CMAF - Research Methods

What is content analysis?

Counting frequency of occurrence in media. It is an unobtrusive form of analysis.

What does Berger say about content analysis?

Content analysis can only make conclusions about the content itself.

Content analysis cannot make inferences about what?

Audiences or producers motivations (Examples: gambling ad does not mean that the audience are gambling addicts or that a frequency of a brand being mention does not mean there is a paid deal)

Deductive content analysis means what?

You determine your categories in advance.

Inductive content analysis means what?

Your categories emerge from your analysis.

What approaches should you use for sampling?

Historical or comparative

What do you call what is being counted in content analysis?

The “unit” of analysis

What does it mean to code something?

Assign material sampled to categories of tags. They are then put in a “coding manual”.

What is inter-coder reliability?

The degree to which multiple researchers categorize analyzed units in the same way.

What is the ideal inter-coder reliability?

90% or higher

What might cause issues with inter-coder reliability?

Unclear categories, issues with the coding manual, or poor coder training.

Advantages of content analysis…

Unobtrusive, no ethics approval needed, inexpensive, easy to obtain material, rests on a transparent process, quantifiable data, provides important insights.

Challenges of content analysis…

Finding a representative sample, determining measurable units, defining codes, reliability of coding, human interpretation, manifest content, and cannot answer “why” questions

An example of historical content analysis

The frequency of occurrence over time. Black women in leading roles in top box office films.

An example of comparative content analysis

The frequency of occurrence from two different media sources / platforms. The amount of times terrorist is used in 2 different newspapers.

Define semiotics

The “science” of signs

Content analysis versus semiotics analysis?

CA deals with the manifest (visible) categories and semiotics deals with the latent (below the surface).

How do CA and semiotics work together?

Semiotics fills in the gaps left from CA, and CA overcomes the subjectives of Semiotics

What equals SIGN?

Signifier (what we perceive) + signified (mental image)

Denotation

Literal meaning

Connotation

Implied meanings

Syntagmatic

Meaning created through a chain of signs

Paradigmatic

Meaning created through absence and substitutions

Ethnography

A research method in which the researcher is immersed in a natural field research setting for a prolonged periods, making deep observations about social interactions. Called “participant observation”.

Overt Ethnography

Actively involved in the people that are being studied

Covert Ethnography

Merely observers of the people being studied

Public vs Closed Ethnography

Public setting (city streets, public spaces) and Closed setting (workplaces, organizations, social movements)

What is the hardest form of ethnography to perform?

Closed and covert!

What is thick description?

Offer enough detail to take readers there by describing in a nuanced manner smaller but significatn details from the field.

Challenges of ethnography

Time commitment, access, ethics approval, taking accurate notes

Advantages of ethnography

Deep insight, potential benefits to group through interaction and collaboration, holistic, all aspects of human behaviour, social significance

What do we know about open ended survey questions?

They have authentic ideas, unprompted knowledge, contemplated answers. They must be manually entered, coded and analyzed. Respondents are very likely to skip these.

What do we know about closed ended survey questions?

They have limited choices from which to select. Minimized variability can help with response rates. Responses lack spontaneity / authenticity. Might not cover all options. Often gets random clicks.

What is a likert scale?

A scale or spectrum that has participants ranking their responses normally out of 5 to 7 options of agree to disagree. Even number scales avoid a neutral middle option.

What to think when designing survey questions?

short, easy to understand, avoid ambiguous terms, no double barrelled questions, avoid leading questions

Bivariate analysis

Tests is there is a relationship between multiple variables. Doesn’t indicate causation.

Differences between verbal surveys and interview research?

Questionnaire versus a systematic conversation

What are the 3 types of interview research?

Structured (short script), semi-structured (optional follow ups), and unstructured (basic protocol)

What is standard interview protocol?

Explain the instructions and the questions. Be respectful of time, follow a logical order / flow, no leading questions, avoid questions with a socially acceptable answer.

What type of sampling is common for interview research?

Convinience and snowball

How are interviews then studying post?

Manually transcribed which typically takes 5 to 6 hours per recording, and then reviewed for common themes

Focus groups are a type of what research?

Interview - though they are harder to recruit / schedule, participates may dominate others, and it is difficult to transcribe.

Media go along

researchers follow participates going through a normal digital expirience

Basic formula for inter-coder reliability?

Number of identical decisions by coders divided by total decisions made.

EV car research, what did Kyle look at?

Setting and people involved