flashcards content analysis
what is content analysis?
method for systematically describing the meaning of qualitative data (texts) by assigning parts of the material through categories of a coding frame
- way to extract the meaning of the text
- putting texts into categories, summarizing them in some ways
sources/texts we analyze can be any type: interview transcrepts, speeches, literature, social media, written reports
e.g. social values in ‘little orphan annie’ comic strips -> gives narrative, examples and frequency counts (not wordcount but themes)
qualitative content analysis
originated from critique of quantitative content analysis
- they have many procedures in common
quantiative = manifest content/meaning
- assumption that word frequency implies importance
- really efficient, good for much material
critique = overlooks latent meaning
- meaning is often complex, holistic, context dependent = not necessarily apparent at first sigh
- proposition: contextual analysis beyond manifest content and (uncontextualized frequency counts)
now: tons of text -> sometimes infeasible to do qualitative analysis on the whole material (= trade-offs)
!!unit of analysis is not just necessarily the words
main characteristics qualitative content analysis
- systematic:
- analyzes the whole text material (avoid cherry-picking specific parts) to make an argument
- specifies a sequence of categorizing instructions beforehand (but allows adaptation)
- runs the coding twice to verify consistent categorization (pilot vs main phase) - reduces data: analyzes the whole text material BUT focus on specific parts
- what aspects? aspects of meaning (themes, tropes, statements) related to the RQ
- text pieces not relevant to coding framework stay out (by definition) - flexible = coding categories may be modified to adapt a specific tet (but always guided by latent concepts): combines concept-driven (theory-driven) and data-driven categories
= main diff with quantitative content analysis
contrast quali and quanti content analysis
shared procedures -> they look very similar
(bc quali develops from quanti)
- coding to systematically describe data
- definitions for categories
- predefined steps
- pilot phase followed by main analysis phase
- adaptation in second iteration(s) = do it twice to check it
- quality criteria for categories: validity and consistency
- presentation may involve frequency counts (but interpretation is diff)
but also differences
- main = quali goes beyond manifest meaning (not words at face value) = latent and contextual meaning
- collect THEMES not words (diff unit of analysis), themes can also be in phrases or tropes
- to infer a theme is important we require more evidence than sheer frequency -> analyze material holistically and its publication context - flexible coding: concept-driven and data-driven: modify categories to contain all the data (to understand latent meanings + discover info we didn’t know existed)
= we adapt our coding frame
unit of analysis quali CA
quali goes beyond manifest meaning (not words at face value) = latent and contextual meaning
- collect THEMES not words (diff unit of analysis), themes can also be in phrases or tropes
- to infer a theme is important we require more evidence than sheer frequency -> analyze material holistically and its publication context
how to do qualitative CA
8 steps
- decide RQ
- select material: any kind of text
- build coding frame
- segmentation
- trial coding
- evaluating and modifying the coding frame
- main analysis
- presenting and interpreting the findings
step 3: build coding frame
create categories and subcategories
- categories = aspects of the text-material about which we want information (e.g. social values, objectives in life)
- subcategories = specify what is said about the categories (e.g. a specific objective in life)
everything that is relevant is going to be considered
to create the frame =
- take a (purposive, non-random) sample: select parts of the text that reflects the diversity of sources
- to structure categories/sub-categories =
- use concept-driven categories (a theory, prior research, everyday knowledge, logic or an interview guide gives expectations of what you will see)
- AND data-driven categories: inductively = accounting for the characteristics of our data (things you didnt take into account beforehand, before analyzing the text)(data-driven supplements concept-driven)
!this data driven categories you wouldn’t do with quanti
!make procedures transparent and replicable -> define and keep a CODEBOOK: for all categories and sub-categories give:
(explanation how you classify the text)
step 4: segmentation
= need to define the unit of analysis
- formal = element of syntatix, like a word, paragraph, individual interviewee
- thematic = diff themes or discursive pieces
usually thematic = more appropriate for qualitative analysis, but less clear-cut than if we use an element of syntaxis
e.g. unit is a piece of discourse in relation to euthanasia (not necessarily a single sentence/word/paragraph)
- sentence may have two themes: an opinion and a reasoning behind it for example
quality criteria for categories and subcategories
- unidimensionality = each category, one concept (one unit/theme should not go into two categories)
- overall coding frame covers many concepts/dimensions
- e.g. opinion vs reasons for the opinion - mutually exclusive = no ambiguity about where a unit of text should be categorized
!unit is a theme, not a sentence or smth, so a sentence can be in multiple categories, a theme cant - exhaustive: framework allows to categorize all relevant aspects of text-material
- we can always have a miscellaneous or residual category, but this is risky: we should question why the units can’t go elsewhere and if we shouldn’t create a more specific category
step 5: trial coding
+ step 6: evaluating and modifying the coding frame
= pilot phase
- select a sample from the material
- code it
- assess whether the coding framework should be modified
two trials, ideally two coders working independently (or the same person at diff times)
what % of units get equally classified the second time compared to the first?
= look for consistency
+ look for validity (are the categories adequately describing what we’re talking about)
make sure your coding frame is good enough before going into the whole analysis
step 7: main analysis
+ step 8: report results
main analysis = all relevant aspects of the text material is coded
reporting = represent the frame and illustrate through quotations
- quantitative style = frequencies or inferential statistics (e.g. chi-square)
- for quali you can also count the themes, than there is more interpretation than with quanti counting words
applied reading: minority empathy hypothesis in NL 1930-39
RQ = does a minority status affect opinion on the ‘other’?
- compare how catholic elites (north holland they are minority, in limburg they are majority) talk about jewish communities
- hypothesis = religions minorities discriminate less and show more empathy toward persecuted an stigmatized minorities than their majority counterparts
text material = public statements regarding Jewish communities in the 2 most important catholic newspapers of the regions, only the digitalized ones
method = qualitative content analysis (semi-automated = finding text through machines but hand-coded)
coding frame = completely theory-driven, but well adapted to the context with secondary literature
- categories = identity-related frames (pluralism, assimilation, similarity, difference)
- subcategories = ultilitarian frames (opportunity and threats)
segmentation (unit of analysis): acts of public claim making = public articulation of political demands, critiques, proposals and policies targeting specific collective actors
example categories and subcategories - interview about Terry Schiavo’s euthanasia
- category = opinions about Schiavo’s euthanasia
(subcategories of opinion: morally justified, long overdue, morally wrong) - category = reasons in favor for turning off machines
(subcategories of reasons: unneccessary prolongation of suffering, high cost)
!!!if you did this with quanti
RQ: do people reflect on euthanasia in moral or practical terms? -> define words associated with moral and with practical terms -> count which one you see more
- what are you missing = reasons why people oppose/support euthanasia
- is like a survey: extracts data you’re looking for, not possible to get additional info
