Chapter 6-9 Research Methods

Interpret different possible outcomes in cross-lag correlations, and make a causal inference suggested by each pattern.

1) A at time 1 predicts B at time 2 but B at time 1 does not predict A at time 2 2) B at time 1 predicts A at time 2 but A at time 1 does not predict B at time 2 3) A at time 1 predicts B at time 2 and B at time 1 predicts A at time 2

Why experiments are superior to multiple-regression designs for controlling for third variables.

1) problems if predictors are correlated 2) need large sample size 3) doesn't address temporal precedence 4) can't control for variables you didn't measure

Identify a mediation hypothesis and sketch a diagram of the hypothesized relationship. Describe the steps for testing a mediation hypothesis.

1. Test for relationship c. 2. Test for relationship a. 3. Test for relationship b. 4. Run a regression test both as predictor variables to see if relationship c goes away. 5. Causal variable is measured(manipulated) first in the study followed by the mediating variable, followed by proposed outcome variable.

Why a random sample is more likely to be a representative sample and why representative samples have external validity to a particular population.

A random sample is more likely to be representative because all members have an equal probability of being included in the sample. Representative samples have external validity because they can generalize to a larger population.

How longitudinal designs are conducted.

A study in which the same variables are measured in the same people at different points in time.

Unbiased Sampling

An accurate representation of a population

Way to increase the construct validity of Questions?

Avoid leading, double barreled, and double negatives

Double Negatives

Avoid these! Ask both ways to test for internal consistency. By using negative wording, this can cause the participant to get confused and be misled, resulting in them answering inaccurately.

Which of the following is not one of the tell tale signs that a researcher has used multiple regression in the research? A. Controlled for B. As caused by C. Correcting for D. Taking into account

B. As Caused by

Fatima is studying Ohio high school students. Shee samples participants by selecting a random sample of high schools in Ohio and surveying every student at each of the schools. Which sampling techniques is she using? Simple random sample- randomly choosing purposive sampling Cluster sampling

Cluster sampling because she chose each person from the clusters

How question wording can change the results of a survey or poll [6.1]

Double barreled, Question order, Double negatives, Biases

How to identify the best estimate of a correlation from a correlational study with two quantitative Variables If given a scatter plot

Effect size, cohen's d, and r stat

Understand correlational studies in which at least one variable is categorical.

Finding the correlation between variables that are categorical can be measured through r or through t-testing. Using a bar graph would be the most beneficial representation for this measure.

Do you like the student sitting next to you? Yes or No? What response scale is this?

Forced, they only have 2 options, they are forced to make a choice here.

How to interrogate the external validity of an association claim by asking to whom the association can generalize?

How was the sample selected? Is it generalizable?

A survey question that has you rate from "strongly agree" to "strongly disagree" is utilizing what? Likert scale Double barreled approach Semantic differential format

Likert Scale

Ways to improve the construct validity of observations by reducing observer bias, observer effects, and target reactivity [6.7]

Masked, Unobtrusive, Wait it out, Measure behavior

How observational techniques for measurement differ from survey techniques.[6.6]

More frequency claims, more accurate polls, more direct, and no bias or confounds

Ways to increase the construct validity of questions by preventing respondent shortcuts (e.g.,yea-saying), biases (e.g., trying to look good), or simple inability to report. [6.5]

Nay-Say, fence sitting, social desirability

How multiple-regression designs are conducted.

Need to measure your independent and dependent, but all several others (participant characteristics)

Unobtrusive

Observers hide

Biased Sampling

One or more parts of the population are favored over the others

Snowball:

Participants tell their friends about the study (recommendations)

Wait it out

Participants used to their presence before beginning the observation

Four techniques of nonrandom sampling

Purposive, Convenience, Quota, Snowball

Five techniques for random sampling: simple random, cluster, systematic, stratified random sampling, and oversampling.

Random, Cluster, Systematic, Stratified random, oversampling

Why it is more important, when assessing external validity, to ask how a sample was collected rather than how large the sample is.

Researchers generalize from sample to population; the margin of error (degree of sampling error) does not significantly decrease after 1000 as long as it is random; sample size NOT external

How question order can change the meaning (and validity) of a question.[6.4]

The first few questions can grab the participants' attention, the way they understand and grasp the concepts of these questions can affect how they answer the later questions based on how they interpret them.

Social Desirability

These can see anonymous surveys, filler items, and computerized tests. By adding these, it can help the participants answer accurately rather than answering based on judgement and the desire for social status.

Nay-Say

This included reverse worded items, this slows it down for people to get them to actually fully read the statement and gain understanding

Fence Sitting

This is a forced choice format. This allows the participants to answer a clear and constructed question. These questions have the participant either approve or disapprove of the statement asked.

Measure data

Unobtrusive data, traces measure left behind

Times when an unrepresentative sample may be appropriate/sufficient for a frequency claim.

When external validity is not vital to a study's goals (ex. shoe shoppers rating for a shoe)

Why representative samples may be especially important for many frequency claims.

You can't directly check accuracy so interrogate sampling technique, if random sampling is used, more confident in external validity

Rating Scales

a common method of data collection that is used to gather comparative information about a specific research subject. Specifically, a rating scale is a type of multiple-choice question and it allows survey respondents to assign a value to a product or service

Forced Choice questions

a format for question responses that requires respondents to provide an answer (e.g., yes or no), forcing them to make judgments about each response option.

Harriet and Lucas now calculated the interrater correlation coefficient to ensure that there is consistency in their observations. What would be the best value for them to obtain? a. 0.85 b. 0.05 c. -0.92 d. -0.03

a. 0.85 Want a positive correlation coefficient as they should be going in the same direction.

Which of the following is an example of observer bias in a study on arm strength and mood? a. Research assistance records the participants as stronger in the happy condition than the sad conditions, because that fits the hypothesis b. Participants performs with more strength in the happy mood condition because of subtle, encouraging caused from the research assistance c. A participant feels self-conscious in the experiment

a. Research assistance records the participants as stronger in the happy condition rather than the sad conditions, because that fits the hypothesis - The researchers are recording the participants and they are being biased as to how they rate them.

You can tell when a study is correlational because? a. The variables are all measured b. The researchers used the correlation coefficient r c. Used a t test d. Used a scatter plot

a. The variables are all measured They are all measured and none are manipulated

A correlational study that suffers from a restriction of range problem will probably accidentally find a correlation that is: a. Too small b. Too large c. Negative d. Positive

a. Too small - Make it look smaller with a restriction range, only accepting certain scores, people, etc.

A democratic national Committee mail survey included questions that read: "which aspects of the Trump presidency do you find the most disturbing?" followed by a list of 15 suggestions, such as "his erratic temperament and judgement" a. Leading b. Double barreled c. Negatively worded d. No issue

a. leading

open-ended questions

allow to collect qualitative answers from customers that are, for the most part, full of information. By asking an open-ended question, you are giving your customers the opportunity to answer whatever they like, without limiting or influencing them with predefined answers

How to interrogate the construct validity of an Association claim, asking whether the measurement of each variable was reliable and valid.

ask questions about the operationalization of the variables measured

Double Barreled

ask questions separately instead of asking multiple questions in one sentence, this way the participants are not getting confused and are giving accurate answers.

How to interrogate the statistical validity of an association claim

asking about features of the data that might distort the meaning of the correlation coefficient, such as outliers in the scatterplot, effect size, and the possibility of restricted range (for a lower than expected correlation). When the correlation coefficient is zero, inspect the scatterplot to see if the relationship is curvilinear. 1. What is the effect size? 2. Is there restriction of range? 3. Outliers? 4. Curvilinear? 5. Correlation statistically significant?

The why in the relationship: a. moderator b. mediator c. subgroup d. third variable

b. Mediator

Which of the following is an example of Observer Effects in a study on arm strength and mood? a. Research assistance records the participants as stronger in the happy condition than the sad conditions, because that fits the hypothesis b. Participants performs with more strength in the happy mood condition because of subtle, encouraging caused from the research assistance c. A participant feels self-conscious in the experiment

b. Participants performs with more strength in the happy mood condition because of subtle, encouraging caused from the research assistance - Encouraging cues are affecting the way the participant reacts, encouraging them.

An effect size of d=.53 would be: a. weak b. moderate c. strong

b.Moderate

Masked

blind study where the observers are unaware Change the wording, make it open ended, avoid biases

To establish that her and Lucas are coding behavior of the children in a similar way (they are consistent with one another), what statistic would they calculate? a. Beta b. P-value c. Interrater correlation coefficient (ICC) d. Cronbach's Alpha

c. Interrater correlation coefficient (ICC)

Which method is most likely to produce a random sample of the members of your class? a. Listing the first 6 students that come to mind b. Choosing 5 oldest students in the class c. Writing the name of each student on a separate piece of paper and drawing these slips from a hat d. Selecting the first 6 students that arrive at class

c. Writing the name of each student on a separate piece of paper and drawing these slips from a hat - Everyone has an equal chance at being selected, no bias as you do not know who you will pick

Systematic:

create a list of every member in the population and randomly select the first sample element from the first n elements population list. After, select every nth element on the list.

Three types of correlations in a longitudinal correlational design

cross- sectional correlations, autocorrelations, and cross- lag correlations.

What's the best sampling technique to use for determining the average speed of cars on a section of highway? a. Simple random sample b. Systematic sample c. convenience sampling d. A and B

d. A and B

In a study in which participants must provide feedback directly to others on their work, Joe notices that many people give positive evaluations of all of the work, even if the work is very poor. Which bias may be a problem in his study? a. Retrospective bias b. Participant reactivity c. Demand characteristics d. Social desirability

d. Social Desirability

What type of sample is this? "A large bakery mass produces cakes on an assembly line. Each shift, a quality control expert randomly selects one of the first 10 finished cakes, and every 10th cakes thereafter. Employees weight those cakes and give the cakes a detailed visual check a. Simple random sampling b. Stratified random sample c. Cluster random sample d. Systematic random sampling

d. Systematic random sampling - Because this is a systematic approach as they are randomly selecting every 10th cake.

Harriet wants to observe a kindergarten classroom during recess to examine gender differences in aggression. She recieves IRB and parental approval. For her observations, she dei ces that it would be best to sit in her car next to the parking lot close to the playground to observe the children's behavior, which she will code for further analysis later. Her colleague, Lucas will alternate days with her. a. Obtrusive b. Reactive c. Random selection d. Unobtrusive

d. unobtrusive

Stratified random:

demographic category is identified and then individuals are selected from each category

Dependent (outcome) variables and predictor variables in the context of multiple-regression data.

dependent/criterion: the variable you're most interested in understanding or predicting predictor: the rest of the variables measured

Explain the value of pattern and parsimony in research.

examination of a variety of correlational studies that all point in the same causal direction, idea that science is cumulative

How the correlation coefficient, r, represents strength and direction of a relationship between two quantitative variables.

if it's positive or negative shows direction, greater = greater strength, unless curvilinear then it doesn't explain it

oversampling:

in which the researcher over represents one or more groups

Articulate the difference between mediators, third variables, and moderating variables.

mediator: internal to the causal variable, not problematic, explain why the variables are linked third: external to the bivariate correlation, problematic moderating: ask for whom and when are the variables linked

Consider why journalists might prefer to report single studies rather than parsimonious patterns of data.

more validity when taken from a single study-- new information gets attention

Different ways questions can be worded, such as open-ended, forced-choice, and using rating scales. [6.2]

open ended, forced choice, and rating scales

Identify and interpret data from a multiple-regression table understand what each coefficient means.

read across, look at beta and significance to see if there is a relationship between that predictor variable and the criterion, and if it's significant

Cluster:

researcher divides the population into separate groups, then a simple random sample of clusters is selected from the population

Random:

sample is chosen completely at random from population of interest (ex. drawing names out of a hat)

Quota:

sampling a subgroup that is of great interest to a study

Purposive:

study a certain kind of person or group of people

Convenience:

study those who are easy to access

Distinguish an association claim, which requires that a study meet only one of the three rules for causation (covariance) from criteria is required for a causal claim

the study must meet only one of the three rules for causation (covariance), from a causal claim, which requires that the study also establish temporal precedence and internal validity.

When do we do multiple regression?

when it is unethical to do an experiment (like smoking effects on cancer)

Leading Questions

words questions neutrally or word them differently and look at the results to see if the wording affects the response. Responses should be neutral and not leading the particiaptns a certain way

A response set can be in the form of: Leading questions Yea-saying answers (fence sitting) Likert Scale

yeah-saying answers