Chapter 2 Research Flashcards
Types of psychological research
Applied and basic research
Hindsight bias
Tendency for ppl to think they knew the conclusion of an experiment all along after already hearing about it
Applied research
Research whose conclusions have clear practical applications (what methods students learn best by)
Basic research
Research whose conclusions are of interest but don’t have immediate practical applications
Hypothesis
Expresses relationship bw 2 variables
Independent Variable
The factor that is purposely varied or manipulated in an experiment; what is changed
Variables
Things that are different across the different participants of the experiment
Dependent variable
The factor affected by the independent variable
Theory
Explanation that Aims to explain a phenomenon
Operational definition
Explanation of how a variable will be measured (ex: how do you measure aggressiveness?)
Valid research
How Accurate it is
Reliable research
If the experiment can be replicated and get the same consistent results
Participants/subjects
Individuals being researched in a study
Sampling
Process where participants are chosen
Sample
The complete group of participants
How to do sampling
- identify the population
- do a random selection so that it can be representative of a larger population
- the bigger the sample, the more representative it is of the population (but it’s a lot more time and money)
Population
Anyone that could possibly be selected in the sample (ex: student body of a college)
Random selection
- Every member of the population has an equal chance of being selected
- increases chances that the sample can be generalized to a larger population
- best done using a computer or picking names out of a hat to be truly random
Stratified sampling
If you think a certain factor or variable of the participant would affect their results, get the different proportionate variations of this factor in your sample according to the population it represents
Types of experiments
Laboratory and field experiments
Laboratory experiments
In a controlled lab environment
Field experiments
Conducted out of a lab; more realistic than lab
Experiment
Preferred method of psychological research, only research method that can show cause-effect relationship
Confounding variable
A factor other than the independent variable that may affect the dependent variable thus making experiment inaccurate
- can be controlled
2 types:
Participant-relevant (age, sex, iq)
Situation-relevant (time of day, environment)
Assignment
Process where participants are put in either experimental or control group
Group matching
Dividing a certain factor within the sample equally among the experimental and control groups
Random assignment
- diminishes chances of participant-relevant confounding variables
- everyone in the sample has an equal chance of being in experimental or control group
Experiment bias
Inadvertent tendency for researchers to treat experimental and control groups differently so that their hypothesis can be supported
- type of situational confounding variable
- can be removed using double blind procedure
Double blind procedure
When neither participants or researcher can affect research outcome, participant doesn’t know which group they’re in, researcher doesn’t know which group participant is
Single blind procedure
Only the participants don’t know which group they are assigned to, minimizes effect of demand characteristics and participant bias
Response or participant bias
Tendency for subjects to behave in certain ways different than they would normally when they know what group they are in (experimental ir control)
Social desirability
Kind of response bias when participants act a certain way that reflects well on them when they know what group they are in (control/experimental)
Demand characteristics
Cues to the participants about the purpose of the study
Experimental group
The group that is exposed to varying amounts of the independent variable
Control group
Group with no presence of the independent variable, used to compare if the dependent variable was really caused by the independent variable by altogether removing it
Hawthorne effect
How selecting a group of ppl to be in an experiment and be observed is already affecting how naturally they will be acting
Placebo method
When psychologists are testing a drug, the control group is told they will be taking a drug when really it does nothing while the experimental group takes the working pill. This will separate the actual physiological effects of the drug from the psychological effects.
Method of control
Counterbalancing
When using same participants for control and experimental group, split them in half. Send one group to be the experimental and then switch to the control. Have the other group start at control, then to experimental. This will prevent the order effects
Order effects
When the same subjects are part of the control and experimental group, the skewing of results
Correlation
Expresses relationship bw 2 variables wo saying the cause of why they are related
- can be strong or weak
- can be positive or neg
- can be graphed with a scatter plot
Positive correlation
Presence of one thing predicts presence of another
Negative correlation
Presence of one thing predicts the absence of another
Ex post facto study
Seeking to control all aspects of a research process
Survey method
- There are 2 variables but they cannot be manipulated or made into a cause-effect relationship like an experiment would. Aka, it is not an independent and dependent variable (only shows correlation)
- no control over participant related confounding variables
- situation relevant confounding variables can be controlled but it is rarely done (too inconvenient)
- low response rates that are unlikely to be representative of a larger pop
Naturalistic observation
Observing participants to get a more realistic view of behavior; no interaction means that control is sacrificed
Case study
Gets a full detailed description of a person or small group of ppl; cannot be generalized to a large pop
Types of research methods
- experiment
- survey
- naturalistic observation
- case study
Descriptive statistics
Set of data
Frequency distribution
A graph measuring the varying frequencies of different factors
- Histograms
- frequency polygons
Frequency polygons
Line graphs of frequency distributions
Histograms
Bar graphs of frequency distributions
Measure of central tendency
Marks the center of distribution in a graph
- mean
- median
- mode
Best not to use mean when there are outliers in the data (median is best)
Mode
Most frequently seen score
Graphs may be bimodal (2 modes)
Outlier
Extreme score in data that deviates from others
- skews distributions
Data skews
Caused by outliers
- positively skewed- outliers that are higher than the others, mean is higher than median
- negatively skewed- outliers that are lower than others, mean is lower than median
Types of statistical measures
Measure of central tendency
Measures of variability
Measures of variability
the diversity of the distribution
- range
- variance
- standard deviation
Standard deviation
Avg distance from the mean, the higher it is the more spread the distribution
- subtract the mean from one of the numbers of the set
- raise the expression to the power of 2
- do this for all other numbers, add them together separating with parenthesis
- put a huge square root umbrella over it all
Variance
Avg distance from mean, the higher it is the more spread the distribution
- subtract the mean from one of the numbers of the set
- raise the expression to the power of 2
- do this for all other numbers, add them together separating with parenthesis
Z score
Measures a score’s distance from the mean in units of standard deviation
Ex: score of 72% on a test with a Mean score of 80% and a standard deviation of 8, z score is -1
Normal curve
Theoretical symmetrical Bell shaped curve predicting data score distribution
- 68% of results fall within one standard deviation
- 95% fall within 2
- almost 99% fall within 3
Percentiles
Indicate the average distance of a score from 0
Ex: a person who scores in the 50th percentile in a test scored better than 50% of people who took the test. The person’s z score would be 0.
Correlation coefficient
Measures strength of a correlation
- range from -1 to +1
- 0 indicates no correlation at all
Scatter plot
Used to graph correlations
- closer the points come to falling in a straight line the stronger the correlation
- line of best fit or regression line is the general direction the point goes
- positive- slopes upward
- neg- slopes downward.
Inferential statistics
- Sees whether findings from a sample can be applied to the larger pop
- takes into account magnitude of difference between control and experimental group, also sample size
- many inferential tests exist- all yield a ‘p’ value
- p value is the percentage probability that the results were just gotten by chance, no such thing as 0 p value
- p value of .6 means 60% chance the results were coincidental
- cutoff for statistically significant results is .05 or 5%
Sampling error
Extent to which a sample differs from the pop it’s trying to represent
Ethical guidelines in research
- any academic research must first go through the institutional review board (IRB)
- must follow ethical guidelines by the APA when experimenting
Apa ethical animal guidelines
- Clear scientific purpose
- must be best suited to answer the question at hand
- care and house animals humanely
- acquire animals legally, wild animals trapped humanely
- procedures that inflict least amount of suffering
Apa ethical human guidelines
- only voluntary participation
- informed consent, a little bit of deception allowed in what the nature of the study is but not enough to invalidate the general informed consent
- anonymity(no personal info given) or confidentiality(any personal info given is not released to public)
- participants can’t be placed at significant or long term mental/physical harm (some institutions allow more than others)
- participants have to be debriefed on the purpose of the study, results, etc after esp if deceit is involved
Extraneous variable
Factor other than the independent variable that can effect the dependent variable making experiment inaccurate
- cannot be controlled
