Data Analysis And Reporting On Investigations Flashcards
Content analysis - top down
Systematic research tech for turning qualitative into quantitative
Uses coding system of predetermined categories which are applied to material in a consistent manner
Pilot study often needed to generate and test the coding system
Easier to analyse and compare data BUT time consuming, reductionist, biased and subjective
Content analysis - bottom up
Thematic analysis/inductive analysis - data dictates - no list of themes when observing - let them emerge
Accessible and understandable - easy to share findings with those who may benefit
Valid as not imposing a structure - BUT subjective and no control
Content analysis - cycle of review
Repeat
Read pilot -> add improvements
Review
Refine
Thematic analysis
Transcribe data
Divide into meaning units
Search text for meanings
Keep adjusting themes
Define and name each theme
Write up report
Levels of measurement
Nominal - named categories, mutually exclusive
Ordinal - ranked or rated data
Interval - equal intervals on a set standard of measurement - no 0 point eg temp, IQ
Ratio - has absolute 0 eg. Height, time etc
Statistical tests
If looking for a correlation - SPEARMANS RHO - for result to be sig observed > critical
If looking for a difference and it’s nominal data - CHI SQUARED - for result to be sig observed > critical
If looking for a difference, it’s at least ordinal and it’s independent groups - MANN WHITNEY U - for result to be sig critical > observed
If looking for a difference, it’s at least ordinal and it’s repeated measures - WILCOXON T - for result to be sig critical > observed
Type 1 and 2 errors
Type 1 - accept alt where there is none - caused as level of sig set too leniently eg 0.2 - easier to get a sig result
Type 2 - accept null when there is an observed effect - caused when level of sig set to low eg 0.001 - harder to get a sig result
Content analysis
Analysing data so it can be placed in categories and counted (quant) or analysed in themes (qual)
