Flashcards in Research Methods Deck (56):
is a process of developing an explanation of a question in the natural world by testing, investigating and collecting data that will either support or reject your idea.
Non Scientific Inquiry
Is one that does not use a systematic collection of evidence or one that tries to find answers to questions about things other than the external world
7 Steps of Psychological Research
Identify the Research Problem
- select partipants
- how many participants
- which data collection method
A sample is the group of people who take part in the investigation
The target population is the total group of individuals from which the sample might be drawn.
Confidentiality, Withdrawal Rights, Voluntary Participation, Debriefing, Deception, Informed Consent.
A general prediction about the direction of interactions between the IV, the DV and the population.
A testable prediction that explains exactly how the variables will be measured/manipulated + the population from which the sample has been drawn (workable/testable/repeatable)
4 Elements of a Operational Hypothesis
- testable prediction
- operationalised IV
- operationalised DV
Extraneous Variables + Types
Any variable other than the IV that causes change in results and an unwanted effect on experiment.
Confounding Variable and Effect
is an uncontrolled variable that has affect on DV.
If EV is not controlled then EV not IV had effects on results leading to wrong assumptions on what has effected the DV
Experimental vs Control Groups
An experimental group is the group in an experiment that receives the variable being tested (IV). The control group does not.
Participants know they are being studied and so they change.
Studies same group of people at different points in time.
- expensive/time consuming
- loose participants (is sample still representative)
- practice effects can distort findings
- cross generalisation problems ie children in differing cohorts may experience different at each point in life span than other children in earlier/later cohorts.
- no cohort effect
Cross Sectional Design
Studies cohorts who differ in age at the same time. Takes into account developmental differences amongst age and can compare children's abilities of different ages.
- Cohort effect: cannot be sure if its not due to developmental or participant differences ie cultural/environmental
- cannot determine whether children in youngest group will reason like older children that age
- easy, cheap
Longitudinal Sequential Design
Groups of participants are studied over time, and at each measurement a new group is added that was the same age as the first group when beginning the study.
Lets us look for changes in individuals across time (as a longitudinal) and age differences (as in cross sectional).
- complex and expensive
- no cohort effect
group of people at same age who have experienced same cultural conditions and environment
Independent Measures Design
Randomly sampled to control or experimental groups, with equal chance of being allocated to each.
- doesn't minimise potential differences in characteristics
- popular and easy to administer
Matched Participant Design
Involves pairing each participant based on a characteristic they share. Randomly allocated to control and experimental group.
- pretesting required
- seeks to eradicate participant differences by separating into 2 groups
Same participants exposed to control and experimental groups.
- creates order effects; occur when their is change in results due to the sequence in which 2 tasks are completed
- eliminates partipant differences
Involves dividing the groups of participants in half and arranging them in order of the conditions so that each condition occurs equally as often in each position e.g half exposed to control then experimental condition and the other half experimental then controlled
Is the information or observable facts that a psychologist systematically collects in a experiment, studies and investigations.
Collected through observations of behaviour or information based on participants self reports. Can be biased and difficult to statistically analyse.
Collected under controlled conditions, are easily measured and compared with other data. Often numerical and can be statistically analysed.
Describes changes in the quality of behaviour and often expressed in words. Difficult to statistically analyse as responses take a wide variety of forms.
Describes changes in numerical or categorical form. Can be statistically analysed,easily measured and compared with other data.
Non Experimental Designs Examples
Case studies, Correlational, Interviews, Scales, Observational, Self Reports, Brain Imaging, Archival Research
Provide detailed knowledge about single case or small number of related cases by studying an individual or small group. Uses a variety of data collection processes.
- not generalisable/expensive
- detailed/specific data
- can offer insight into cases otherwise unethical
Has open ended questions and focussed on particular area of interest
- interviewer bias
- comparable data
provide scale on which individuals standing on an issue can be measured. Quantitative subjective data can make inferences/stats.
- limited scales
- easy to understand
Involves individual observing another individual or a group of people in natural environment (no interference or observation) and recording behaviours.
- Observer bias/ Experimenter Effect
- Expensive, Time Consuming
- No scientific control of extraneous variables
- Hawthorne Effect
- forms a hypothesis that guides future studies
- qualitative subjective data
Aware of observer and so acts different (side effect of artificiality).
Experimenter effect + strategy to fix
Experimenter treats one individual different to others (less interaction thus reduced). Controlled by standardised procedure/single and double blind side.
A phenomenon that occurs when a person believes he or she is receiving real treatment and reports an improvement in his or her condition.
Single Blind Procedure
Experimenters are aware of which subjects are receiving the treatment or independent variable, but the participants of the study are not - a technique for eliminating subjective bias
Double Blind Procedure
Both the experimenters and participants are unaware of who is receiving the independent variable
Individuals asked to comment on own thoughts/emotions etc by answering series of questions on a particular topic
- partipant variables
- subjective data that cannot be overtly seen
- difficult to compare
- lots of data and flexible
Correlational Studies Definition and 3 types
Looks at relationships between different variables and are usually used in cross sectional studies.
- naturalistic observation
- survey methods
- archival research
Limitations and Strengths of Correlational Studies
- allows for researchers to determine strength/direction of relationship so later studies can narrow down findings
- only uncover relationships not provide a conclusive reason for relationships
Reliability and Types (Internal/External)
the extent to which an assessment tool measures what it is supposed to measure consistently each time it is used
- Internal Consistency
- Test- Retest Reliability
- Parallel forms of reliability
- Inter-rater reliability
Validity and Types (Internal/External)
The extent to which an assessment tool actually measures what it is designed to measure
- Content Validity
- Construct Validity
- External Validity
- Criterion-related Validity
is research involving primary sources held in an archives ie reports/documents etc
- dates may be missing
- unreliable research
- cannot change anything
- analyses studies
- less expensive
Allows us to make inferences, form conclusions, generalise findings and determine validity. Can determine whether differences in results due to manipulation of IV or to chance (P Values, Probability, Correlations)
All items contribute equally to what is being measured
Test Re-Test Reliability
Same results in same condition by same person at a different time
Inter Rater Reliability
Same results when assessment tools used by different administrators
Parallel Forms Reliability
Do two tests developed from the same content produce the same results
Are research tools assessing what they are meant to
Are research tools assessing the content/theories they are meant to
Can finding be generalised to wider population
Criteria Related Validity
Consistent with other research
used to summarise, organise an describe data obtained. Allows data to be easily interpreted (percentages, graphs, measures of central tendency)
Measures of Central Tendency
Mean, Mode, Median, Range, Variance, SD)
- Refers to the significance of the difference between two scores
- whether we can attribute the results to the IV or to chance alone
- Demonstrates the cause and effect relationship
is the level of probability that the results (differences) are due to chance alone, and determines the statistical significance = 0.05%,p