_(:τ」∠)_ im so tired rn dude cmon

UGGHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH OHHHMYGOOODDDDDDD

True experiments

Require equivalent groups and all extraneous variables are controlled and counterbalanced.

Correlational studies

Studies that focus on the effects of subject variables where researchers can identify relationships but typically not casual effects.

Preexperiment

When a single group or multiple groups are observed after being given treatments prior to the true experiment. There’s no equivalent, if at all, comparison group (control) so researchers cannot conclude that any change exhibited by participants was due to the treatment. It can be used to test the feasibility for further study.

Studies that manipulate an independent variable, but for which equivalent groups are not possible use:

Preexpermental or quasi-experimental designs

One-shot study

A type of preexperiment where one group is tested only one time.

• Ex: A researcher wants to know if CPAP machine improves memory function for sleep apnea patients. The researcher then collects data from sleep apnea patients already using a CPAP machine. These patients are then asked to take a digital span memory test.

Nonequivalent control group design

(AKA Nonequivalent control group posttest only design or static group design)

Preexperiment type: Where a comparable control group is “identified,” it is comparable but not equivalent as random assignment was not used. Where an experimental and nonequivalent control group are tested once.

• Ex: A researcher identifies patients using a CPAP machine and a control group of people with apnea who don’t use it. Both groups took the digit-span test and the researcher found that those who used the CPAP machine performed better.

Pretest-posttest design

Preexperiment type: A design where one group is tested twice. Does not yield findings that researchers can assume are the result of causality.

• A group of apnea patients perform the digit-span test before using the CPAP machine and then again afterward. The researcher found that those patients performed better on the test after using the machine.

How is causation determined in a quasi-experiment?

Casual data can only be found when all confounds are eliminated in order for data to be confidently claimed as significant.

Pretest-posttest nonequivalent control group design

A form of quasi-experimental design that involves comparing the experimental group with the comparable, but not equivalent control group. Unlike the preexperiment version, both groups are tested before and after the introduction of the independent design.

• Ex: A researcher may identify apnea patients who haven’t begun treatment with the CPAP machine and a group that already has. They have both groups take the digit-span test twice, once before sleep and once 8 hrs later. The pretest measures the equivalency of both groups and the posttest will measure the relative effect of the independent variable.

Time-series design

A quasi-experimental design when there’s no appropriate nonequivalent control group. Researchers take multiple measurements of a single group before and after the independent variable manipulation is made.

• Ex: Instead of comparing the apnea patients who use a CPAP machine to others who don’t, a researcher tests the apnea patients with the digit-span test several times both before and after the CPAP machine.

Multiple time-series design

A quasi-experimental design where the researcher makes multiple observations of the experimental group and the nonequivalent control group. Essentially, two nonequivalent groups are tested and measured several times before and several times after treatment.

Interaction of selection with other threats to internal validity occur when

Extraneous variables affect one group but not the other.

Quasi-experimental design that may control threats to internal validity for history effects

Time series design, multiple time-series design

Quasi-experimental design that may control threats to internal validity for maturation effects and testing effects

Nonequivalent control group design, multiple time-series design

Quasi-experimental design that may control threats to internal validity for selection bias and interaction for selection

Time-series design

Quasi-experimental designs have ____ threats to their internal validity than preexperimental designs. However, as opposed to a true experiment, quasi-experiment designs are likely to have ______threats to internal validity in comparison.

Fewer; More

In quasi-experimental research, which is preferred, and why?:

A pre-test/post-test design, or a pre-test/post-test with a non-equivalent control group design?

A pre-test/post-test with a non-equivalent control group design, because this provides an additional point of comparison for establishing causation.

If the data from the non-equivalent control group are different than that of the experimental control group

Then it is still no guaranteed that there is an effect of the independent variable, because it is the relative change in scores that matters.

What is the logic behind running a time series design?

It can identify trends in the data before and after the independent variable manipulation.

Quasi-experimental designs are susceptible to carryover effects because

They involve measuring the same individuals multiple times.

True or False: Selection bias can be avoided when choosing participants for a quasi-experimental design

False

Factors

Another name for an independent or subject variable in the design

Factorial design

A type of design where more than one independent variable is tested simultaneously.

Main effect

The effect of each factor on the dependent measure. Calculated by comparing the marginal means. There’s one for each independent variable.

Level

The different conditions for each factor.

• Ex: If a factor is time of day, the different conditions may be morning or night.

Interaction effect

Whether or not each level of the independent variable is affected the same way by the other independent variable.

• Ex: In an experiment measuring how well participants score on either a reading or math test when given positive or negative feedback feedback, an interaction could be seen for the type of feedback and the type of test where perhaps type of feedback doesn’t affect reading test confidence levels, but has a considerable effect on confidence about the math test.

Marginal means

The means for each level of the independent variables. Used to calculate main effects.

Pilot study

A smaller version of a study with just a few participants to find and eliminate confounds.

What statistical procedure is most commonly used to analyze factorial designs?

Analysis of variance tests (ANOVA)

On a graph for a factorial design, the dependent variable is always assigned to which axis?

The y-axis

While there’s no particular rule for deciding which independent variable is presented on the x-axis of a factorial design, a general guidelines states:

The independent variable with the greatest number of levels is on the x-axis so there are fewer lines on the graph. However, if one variable is a continuous factor such as mood and the other is a categorical factor such as sex, then the continuous one goes ono the x-axis.

How will the lines on a graph of a factorial design look if there is no interaction?

They will be nearly parallel.

What are the types of two-factor designs?

• Mixed design (Split-plot design)

• Within subjects design

• Between subjects design

Mixed design (Split-plot design)

Where one factor is a between-groups factor and one is a within-subjects factor.

How do you find the number of conditions?

Multiply the number of levels by the number of independent variables

Higher-order factorial designs

Used for research studies with 3 or more independent variables.

Two-way interaction

Involves two independent variables.

• Ex: A research designs a project with 3 independent variables, A, B, C. ANOVA tells us that there’s 3 main effects, one for each variable. The 3 interactions are also analyzed: interactions between A & B, B & C, and A & C.

Correlations

Provide information about a relationship between variables, but can be used to predict future scores.

Regression

A statistical technique that’s used to predict and estimate the relationship between variables. Used to make predictions where correlational data does not already exist

• Ex: If there’s a correlation between a student’s interest level in a psychology-related career and taking an intro psychology course, then this technique can be used to predict what this student’s interest level in a career in psychology will be after taking an intro course.

Multiple correlation

Measures the relationship between multiple measures and a particular measure.

Multiple regression

A formula used to predict a particular measure using multiple correlation coefficents. Also provides information about the degree to which each individual measure contributes to the prediction of the participant of interest.

In factorial designs, we measure

The individual effects of each independent variable, as well as the interactions between them.

Using a factorial design generally requires __________ participants and _________ time than running separate studies for each independent variable.

Fewer; Less

In a factorial design, you will have a number of cell means equal to

The number of possible unique condition combinations

The number of unique conditions in a factorial design will increase

If the number of factors and/or the number of levels of factors are increased

Unique condition

Unique combination of factors. Equal to levels of all factors multiplied.

In a factorial design, you will have a marginal mean for

Each level of each factor

If, in a factorial design, different participants experience each possible combination of the independent variables, it is a

Between Groups Design

The number of interaction effects in a factorial design will increase

Only if the number of factors is increased

How is observational used?

It’s a starting point for research on a new topic. It can also be used to see if laboratory results also apply in the natural environment.

Two types of research in natural settings:

Field experiments and Observational studies

Field experiments

Controlled experiments involving random assignment and the manipulation of an independent variable conducted in a natural setting. Can provide greater external validity than what is obtained in laboratory research alone.

Observational studies

Investigations involving no manipulation of an independent variable. No independent variable is manipulated, and results may suggest relationships among variables, but casual conclusions cannot be drawn.

Naturalistic observation

Researchers unobtrusively observe behaviors in their natural setting; the investigator does nothing to interfere with the participants’ behavior. Focus of this research is often broad.

Systematic naturalistic observation

Unobtrusive observational study focused on particular behaviors in a particular setting

Desensitization

The researcher gradually moves closer to the participants until they can sit near or even among ඞ them; Often used in animal studies.

Habituation

When the researcher appears in the setting numerous times until their presence no longer appears to affect the participants’ behavior.

Participant observation

The researcher is an active participant in the situation.

Undisguised participant observation

Where the other participants are aware that the researcher is observing their behavior. Common among anthropologists, who join and work with a society while also observing it.

Disguised participant studies

Where the other participants are not aware that the researcher is observing their behavior.

Potential pitfalls for observational research

• Observer bias

• Lack of clear operational definitions may make it difficult to record observations

• Ethical concerns

Hawthorne effect

A term used in the social sciences, it is the effect of the observer on the behavior of the subjects.

In order to minimize the effect of their presence during observations, researchers must

Make observations surreptitiously (in a way that doesn’t draw a lot of attention)

Reactivity/Reactive measure

When a participant’s behavior changes with the awareness that they are being studied.

Observer bias

A person’s conscious or unconscious preferences affecting their perception of a situation.

Ways to avoid observer bias:

• Training research assistants to make observations for a study, without mentioning the expected results.

• Use of nonhuman observers

Ecological validity

The results can be readily generalized to real life.

• Naturalistic observations have the advantage of a high level of this.

Interobserver reliability

The degree to which a measurement procedure yields consistent results when used by different observers.

Formula for interobserver reliability

Scenario example: 2 observers watch children play and categorize the play activity as independent play, parallel play, and cooperative play. They watch the same child for 5 minutes and categorize the play activity every 30 seconds, for a total of 10 observations. However, they disagree on one occasion where one observer interprets play activity as parallel and the other interprets it as cooperative.

Two techniques for data collection for observational studies

Narrative records and Checklists.

Narrative records

Running records of behavior in a given situation. They can vary from being either very complete or rather sketchy. Can be created by audio or video recording a situation or by writing notes by hand.

Data reduction

Reducing the amount of information available to be more manageable.

• Ex: “The couple engaged in a warm embrace” becomes physical act of affection.

Checklist

Efficient ways of organizing record making. Works well when there is a focus on specific operationally defined behaviors.

Static checklist

Used to record characteristics that will not change during the course of the observations.

• Ex: Characteristics about the setting, how many people are present, gender and age, etc.

Active checklist

Used to record the presence or absence of specific behaviors and characteristics over time.

• Ex: The types of play behavior a child demonstrates on the playground over a period of time, or what a small group of people do in a library.

Disadvantages of checklists

They focus on a relatively small subset of behaviors or characteristics and ignore all others. Data collection by narrative records does not have this issue.

Advantages of checklists

Compared to narrative records, checklists don’t need to be reduced by coding and are already organized and can be more easily summarized.

Behavior sampling

Where a researcher observes subsets of a participant’s behavior at different times and/or in different situations.

Time sampling

The times at which observations will be made are chosen in an effort to obtain a representative sample of behaviors. May be done randomly or systematically.

Random time sampling

Each interval of time is equally likely to be chosen for when observations are to be made.

Systematic time sampling

When time intervals for observations are chosen purposefully.

When is time sampling appropriate?

When the behaviors of interest occur on a nearly continuous basis.

Event sampling

Used when behaviors occur less frequently. It is the random or systematic sampling of events that include the behavior of interest.

• Ex: A research interested in neighbors’ reactions to a building fire may need to sample among fires occurring in a city.

Systematic event sampling

Events are chosen in a purposeful manner. Risks having an unrepresentative sample as observations may be made only when it’s convenient for the researcher.

• Ex: Attending every fifth fire announced on the police radio to sample from.

Random event sampling

Events are chosen in a manner so that each event is equally likely to be chosen.

Situation sampling

Observations are made in different settings and circumstances. Can greatly enhance the generalizability of an observational study.

• Ex: A researcher interested in how children play together on a playground might make observations at a number of playgrounds in different neighborhoods.