expe lesson 1.2-1.4

natural settings,

natural settings,

Study behaviors in— explore unique or rare occurrences, or sample personal information.

High in internal Validity

• Degree to which a research design allows us to make casual statements

High in External Validity

•

• Generalizability or applicably to people and situations outside the research setting

naturall to controlling 

the degree of manipulation of antecedent  conditions varies from low to high (letting whatever happens — to — it

responses

the degree of imposition of units

• the extent to which the researcher limits the — a subject may

contribute to the data

the degree of manipulation of antecedent  conditions

varies from low to high (letting whatever happens naturally to controlling it

the degree of imposition of units

• the extent to which the researcher limits the responses a subject may

contribute to the data

Phenomenology

• the description of an individual's immediate experience;

Case study

• descriptive record of a single individual's experiences or behaviors or both kept by an outside observer

Field studies

• are nonexperimental approaches used in the field or in real-life settings.

Archival study

• a descriptive method in which already existing records are reexamined for a new purpose.

Qualitative research

• relies on words rather than numbers; study phenomena that are contextual

Survey research

• useful way of obtaining information about people’s opinions, attitudes, preferences, and behaviors simply by asking

Step 1

: map out your research objectives, making them as specific as possible;

Step 2

design the survey items; decide how you are going to address the imposition of units

Step 3

 once the questions have been designed they need to be pretested

Levels of Measurement

• the kind of scale used to measure a response for a closed question; different statistical tests are required for different levels of measurement; four kinds of scales:

Nominal

• classifies response items into two or more distinct categories on the basis of some common feature.

Ordinal

1.

• Rank ordering of response items.

Interval

• measures magnitude or quantitative size using measures with equal intervals between

the values and has no true zero point.

Ratio

• Equal intervals between all values and a true zero point.

Continuous Dimension

when variables lend themselves to different levels of measurement; traits, attitudes, and preferences are all continuous

Semantic Differential

q evaluating variable on a number of dimensions; two adjectives

Likert

q present a positively worded statement with a negatively worded statement

Response Styles

• are tendencies to respond to questions or test items in specific ways, regardless

of the content

Reliability

q the extent to which the survey is consistent and repeatable

Validity

q the extent to which a survey actually measures the intended topic

Sampling

q

q deciding who or what the subjects will be and, then, selecting them

Population

q

q all people, animals, or objects that have at least one characteristic in common; ex: all undergraduate student

Sample of subjects

q

q a group that is a subset of the population of interest

Representativeness

q

how closely the sample mirrors the large population

Probability Sampling

q

• involves selecting subjects in such a way that the odds of their being in the study are known or can be calculated.

Simple random sampling

1.

• a portion of the whole population is selected in an unbiased way

Systematic random sampling

• all members of the population are known and can be listed in an unbiased way

Stratified random sampling

1.

• obtained by randomly sampling from people in each subgroup in the same

proportions as they exist in the population

Cluster sampling

1.

• sample entire clusters or naturally occurring groups that exist within the population

Correlational Designs

• used to establish relationships among preexisting behaviors and can be used to predict one set of behaviors from others

Quasi-Experimental Design

q Called “natural experiment”, but they

lack one or more of the essential elements, such as manipulation of antecedents or random assignment to treatment conditions.

Correlational Study

designed to determine the correlation, or degree of relationship, between two traits, behaviors, or events.

Simple Correlation

•

• Relationships between pairs of scores from each subject

Scatterplot

•

• visual representations of the scores belonging to each subject in the study

Regression lines

•

• direction of line corresponds to the direction of the relationship

Coefficient Of Determination (R2)

• estimates the amount of variability in scores on one variable that can be explained by the other variable; how much one variable can explain the variability in scores of the other variable

Path Analysis

• method that can be used when subjects are measured on several related behaviors

Cross-lagged Panel Design

•

• uses relationships measured over

time to suggest the causal path

Ex Post Facto Studies

q

• Explores characteristics, behaviors, or effects of naturally occurring events in preexisting groups of subjects.

Nonequivalent Group Designs

q

• Compares the effects of different treatment conditions on preexisting groups of subjects.

Longitudinal Design

q

• Investigates changes across time by measuring behavior of same group of subjects at different points in time.

Cross-sectional Studies

q

• Investigates changes across time by comparing groups of subjects already at different stages at a single point in time.

Pretest/Posttest Design

q

Explores the effects of an event (or treatment) by comparing behavior before and after the event (or treatment).

Solomon 4-group Design

q

number of comparison groups are needed --> a nonequivalent control group (took both pre and post tests but was not exposed to the "treatment", a group that received the treatment and took only the posttest, and a posttest-only group