Search
Browse
Create
Log in
Sign up
Log in
Sign up
Upgrade to remove ads
Only $2.99/month
317 exam 4
STUDY
Flashcards
Learn
Write
Spell
Test
PLAY
Match
Gravity
Terms in this set (44)
Difference between positive and negative correlations
-Positive: as one variable increases, the other variable increases
-Negative: as one variable increases, the other variable decreases
Magnitude of correlations
- Expresses the strength of the relationship
- Unrelated to the signs (+ or -)
- +/- .10 = small
- +/- .30 = medium
- +/- .50 = large
Scatter plots and what they represent in terms of correlations
- Graph of participants scores on two variables
- Visualize strength of the correlation
Meaning of .00 correlation and how it relates to curvilinear relationships
- Indicates no linear relationship OR a curvilinear relationship
Coefficient of determination and how to calculate it
- Coefficient of determination (r^2): Proportion of variance in one variable accounted for by the other variable (systematic)
- square r to find c.o.d.
What a p-value means in terms of correlations
- Index of how likely the true correlation in the population is different from .00
- If less than 5% probability of the size of the smaple's correlation is due to chance (p<.05), rules are considered "statistically significant" and reflect the larger population
Three things that influence statistical significance
- Sample size: more participants = more significant
- Correlation magnitude: larger r = more significant
- How careful we want to be in our conclusion that the correlation in the population is not .00 = move alpha from .05 to .01 or .001
Directional v Non-Directional hypotheses
- Directional: Predicts the direction of the correlation (i.e. positive or negative)
--- More powerful and can reach statistical significance with fewer participants
- Non-directional: Predicts that two variables will be correlated but does not specify whether r will be positive or negative
The effects of a restricted range on correlations
- Participants' scores are confined to a narrow range of the possible scores on a measure
- Artificially lowers correlations below what they would be if the full range of the scores was present
- More serious distortion if curvilinear relationship
On-line vs Off-line outliers and what they do to correlations
- On-line outlier: fall in the same pattern as the data and artificially inflates r
- Off-line outlier: Fall outside the pattern of the data and artificially deflate r
Relationship of scale reliability to correlation size
- The less reliable a measure, the lower its correlation with other measures
- A correlation between two scales can never be higher than either scale's Cronbach's Alpha
CORRELATION DOES NOT MEAN CAUSATION!
CORRELATION DOES NOT MEAN CAUSATION
Three criteria for inferring causality
- Covariation: changes in one variable are associated with changes in the other variable
- Order in time: the presumed causal variable must occur before the other variable
- Removal of all other variables: all other variables that may affect the relationship between the two variables are controlled or eliminated
--Correlation is criteria but cannot determine causality alone
Spurious correlations
- Correlation between two variables due to their relation to other variables
Partial correlations
- Correlation between two variables with the influence of one or more other variables statistically removed
Different types of correlations
- Spearman rank-order correlation: used when one or both variables are on an ordinal scale
-- Placing in a marathon and age
- Point-biserial correlation: used when only one variable is dichotomous
-- Gender and height
- Phi coefficient: used when both variables are dichotomous
-- Gender and virginity
Three components of a well-designed study
- Vary at least one independent variable to assess its effects on participants' responses
- Assign participants to experimental conditions to ensure their initial equivalence
- Control extraneous variables that might influence the outcome
Difference between an independent variable and dependent variable
- Independent variable: Variable that the researcher manipulates to assess its effects on the participants behavior
- Dependent variable: response being measured in a study that is theoretically affected by the IV
Levels of an independent variable
- Must have 2 or more levels (values of the IV)
- Levels can be quantitative or qualitative
Three types of independent variables
- Environmental manipulation: modifications of the participant's physical or social environment
- Instructional manipulation: vary the IV through the verbal instructions that participants receive
- Invasive manipulation: create physical changes in participant's body through surgery or drugs
Experimental v. control groups
- Experimental groups: participants who receive a non-zero level of the IV
- Control group: participants who receive a zero level of the IV (absence of the variable of interest)
Purpose of a pilot test
- Preliminary study ensuring the levels of the IV are different enough to detect an effect
Purpose of manipulation checks
- Questions that determine whether the IV was manipulated successfully
Participants (or subjects) variables and how they are different from traditional independent variables
- Personal characteristics of research participants, such as age, gender, self- esteem, weight, etc.
- Not technically IV because they are not manipulated
Difference between simple and matched random assignment
- Simple: participants are placed in experimental conditions such that every participant has an equal probability of being in any condition
-- Great equalizer
-- Used to make conditions roughly equivalent at the start of the study
- Matched: participants are matched into groups according to scores on a particular variable, and then assigned randomly to conditions
Difference between between-subjects and repeated-measures designs, as well as the strengths and drawbacks of both
- Between-subjects: each participant serves in only one experimental condition
- Repeated- measures: experiment in which each participant serves in all conditions of the experiment
-- Eliminates the need for random assignment
-- More powerful than between subjects because every participant serves in every condition
--- Ability to detect effects of the IV
--- Requires fewer participants
-- No pre-existing differences in participants in different conditions: they are the same
-- Order effects, counter balancing, and carryover effects
Treatment variance v. error variance v. confound variance
- Treatment: primary variance, due to the IV
- Error: within-groups variance, variability among scores within a particular experimental condition, unrelated to the IVs under investigation
- Confound: secondary variance, due to the extraneous variables that differ systematically between experimental groups
Types of order effects
- Practice effects: responses are effected by completing the dependent variable many times
- Fatigue effects: participants become tired or bored
- Sensitization: participants become suspicious of the hypothesis as the experiment progresses
Purpose of counterbalancing
- Presenting the levels of the independent variables in different order to different participants
- Protect against order effect
- Use all orders if possible
Carryover effects
- Occur in within-subjects designs when effects of one condition are still present when the participant is tested in another condition
Internal v. external validity and the experimenter's dilemma
- Internal: degree to which a researcher draws accurate conclusions about the effects of the IV on the DV
- External: degree to which the results obtained in one study can be replicated or generalized to other samples, research settings, and procedures
- Experimenter's dilemma: more tightly controlled = stronger internal validity but less external validity, tight control makes experiment less like real world, experimenters almost always opt or internal over external validity
Threats to internal reliability
- Biased assignment of participants to conditions, effects due to initially non-equivalent groups rather than the IV
- Differential attrition: participants drop out of conditions at different rates, making experimental groups no longer equivalent
- Pretest sensitization: completing a pretest leads participants to react differently to the IV than if they had not been pretested, pretest measure of gender stereotypes alerts participants the study will involve gender stereotypes
- History effects: extraneous events outside of the research setting affect responses, interact with experimental condition
- Miscellaneous design confound: something other than the IV differs systematically between the experiment conditions
Confounding
- Something other than the IV that differs in a systematic way
Placebo effect
- Physiological or psychological change that occurs as a result of the belief that an effect will occur
Why differential attrition is a problem
- Participants drop out of conditions at different rates, making experimental groups no longer equivalent
Possible sources of error variance
- Individual differences: pre-existing differences between people
- Transient states: at the time of the experiment, participants differ in how they feel (mood, health, fatigue, interest, etc.)
- Environmental factors: differences in the conditions under which the study is conducted (noise, time of day, temperature, etc.)
- Differential treatment: treating different participants in different ways
- Measurement error: unreliable measures contribute to error variance
Ways to eliminate demand characteristics
- Aspects of a study indicate to the participant how they should respond
- Use double-blind procedures: either participants nor experimenters know which condition participants are in
Structure of one-way designs
- Experiment in which only one IV is manipulated
Difference between pre-test post- test, and post-test only designs, strengths and drawbacks of each
- Pre-test post-test: DV is measured twice, before and after the experimental measure
-- Can determine that the conditions did not differ on the DV at the beginning of the experiment
-- Can see how much the IV changed behavior from pretest to posttest
-- More powerful that posttest only
-- Not necessary: posttest-only designs can determine whether IV affected DV
- Post-test only design: DV is measured only after the manipulation of the IV
Be able to identify different types of factorial designs, and know what each number in factorial nomenclature means
- Two or more IVs are manipulated
- Factorial nomenclature
-- 2 X 2 factorial - 2 IV's, each with 2 levels
-- 3 X 3 factorial - 3 IV's, each with 3 levels
-- 2 x 2 x 3 - 3 IV's, 2 with 2 levels, 1 with 3
Difference between main effects and interactions
- Main effects: Effect of an IV while ignoring the effects of all other IV's
-- Factorial design will have as many main effects as there are IV's
- Interactions: effect of one IV differs across the levels of another IV
Two-way and three-way interactions and how many possible main effects and interactions are in each
- Three-way designs examine the main effects of three IV's
- Three two way interactions: A x B (ignoring C), A x C (ignoring B), and B x C (ignoring A)
- The three-way interaction of A x B x C
Difference between median-split and extreme-groups procedures and their drawbacks
- Median-split: participants who score below the median are classified as low, and participants scoring above the median are classified as high
- Extreme-group: use only participants who score very high or low
Causality and manipulated vs participant variables in an experiment
- If the manipulated IV affects the DV, can conclude that the IV caused the effects
-- BUT, participant variables are not manipulated, so cannot infer causation
-- If a participant variable is involved in an interaction, it moderates participants' reactions to the IV
YOU MIGHT ALSO LIKE...
Research Methods Test 4
96 terms
Psych 12 Midterm 3
73 terms
psych test 3
48 terms
Research Methods Test Three
60 terms
OTHER SETS BY THIS CREATOR
arth 104 exam 4
77 terms
Art History 104 Exam 3
74 terms
317 Exam 3
49 terms
Art History 104 Exam 2
76 terms