an interaction look like?

-*crossover interaction*: lines cross over eachother (when people eat icecream, they like their food cold more than hot; when people eat pancakes, they like their food hot more than cold) -*spreading interaction*: lines are not parallel, and they do not cross over eachother; When an efect exists at one level of a second independent variable but is weakter or nonexistent at a different level of the second independent variable. (when i say nothing, my dogs probability of sitting is the same in both conditions: treat or no treat. But when i say sit his probability of sitting is higher in the treat than the nontreat condition) OR (when i say nothing, there is zero difference between the treat and no treat conditions. when i say sit there is a large difference between the treat and no treat condition)

design confound and selection effects

-*design confounds*: an experimenters mistake in designing the independent variable; it is a second variable that happens to vary systematically along with the intended independent variable and therefore is an alternative explanation (ex: if van kleef had accidentally served a more appetizing pasta in the large bowl than the medium bowl, the study would have a design confound, because the second variable, pasta quality, would have systematically varied along with the independent variable) -*selection effect*: occurs when the kinds of participants in one level of the independent variable are systematically different from those in the other; can occur when the experimenters let participants choose which group they are in; can also result if the experimenters assign one type of person (all women) to one condition, and another type of person (all men) to another condition (ex: autistic group)

why can't some bivariate correlational study meet all three criteria for establishing causation? (CYU)

-cant establish temporal precedence or internal validity (1) Covariance: yes, proved that there is a correlation! (2) Temporal Precedence: cannot say which variable came first and caused the second variable (3) Internal Validity: third variables are not usually controlled for!

how does pattern and parsimony help us understand that the relationship between cigarette smoking and lung cancer IS causal? / how is this similar to a psychologist focusing his/her entire research career on one topic?

-A psychologist would study all types of different variables that all point in one direction --even though multi-regression analyses can control for these 3rd variables, critics could always argue that regression can't control for every possible 3rd variable -ex. ppl who smoke are also nervous→predisposition to cancer -even though experiments can rule out many 3rd variables, can't ethically assign ppl to smoke and then see what happens so can only do correlation -must specify mechanism for causal path (that covers all patterns of the data) --more smoking→more toxic chemicals contact human tissue→incr chances for cancer --set of predictions that can all be explained by single, parsimonious theory that chemicals in cigarettes cause cancer -Pattern and parsimony shows us that through the multitude of studies, many different studies point in the same direction

what is the relationship between a dependent variable score, a participants true score, and random error? (what causes variability within groups and obscure true group differences) *CHAPTER 11*

-Dependent Variable score=participant's true score +/- random error of measurement -measurement error: the more sources of random error there are in a dependent variables measurement, the more variability there will be within each group in an experiment; in contrast, the more precisely and carefully a dependent variable is measured, the less variability there will be within each group (measurement error) -individual differences: some people might be more cheerful or grumpy -situation noise: external distractions of any kind -power: when researchers design a study with a lot of power, they are more likely to detect true patterns, even small ones

under what circumstances does it seem appropriate to use small-n designs? (multiple baselines, reversals, etc)

-Even a small sample can detect detect a large effect size, meaning that statistical validity is not necessarily undermined by a small design. (split brains) -*small base line designs*: researcher observes behavior for an extended baseline period before beginning treatment -*multiple baseline design*: researchers stagger their introduction of an intervention across a variety of contexts, times, or situations (ex: disabled girls) -*reversal design*: a researcher observes a problem behavior both with and without treatment, but takes the treatment away for a while to see whether the problem behavior returns (if the treatment was really working, the behavior should worsen when the treatment is discontinued) -Behavior-change Studies in clinical Settings (There are three) Convenience .

explain why each of the five steps in a mediation examination is important to establishing evidence for a mediator (CYU)

-It establishes a correlation and temporal precedence while taking into account this third variable. With all causal claims being confirmed, it can establish evidence for the proposed variables. 1.) test for relationship c -is __ associated with __? --> if not, there is no relationship 2.) test for relationship a -is __ associated with the proposed mediator? 3.) test for relationship b -is mediator associated with __? 4.) run a regression test -to see whether relationship c goes away 5.) establish temporal precedence

Summarize three threats to internal validity discussed in this chapter (CYU) *CHAPTER 11*

-Maturation, your participants naturally change over the course of the study. -Instrumentation: the measurement itself cause change in DV -Attrition: when you lose participants over the course of the study

what is the difference between a moderator and a mediator?

-a *mediating variable* is internal, it asks "why are these two variables linked?" (ex: a mediation hypothesis could propose, for instance, that medical adherence is the reason why conscientiousness is related to better health) -a *moderating variable* is external, and asks "are these two variables linked the same way for everyone?, or in every situation?" (ex: a moderation hypothesis could propose that the link between conscientiousness and good health is strongest among younger people)

what is carryover? and what are the 6 different types of carryover effects for within-subject designs***?

-a carryover happens when some form of contamination carries over from one condition to the next (ex: imagine sipping orange juice right after brushing ur teeth; the first taste contaminates your experience of the second one) -TYPES 1.

in a factorial design what is a moderator?

-a moderator is an independent variable that changes the relationship between another independent variable and a dependent variable -in other words, a moderator results in an interaction; the effect of one independent variable depends on (is moderated by) the level of another independent variable (ex: they found that driver age did not moderate the impact of cell phone use on braking onset time)

What is a latin square design? and how is it useful?

-a technique for partial counterbalancing is *latin square*, a formal system of partial counterbalancing that ensures that each condition appears in each position at least once -Treatments are assigned at random within rows and columns, with each treatment once per row and once per column. -Useful where the experimenter desires to control variation in two different directions

adv and disadv of between groups design

-adv: a way of avoiding the carryover effects that can plague within group designs; random assignment of participants to groups eliminate bias -disadv: potentially differences in groups; more resources are needed; less statistical power; matching takes time and effort and assumes no transfer from matching operation

what are the adv and disadv of within subject designs

-adv: ensures that participants in the two groups will be equivalent; requires fewer participants overall; gives researchers more power to notice differences between conditions -disadv: carryover effects(order effects), might not be possible or practical(teaching kids how to ride the bike); when people see all levels of the independent variable and then change the way they would normally act

What is a mixed factorial design? why might it be useful?

-an example is the cell phone experiment; the age was a between groups because the ages varied but the cell phone condition independent variable was manipulated as within groups (each participant drove in both the cell phone and the control conditions of the study) -A research design that mixes between-subject & within-subject in the same design.

Recognize how threats to internal validity are slightly different when being applied to quasi-experiments. (selection effects, design confounds, maturation effect, history threat, regression to the mean, attrition threat)

1. *selection effects*: only relevant for between groups design; it is not clear whether it was the independent variable or the different types of participants in each group that led to a difference in the DV between the groups (ex: in the head start study the children in 1 group had different backrounds than the other so you dont know whether the achievement scores were lower because of the head start program or because of the this group had different kinds of children in it); use matched groups or wait-list design 2. *design confounds*: some outside variable accidentally and systematically varies with the levels of the targeted independent variable (ex: in the food break experiment, were the prisoners the judges saw right after they ate lunch guilty of a less serious crime?) 3. *maturation threat*: relevant in pretest/posttest studies; participants show an improvement over time but we don't know what caused it (ex: did the women who got the surgery just naturally get higher self esteem?) a way to avoid this is by using a comparison group 4. *history threat*: when an external, historical event happens for everyone in a study at the same time as the treatment variable (ex: did the rise in larceny have to do with TV or because of an economic recession?) 5. *regression to the mean*: the phenomenon that if a variable is extreme on its first measurement, it will tend to be closer to the average on its second measurement—and if it is extreme on its second measurement, it will tend to have been closer to the average on its first. (ex: in the head start study the head start students might have done so well at first because of a combination of lucky factors that they didnt have when they took the post test) using random assignment to place people in groups eliminates regression to the mean 6. *attrition threat*: when people drop out of a study over time (ex: in the cosmetic surgery study, you might wonder whether peoples self image improved only because those who were dissapointed with their surgery outcomes stopped responding to the study over time; but because the dropouts were not systematically different from completers, there wasnt a threat to internal validity) 7. *testing and instrumentation threats*: kind of order effect in which participants tend to change as a result of having been tested before (ex: judges objectively changed over time with fatigue) -demand characteristics: when participants guess what the study is about and change their behavior in the expected direction -placebo effect (people who got the surgery expecting to feel better) -observer bias

How can a study minimize variability within groups? (CYU) *CHAPTER 11*

1. change to a matched group design (measurement error) 2. add more participants (individual differences) 3. control for situation noise

How can a study maximize variability between independent variable groups? (CYU) *CHAPTER 11*

1. reliable and precise measurements 2. measure more instances 3. change the design 4. add more participants

in a 2 x 2 x 2 design, what is the number of main effects and interactions? (CYU)

3 main effects and 3 two-way interactions and 1 three-way interactions

what is a responsible way for journalists to cover single studies on a specific topic? (CYU)

Ideally, journalists should report on the entire body of evidence, as well as the theoretical background, for a particular claim

why is it difficult to simultaneously have strong internal AND external validity? *CHAPTER 11*

In an experiment, researchers often prioritize experimental control-internal validity. to get a clean manipulation, they may have to conduct their study in an artificial environment, such as a University laboratory. Many scientists choose to sacrifice representatives for internal validity

Why do many researchers find pattern and parsimony an effective way to support a causal claim? (CYU)

In psychology, researchers commonly use a variety of methods and many studies to explore the strength and limits of a particular research question. Often using a broader, simple explanation for how variables relate can keep confounds/third variables limited that occur when one gets more specific with a question.

what are two common reasons to use a factorial design? (CYU)

In the most common factorial design, researchers cross the two independent variables; that is, they study each possible combination of the independent variables. The process of using a factorial design to test limits is sometimes called testing for moderators. (moderator: variable that changes the relationship between two variables) / *to test the limits of an effect and to test theories*

What is a one-group, pretest/ posttest design, and which threats to internal validity are especially applicable to this design? (CYU) *CHAPTER 11*

One-group pretest posttest design is an experiment in which a researcher recruits one group of participants, measures them on a pretest; exposes them to a treatment intervention, or change, and then measures them on a posttest. The threats to internal validity include participants figuring out what they are testing and changing their behavior thus skewing the results.

Describe why both the design and the results of a study are important for assessing a quasi-experiment's internal validity. (CYU)

Reveals its vulnerability to alternative explanations, such as selection, maturation, history, attrition, testing, and instrumentation effects; observer biases, demand characteristics, & placebo effects.