LSU PSYC 2017 Cobb Terms You Should Know Exam 2

maturation threat

when participants naturally change over time, and those changes affect the study's results

Ex: Students get better at math over the school year simply because they grow and learn over time, not because of a new teaching method.

history threat

when an outside event happens during a study and affects the results

Ex: Students feel less stressed on a test because school was canceled the day before, giving them more time to study.

regression threat

when participants selected because they have extremely high or low scores move closer to the average on later measurements

Ex: Students who scored extremely low on their first test score higher on the second test simply because their scores naturally move closer to average, not because they improved from studying.

regression to the mean

the statistical tendency for unusually high or low scores to be closer to average when measured again

attrition threat

when participants drop out of a study, potentially making the remaining participants different from those who started

testing threat

when taking a test once affects participants' scores on a later administration of the same or similar test

instrumentation threat

when the measurement tool, scoring method, or observer changes during a study, affecting the results

selection-history threat

when different groups experience different outside events, which could explain differences in their outcomes

selection-attrition threat

when participants drop out of groups at different rates, potentially making the groups less comparable

observer bias

when a researcher's expectations influence how they observe, interpret, or record participants' behavior

demand characteristics

clues in a study that help participants guess its purpose and influence how they behave

double-blind study

a study in which neither the participants nor the researchers interacting with or assessing them know who is in which condition

masked design

a research design in which important information, such as participants group assignments or the study's purpose, is concealed to reduce bias

placebo effect

when participants experience a change because they believe they are receiving a treatment, rather than because of the treatment's active ingredient

double-blind, placebo control study

a study in which participants are compared with a placebo group, and neither participants nor relevant researchers know who receives the real treatment

ceiling effect

when a test is too easy or scores are already near the maximum, making it difficult to detect further improvement

floor effect

when a test is too difficult or scores are already near the minimum, making it difficult to detect further declines or differences

noise / measurement error

random or irrelevant influences that cause measurements to differ from participants' true values

power

the probability that a study will detect a real effect when that effect truly exists

design confounds

when two variables change together in a study, making it impossible to tell which variable caused the results

selection effects

when the groups in a study differ from the beginning because participants were selected into different groups in different ways

order effects

when the order in which participants experience conditions affects their performance or behavior in a study

measurement error

inaccuracies in how a variable is measured, causing scores to differ from the true value

individual differences

differences between participants, such as personality, intelligence, motivation, or prior experience, that affects their scores

situation noise

unplanned differences in the environment or situation that effect participants' performance

interaction (effect)

when the effect of the independent variable on the dependent variable depends on the level of another independent variable

explain how comparison groups, double-blind studies, and other design choices can help researchers avoid many of these threats to internal validity.

Comparison groups, double-blind studies, random assignment, standardized procedures, and counterbalancing help researchers rule out alternative explanations and increase confidence that the independent variable caused the observed results.

articulate the reasons why a study might result in null effects: not enough variance between groups, too much variance within groups, or a true null effect.

not enough variance between groups = the group averages are too similar

too much variance within groups = individual scores are too spread out

true null effect = the independent variable has no real effect

Describe at least two ways that a study might show inadequate variance between groups and indicate how researchers can identify such problems.

1. Weak manipulation: conditions are too similar. Identify with a manipulation check.

2. Restricted range: participants' scores or experiences don't vary enough. Identify by examining group scores and variability.

Why can large within-groups variance hide a between-groups difference?

Scores vary widely within each group, making the difference between group averages harder to detect and reducing statistical power.

factorial design

a research design that includes two or more independent variables so researchers can study their individual effects and interactions

cell

one specific combination of levels of the independent variables in a factorial design

participant variable

a characteristic of participants that researchers measure or use to categorize them rather than directly manipulate

main effect

the effect of one independent variable on the dependent variable, averaged across the levels of the other independent variable(s)

2 reasons to conduct a factorial study

1. to study main effects - researchers can examine the individual effect of each independent variable on the dependent variable.

2. to study interaction effects - researchers can determine whether the effect of one independent variable depends on another independent variable

describe an interaction as a "difference in differences."

An interaction occurs when the difference between groups on one independent variable changes across the levels of another independent variable. In other words, the effect of one variable depends on the other.

describe interactions using terms such as "it depends" or "especially for."

An interaction occurs when the effect of one variable depends on another variable or when an effect is stronger especially for a particular group.

How do you estimate marginal means in a factorial design to examine main effects?

Average the scores across the levels of the other independent variable(s). Compare these marginal means to determine the main effect of each independent variable.

How do you identify interaction effects in a table and a graph?

Table: Compare the differences between groups across conditions. If the differences change, there is an interaction.

Graph: Look for lines that are not parallel. The more they differ in slope, the stronger the interaction may be.

What is the basic logic of a three-way factorial design?

A three-way factorial design has 3 independent variables (IVs). Researchers examine the main effect of each IV, the interactions between pairs of IVs, and the three-way interaction.

How do you identify a three-way interaction on a graph?

The interaction between two variables changes depending on a third variable.

What keywords indicate a factorial design in a journal article?

Words like "depends on," "especially for," and "taking into account" suggest a factorial design because they indicate researchers are studying how multiple variables work together.

What keywords in popular media indicate a factorial design?

Phrases like "it depends," "especially for," or "the effect changes depending on" suggest researchers are studying how two or more variables work together.

experiment

a study where researchers manipulate a variable to see its affect on another variable

manipulated variable

a variable the researcher changes

measured variable

a variable the researcher observes or measures

independent variable (IV)

the variable that is manipulated to see its effect

condition

a specific version or level of an independent variable

dependent variable (DV)

the outcome that is measured

control variable

a variable kept constant so it does not effect the results

comparison group

a group used to compare results with another group

control group

the group that does not receive the main treatment being tested

treatment group

the group that receives the treatment or manipulation

placebo group

the group that receives a fake treatment that looks like a real treatment

confound

an outside variable that changes along with the independent variable and could explain the results

design confound

a problem where the way a study was designed creates an unintended difference between conditions

systematic variability

differences between conditions that are consistent and related to the study's variables

selection effect

a difference between groups that exists because different types of people are in each group

random assignment

using chance to place participants into different groups

matched groups

groups created by matching participants on important characteristics before assigning them to conditions

independent groups design

different participants take part in each condition

within-groups design

the same participants take part in every condition

posttest-only design

researchers measure the outcome after the treatment, without measuring it beforehand

pretest/posttest design

researchers measure the outcome before and after the treatment

repeated measures design

the same participants are measured across multiple conditions or times

concurrent measures design

participants experience multiple conditions and report their responses to them at the same time

order effect

the order in which conditions occur effect participants' responses

practice effect

participants improve because they have done a task before

carryover effect

one condition effects a participants' response to the next condition

counterbalancing

changing the order of conditions across participants to reduce order effects

full counterbalancing

using every possible order of conditions

demand characteristic

a clue that makes participants guess the study's purpose and change their behavior

manipulation check

a measure used to see whether the independent variable manipulation worked as intended

pilot study

a small trial run of a study conducted before the main study to identify problems

survey

a research method that asks people about their thoughts, feelings, or behaviors

open-ended question

a question that lets people answer in their own words

forced-choice question

a question that requires people to choose from specific answer options

Likert scale

a scale that measures how much someone agrees or disagrees with a statement (1 being strongly disagree, 5 being strongly agree)

leading question

a question worded to encourage a particular answer

double-barreled question

a question that asks about two things at once

negatively worded question

a question containing negative wording that can confuse respondents

acquiescence

the tendency to agree with statements regardless of their content

fence sitting

choosing a neutral or middle answer instead of taking a side

social desirability

giving answers that make you look good or socially acceptable

observational research

studying and recording peoples behavior without manipulating variables

observer effect

when people change their behavior because they know they're being observed

reactivity

when people change their behavior because they know they're being studied

population

the entire group of people researchers want to study

sample

a smaller group selected from the population to participate in a study

census

collecting data from every member of a population

biased sample

a sample that does not fairly represent the population

unbiased sample

a sample selected in a way that avoids systematically favoring certain members of the population

convienence sampling

choosing people who are the easiest to reach

self-selection

people choose for themselves whether to participate

probability sampling

sampling where every population member has a known chance of being selected

nonprobability sampling

sampling where members' chances of being selected are unknown

simple random sampling

every population member has an equal chance of being selected

systematic sampling

selecting every x'th person from a list after a staring point is chosen

(ex: every 5th person from a list)

cluster sampling

randomly selecting entire groups from a population and sampling people within those groups

multistage sampling

selecting a sample in multiple steps, often by randomly selecting groups and then people within them