Data Handing and Analysis Flashcards
define correlation
used to investigate an association betwen two variables
* measured not manipulated
* no cause and effect relationship found
correlation coefficients are calcualted
* negative = less than 0
* postive = more than 0
* zero = equal to 0
curvilinear relationship - one variable increases so does the toher but only utp a certian point- one variable continues to increase the other begins to decrease
evaluation of correlation
- used as starting points to asses patterns between co variables before conducting an study
- quick and economical
- secondary data can be used = less time consuming
- difficult to establish cause and effect- only association is found
- third variable problem - researcher is unware of that is responsible for relationship between variables
- miusesed or minsinterpreated - presented as causation
define and evaluate qualitative data
displayed in words and non numerical
* richness and depth
* allows particpants to develop their opinions = greater external validity
* meaningful insight
- difficult to analuse
- difficult to make comparisons with other data
- researcher bias - conclusions rely on subjective interpretations
* interviews
* open ended questions
* observations
define and evaluate quantitative data
displayed numerically
* analysed statistically - converted into graphs
* easier to make comparisons
- lack of depth in detail
- no meaningful insight
- particiapnts not able to develop opions= low external validity
*** closed rating scales
* closed questions/ surverys **
define and evaluate primary data
information is obtained first hand by researcher
* targets excat information researcher needs = fits aims and objectives
- requires time and effort and can be expensive
define and evaluate secondary data
information is collected by someone else other than the researcher and is used for investigationn
* less expensive
* requires minmal effort
- data can be outdated or incomplete
- data may be unrelaible - researcher not there when study was done
define and evaluate meta analysis
researcher combines results from many different studies and uses all dara to form an overall view of the subject
- more generalisability due to larger amount of data studied
- able to view evidence with more confiendence
- publication bias such as file drawer problem- researchre may intentionally does not publish all data from studies and chooses to leave out negative results = false representation
define and evaluate mean
- makes use of all values
- good for interval data
- infuleunced by outliers= can be unrepresentative
define and evaluate median
- not affected by extreme scores
- goof for ordinal data
- not as sensitive as mean - does not use all data
define and evaluate mode
- useful for norminal data
- not useful when there are several modes
measures of dispersion
define and evaluate range
- easy to calculate
- affected by extreme values
- does not use all data
measures of dispersion
define and evaluate standard deviation
- precise measure where all data values are taken into accoutn
- affected by extreme values
bar charts
discrete data - data has been divided into categories
histograms + line graphs
continous data
scattergrams
show assocation rather than differences
define normal distribution
symmertical pattern of data
define skewed distribution
spready of frequency data that is not symmetrical
postive skew: concentrated on right
negative skew: concentrated on left
define levels of measurement
quantitative data
norminal: in form of categories - discrete data
- does not enbale sensitive analysis- does not yield a numerical result for each participant
* mode (measure of central tendency)
* N/a (measure of dispersion)
ordinal: represented in ranking form
- no equal intervals
- lacks precision as is based on the subjective opion of people
* median
* range
interval: based on numerical scales which include equal units of preciselt defines size
- based on objective measurements
- parametric test
* mean
* standard deviation
define content analysis
studying human behaviour indirectly by studying things that we produce eg tv adverts
= allows insight into structured values, beliefs and prejudices
how to conduct:
1. identify hypothesis
2. create a coding system (eg: 1= male, 2= female)
3. gather resources
4. conduct content analysis and record data
5. analyse data which is descriptive and qualitative - themantic analysis
6. write up report
evaulate content analysis
strenghts:
* strong external validity as data is already in real world = high mundane realsim
* produces large data set of quantitave and qualitiative data
* easy replication
* data is public domain = no ethical issues
weaknesses
* oberserver bias is presented but can be elimated by achieving inter observer reliability
* content of choice to analyse may be biased
* interpretaive bias - may ignore some things
define themantic analysis
identifying and reporting patterns within the material
