Research Methods A.1 Flashcards
What is a hypothesis?
A testable prediction derived from theories
What is reactivity?
The adjustment of behaviour to knowledge about the study
Examples of non-experimental methods:
naturalistic observations
case studies
surveys
correlational research
Qualities of a naturalistic observation:
Mainly categorisiation, observer stays in the background
Problems and solutions of naturalistic observations:
Problem: the reliability of categorisation
Solution: compare notes to for inter-observer reliability
Problem: Reactivity
Solution: Participant observation
What are case studies?
observations of a single person, often used in clinical settings
Problems and solution on case studies:
problems: lack of generalisation, lack of reproducibility, lack of cause and effect, which all means limited application
What are examples of surveys?
Questionares, interviews and diary studies
What are structured interviews?
a type of survey, with fixed questions in a fixed order, usually either multiple choice or ratings
Strengths and limitations of structured interviews:
S: easily quantifiable, can compares answers, can ensure all topics will be covered
L: Rigid structure, not adaptable to ppts, surface info only
Strengths and limitations of unstructured interviews:
S: more in depth info, relevant to each specific ppt
L: lack of generalisability, analysis is time consuming and answers are hard to compare
Strengths and limitations of descriptive research overall:
S: sometimes the only practical or ethical way, real life studies have high ecological validity
L: researcher bias (it’s down to their subjective questions, observations and analysis) reactivity, lack of cause and effect
What is correlational research?
A way of determining the direction and strength of a relationship between two variables without manipulating any variables
What is an experimental method?
a manipulation of one or more variables, where we attempt to determine the effect of this manipulation on the other variable
How can we deal with nuisance variables?
- turn it into a control variable if it varies for levels of the IV
- random allocation, or counterbalancing (ABBA) if it varies across ppts
Strengths of experimental methods:
-relatively strong test of causality
-possibility of a variety of manipulative controls
Limitations of experimental conditions?
-unnatural settings and tasks
- reactivity
ethical limitations (e.g. deception)
-some phenomena cannot be studied under controlled conditions (e.g. social interactions
What is a within subjects design?
ppts do all conditions
What is a between subject design?
ppts do different conditions (ppts differ BETWEEN conditions)
What is a Quasi experiment?
one of IV’s are selected, but not manipulated. Usually a characteristic of ppts.
Why are quasi experiments not truly experimental?
because we are technically not manipulating anything
What is EEG and what are strengths and limitations?
-electrodes being placed on the scalp
- strengths: good temporal resolution
- limitations: poor spatial resolution (only picks up activity on the surface of the brain)
What is an fMRI and what are the strengths and limitations?
-measures amount of oxygenated blood in areas of the brain
- strengths: good spatial resolution
- limitations: poor temporal resolution (5s delay)
What are nominal levels of measurement?
numbers are merely labels, no relationships between size of numbers and attribute measured
What are ordinal levels of measurement?
order of numbers = order of size of attribute measured. Only relative ranking… distances between scores vary e.g. IQ scores
What are interval levels of measurement?
equal intervals on the scales, equal measures of magnitude e.g. Degrees Celsius. Zero isn’t relevant, and we can have minus values
What are ratio levels of measurement?
All properties of interval scale, but we do have an absolute zero, which means absolute absence e.g. reaction times
What level data can you not use the median?
nominal level
What levels of data can you not use the mean?
Nominal and ordinal level
What levels of data are bar graphs used for?
nominal and ordinal data
What type of graphs are usually used to display correlations?
scatterplots
What are the two correlation coefficients we use?
Pearson’s and Spearman’s (always lie between -1 and 1)
Which coefficient is calculated directly from raw scores, and is suitable for interval or ratio level data?
Pearson’s
Which coefficient is calculated from the ranking of raw scores, and is suitable for ordinal level data?
spearman’s
When is it best to use density curves?
when we are dealing with a large sample size and are looking for an overall pattern
In a normal distribution, what is the order of mean, median and mode?
all the same
In a positively skewed density curve, what is the order of mean, median and mode?
mode, median, mean
In a negatively skewed density curve, what is the order of the mean, median and mode?
mean, median, mode
If we had to calculate ‘what percentage of people scored over 50?’ from a density curve, how would we do it?
find 50, then calculate the area of the curve from after that point, then times it by 100 to get a percentage
What is a standard Z score and what is the formula?
- the number of standard deviations that a single value in the data set deviates from the mean
- formula is deviation of _ from mean divided by standard deviation
How do you calculate degrees of freedom?
(no. rows - 1) x (no. columns - 1)
what data is chi-squared usually performed on and how is the data presented?
usually nominal/ categorical data, data is presented in a contingency table
In a density curve what is the value of the total area under the curve?
1 (100%)
In chi-squared what are the observed frequencies and the expected frequencies?
the number or ppts measured in individual categories. the expected frequencies are the frequencies predicted by the null hypothesis. The bigger the gap between these values, the bigger the chi-squared value
In what chi-squared test do we compares these observed values and expected values?
chi-squared test for goodness of fit
if i had two sets of data e.g. scores in RMA exam and scores in RMB exam, and i wanted to ask the question of in which test getting a score of 50 is better, what calculation would i use?
calculate the Z scores: do 50- mean over SD for both groups, then whichever has the value that is closest to the mean is the one in which getting a score of 5o is better
What is standardising?
adjusting values measured on a unique scale to a more common scale, usually by making the mean 0 and the standard deviation 1
what does chi-squared test for goodness of fit tell us?
tells us about the PROPORTIONS of a population distribution e.g. is there gender bias in the psychology department? used to compare different variables of ONE level