Research Methods Exam 3

Maturation threat

Any biological, psychological, or physical change that happens naturally within an individual over the passage of time.
(e.g. participants in a long study perform worse on a final-reaction time test not because the task failed, but because they became tired or bored over the course of the day)

Attrition threat

When study participants drop out before an experiment ends, and those who leave are systematically different from those who stay.
(e.g. participants who are not seeing any weight loss or find the routine too difficult get frustrated and drop out of the study)

Placebo effect

a person’s physical or mental health improves after receiving an inactive or sham treatment, driven by the belief that the treatment will work.

Testing threats

when taking a pre-test influences a participant’s score on a later post-test; familiarity to the testing materials has an effect on score.
(e.g., children take a reading comprehension test (pre-test), complete the storytelling program, then take the exact same reading comprehension test again (post-test)).

History threat

when an external, historical event happens for everybody in the study at the same time. A history threat makes it unclear whether the outcome is caused by the treatment or by the external event or factor.

(e.g., a researcher tests a new reading program in the school. Halfway through the study, the entire school district unexpectedly adopts a new district-wide homework policy. Any improvement in student reading scores could be due to the district-wide policy rather than the new reading program)

Regression threat (or regression to the mean)

occurs when participants or subjects selected for extreme initial scores naturally closer to the population average (mean) on later measurements regardless of any experimental treatment
(e.g. the bottom 5% of students in a school are selected based on their low reading score and placed into a new tutoring program, when you test them again their scores do improvement, but it is partially because their first score was at an unlucky low point, not just because of the tutoring).

Selection effects

when a research sample or methodology is not truly random, causing the study group to differ from the actual target population in a systematic way.
(e.g., a researcher wants to test the general public’s awareness of healthy nutrition products, but collects data outside of a gym or health food store, skewing the data to be high overall knowledge for the entire population)

selection-history threat

the intersection of selection effects (research sample or methodology is not truly random, causing the study group to differ from the actual target population in a systematic way) and history threats (when an external, historical event happens for everybody in the study at the same time.)

Quasi-experiments

Researchers do not have full experimental control. Otherwise known as “natural experiments”. Quasi-experiments each have an “independent” variable and a dependent variable. Aimed to make causal claims.

Non-equivalent control group posttest-only design

Participants not randomly assigned to groups (nonequivalent control group), only tested once after exposure to one level of the independent variable or the other (post-test only part)

Nonequivalent control group pretest/posttest design

participants were not randomly assigned to groups (nonequivalent control group), and were tested both before and after (pretest/post-test part)

Interrupted time-series design

A variable is measured repeatedly before, during, and after “interruptions” caused by some event.

Example: Logic’s 1-800 song (A variable is measured repeatedly (in this example, daily hotline calls and daily suicide rates in the U.S.) before, during, and after “interruptions” caused by some event (the song’s release, the music videos, and award shows)

nonequivalent control group interrupted time-series design

combines: participants were not randomly assigned to groups (nonequivalent control group) and a variable measured repeatedly before, during, and after “interruptions” caused by some event.

multiple-baseline design

researchers stagger their introduction of an intervention across a variety of individuals, times, or situations to rule out alternative explanations.

stable-baseline design

a study in which the researcher observes behavior for an extended baseline period before beginning treatment or other intervention. If behavior during the baseline is stable, the researcher is more certain of the treatment’s effectiveness.

reversal design

researchers observe a problem behavior both with and without treatment, but then take the treatment away for a while ( the reversal period) to see whether the problem behavior returns.

Quasi-experiments versus correlational studies

In Quasi-experiments, researcher tend to select their samples more intentionally than in most correlational designs. In correlational studies, researchers select a sample, measure the two variables, and test the relationship between them. But, in quasi-experiments the researchers might target groups with particular qualities, target certain time periods, or seek out comparison groups provided by state laws.

Demand characteristics

participants guess what the study is about and change their behavior in the ecpected direction

Quasi-independent variable

a pre-existing trait, condition, or characteristic that a researcher uses to group participants, rather than a variable they actively manipulate

Independent variable

Directly manipulated and assigned by a researcher in a true experiment

Participant variable

the individual characteristics and pre-existing traits that each person brings into a study

Temporal precedence

a research design principle stating that a hypothesized cause (independent or predictor variable) must occur or change before the resulting effect (the dependent or criterion variable; used to establish a causal relationship between two variables.
(e.g., a loud noise must occur first, followed by a newborn baby crying, to prove the noise is what caused the crying)

Internal validity

the extent to which a research study accurately shows a cause and effect relationship between its variables, rather than the results being caused by other unwanted factors.

Mediation

how or why an independent variable (X) affects a dependent variable (Y) through an intervening third variable (the mediator, M)
X > M > Y relationship

(e.g., regular physical exercise (X) reduces an individual’s stress levels (Y); mediator (M): better sleep quality; causal sequence: exercise (X) increases sleep quality (M) hence, better sleep quality (M) lowers stress levels (Y).

predictor

often labeled X, a factor used in research and statistical models to estimate, forecast, or explain variations in a dependent or outcome variable (Y).

moderator

a moderator variable changes the strength, direction, or presence of a relationship between an independent and dependent variable.

r

Basic research/fundamental research; to understand how things work, discover fundamental principles, or fill gaps in theoretical knowledge.

b

Applied research/applied/developmental inquiry

To solve concrete, real-world problems, create a new product, or improve an existing system

(e.g., using viral replication data to design a specific vaccine or antiviral drug to stop an outbreak)

Pattern and parsimony

a research approach used to investigate causal relationships when direct experimental manipulation is difficult or unethical. Researchers examine consistent patterns of data across multiple studies and apply the principle of parsimony to find the simplest, most logical explanation that ties them all together.

Longitudinal study

a research design that collects data from the same group of participants repeatedly over an extended period - ranging from weeks and months to years or decades.

Multiple-regression analysis

a statistical technique used to predict a single continuous dependent outcome (y) based on two or more independent predictor variables