Research Methods Midterm

Front 1

Why is science research so important?

Back 1

It can be used in many diff. areas and shows you whether data/ info is reliable

Front 2

Characteristics of science (7)

Back 2

Empirical (based on observation and data), Objective (observations, based on measurable unit), Self-correcting (nothing is final, subject to change), Progressive (always moving forward w/ new discoveries but repetition is imp.), Tentative (cautious about making claims/conclusions), Parsimonious (look for simplest explanation for your data), concerned with theory (if data doesn't fit theory you can speculate, if it does it supports the theory)

Front 3

Broad Goals of Science (2)

Back 3

Discover Regularities/patterns (describe behavior, discover laws/lawful relationships, identify causes of behavior), and Develop Theories (gives some explanatory mechanism to help science progress forward)

Front 4

J Steward Mills rules to determine cause and effect (3)

Back 4

1. Cause must precede effect in time
2. Must be relationship between two factors (method of agreement, method of difference, method of conmitance/correlation
3. All other explanations for effect must be ruled out

Front 5

Theory

Back 5

1. Set of statements that organize a large body of facts or laws into a single explanatory system
2. Includes at least 1 indirect concept/construct that's required to explain the relationship.
3. Testifiable
4. Can't be proven true

Front 6

Why are theories important? (5)

Back 6

1. Help to classify entities, processes, and causal relationships
2. Explain laws and past events
3. Predict new laws and future events
4. Help us to understand what causes events
5. Guide research in useful directions- what we should focus on vs. ignore

Front 7

Hypothesis

Back 7

Specific prediction of a relationship between 2 or more variables that can be tested.

Front 8

Where do hypotheses come from? (5)

Back 8

1. logically developed from theory
2. developed in order to resolve conflict between research results
3. From case studies or systematic observations
4. Failure of a theory to explain what you think you know from personal experience
5. Serendipity

Front 9

Scales of measurement (4)

Back 9

1. Nominal (used for categories, number has no significance)
2. Ordinal (used for ratings, numbers have no significance but rank, analyze using x2)
3. Interval (numbers have significance, = diff between numbers reflects = diff between measures but no meaningful zero). Analyze using t-test.
4. Ratio (has meaningful zero point as well as diff. between numbers. (analyze using t-test)

Front 10

What is reliability?

Back 10

Likelihood you will get the same score if you re-test your subjects. It's your measuring consistency- a minimum of 3 test replications are necessary for reliability

Front 11

What are kinds of reliability? (2)

Back 11

1. Test re-test
2. Internal consistency
A. Split-half reliability test
B. Cronbach's Alpha

Front 12

What factors increase reliability? (4)

Back 12

1. Number of items
2. High variation among individuals being tested
3. Clear instructions
4. Optimal testing situations

Front 13

What is validity?

Back 13

Whether or not you're measuring what you think you're measuring

Front 14

What is error variance?

Back 14

Variability in the dependent variable that isn't associated with the IV. Also known as random error.

Front 15

What are the measurements of measure validity? (4)

Back 15

1. Construct validity- does it measure the construct it's trying to measure and not others
A. Convergent
B. Discriminant
2. Content Validity- does it measure range of behavior represented by construct
3. Criterion validity- does it have similar scoring to other measures of the same construct
4. Face validity- does it appear to measure what it is actually measuring

Front 16

What is systematic error?

Back 16

Measurement error that's associated with consistent bias. Not a problem if it's same for the entire study. It can be a problem though if it's associated with one of the IVs

Front 17

What is internal consistency?

Back 17

Extent to which various items on a test are measures of the same thing.

Front 18

What are three components of Observed score?

Back 18

True score, systematic error, and random error

Front 19

What are 3 criteria for good data description?

Back 19

1. Accuracy or recording and organizing
2. Conciseness
3. Understandability (label accurately, it should be easily understood)

Front 20

What are two methods of data description?

Back 20

1. Numerical methods: data summary in the form of numbers
-measures of central tendency like mean/median/mode
- measures of variability (SD)
2. Graphical methods: presentation of data in a graph
-Frequency distributions
-Frequency histograms
-Frequency polygons
-Bar graph
-Scatterplot (or scattergram)

Front 21

What are the axes of a frequency distribution histogram?

Back 21

Frequency vs. score (DV)
This is a graph showing number of scores falling into specific bins or divisions of the variable. Bars touch each other.

Front 22

What histogram measurements fall into?

Back 22

Bins with specific ranges

Front 23

Frequency polygon is...?

Back 23

Frequency distribution in which frequencies are connected by straight lines.

Front 24

What is a sigmoidal curve on a graph?

Back 24

An s-shaped curve, generally fond in a cumulative frequency curve

Front 25

When is a scattergram used?

Back 25

When you want to show response of a number of subjects on two variables; visual display of correlational data. x and y axis can represent both IVs, both DVs. or an IV and a DV

Front 26

When are bar graphs used?

Back 26

When IV is categorical rather than quantitative

Front 27

What is the box and whisker plot based on?

Back 27

Median and percentiles, rather than on the mean and SD. Useful whenever data are skewed or in situations where median is appropriate for the data.

Front 28

What is validity of research?

Back 28

Extent to which research conclusion corresponds to actual state of the world.

Front 29

What are criteria of evaluating validity of research? (4)

Back 29

1. Internal Validity- can we reach causal conclusions about effect of IV on DV based on research
2. External validity- can we generalize to other populations
3. Construct validity- are the constructs we're interested in successfully operationalized in the research.
-Convergent validity- scores on one measure of construct are similar to same subjects' scores on diff. measure of same construct
-discriminant validity: measure of construct isn't related to measures of other unrelated constructs
predictive: measure of construct is useful is predicting results related to theoretical concept it's measuring
4. Statistical conclusion validity-extent to which research design is sufficiently precise/powerful to detect relationships among operationalized constructs

Front 30

What are threats to internal validity of research? (6)

Back 30

1. History
2. maturation
3. regression to the mean
4. mortality
5. selection
6. testing

Front 31

What is random assignment?

Back 31

process of assigning people or groups to experimental conditions such that each person has an equal chance of being assigned to a particular condition

Front 32

What are subject variables?

Back 32

Those differences between subjects that can't be controlled

Front 33

How can you maximize external validity?

Back 33

1. Choose random sample of subjects
2. Try to generalize on theoretical basis by speculating how subjects might differ on key variables from other people who didn't participate in study
3. Replicate findings in another setting w/ another pop.

Front 34

what is power?

Back 34

probability of detecting a relationship between IV and DV if such a relationship actually exists

Front 35

How can you increase power?

Back 35

1. Increase number of participants, power analysis tells you how many participants you need for significance
2. Decrease error variance by making all possible confounds into IVs

Front 36

What are demand characteristics?

Back 36

Cues within research context that bias subjects, resulting in responses that aren't valid, reliable reactions to variables of interest. Cues subjects get that allow them to make certain assumptions affecting their behavior.

Front 37

What are the four demand characteristics?

Back 37

1. Subject expectancies (desire to please researcher/look their best)
2. Experimenter expectancies- biases of how participants should behave
3. Environment- time of day, weather, etc can all affect behavior
4. Measurement procedures- can tip off subjects to what you're testing so you have to decrease face validity

Front 38

How can you minimize demand characteristics? (5)

Back 38

1. Deception (of subjects and/or of testers
2. Use unobtrusive measures (don't let subjects see you collecting data)
3. Keep experimenters blind to hypothesis
4. Keep experimenters blind to condition that subject is in (they shouldn't know who is control, who is experimental group member)
5. Standardize experimental procedures so all subjects received same instructions and stimulus materials

Front 39

What is control?

Back 39

The elimination of unintended, extraneous factors that might influence the behavior being studied

Front 40

What are strategies of control? (7)

Back 40

1. Random assignment
2. Subject as own control
3. Matching
4. Statistical control
5. Limit population you study
6. Introduce nuisance variables
7. Replication

Front 41

What kind of assignment is needed in experiments to draw causal conclusions?

Back 41

Random assignment

Front 42

What is within-subject design?

Back 42

Subjects are tested after exposure to each of experimental conditions

Front 43

What is between-subject design?

Back 43

Different subjects reactions to diff. IVs compared

Front 44

What are advantages for subject as own control? (2)

Back 44

1. fewer participants needed
2. Higher power

Front 45

What are disadvantages for subject as own control? (3)

Back 45

1. May arouse subjects suspicions about purpose of experiment
2. Treatment may produce long-lasting effect that carries over from one testing to the next
3. Cannot control for subject variables

Front 46

What are the bare essentials for a true experiment? (3)

Back 46

1. Random assignment
2. Treatment and comparison groups
3. Observations after treatment

Front 47

What is matching?

Back 47

Participants in one condition are matched with participants in other conditions on a third variable that's expected to affect the DV

Front 48

What are advantages of matched par designs? (3)

Back 48

1. Increased internal validity because you know other variables aren't confounding your results, you've already equated on imp. extraneous variables
2. More powerful design, better able to detect small differences between conditions
3. More power, probability of detecting a relationship between the IV and the DV if such a relationship actually exists

Front 49

What are limitations of matched-pairs designs? (4)

Back 49

1. Pretesting may increase demand characteristics- lots of pre-screening is required
2. May have to pretest many subjects to find enough subjects who match
3. Difficult to identify good matching variable
4. If matching variable isn't highly correlated with DV. it may weaken the design

Front 50

What is statistical control?

Back 50

using statistical procedures to equate subjects on extraneous variables, that is, to hold the effects of extraneous variables constant statistically. Done after study is completed and data collected when unexpected variable had unexpected effect

Front 51

What is an advantage of statistical control?

Back 51

Provides more precise test of hypothesis when subjects cannot be randomly assigned to conditions

Front 52

What is a disadvantage of statistical control?

Back 52

You can't reach causal conclusions and it's not always accurate. Doesn't tell you which of IV and extraneous variable is cause and which is effect.

Front 53

What is limiting the population?

Back 53

Limiting the type of participants you study such that they're all the same with respect to some third variable.

Front 54

What's an advantage of limiting the population?

Back 54

It reduces variability

Front 55

What's a disadvantage of limiting the population?

Back 55

It limits pool of participants and reduces generalizability of results

Front 56

What is a nuisance variable?

Back 56

an independent variable that was included in experimental design because it couldn't be eliminated. It's any variable that could be a potential confound.

Front 57

What is systematic replication?

Back 57

way of controlling for possible nuisance variables by altering slightly these variables in replications of these tests. Can increase internal validity and external validity (you can test diff. populations)

Front 58

What do t-tests measure?

Back 58

They test the null hypothesis that is used to compare means of 2 groups. Used for interval and ratio scales.

Front 59

What are chi-squared tests used for?

Back 59

Used for nominal and ordinal scales to evaluate whether the observed frequencies for a variable are adequately described by expected frequencies (does our data fit with past data)