Communication science cheat sheet

Communication model

A sender encodes a message, a channel carries it, and a receiver interprets it within a context.

The same message can be interpreted differently by different audiences.

Noise and feedback

Noise interferes with communication. Feedback gives information about whether a message was received or understood.

An unclear audio connection creates technical noise.

Research methods

Surveys, interviews, experiments and content analysis answer different kinds of communication questions.

Content analysis can code recurring themes in a collection of messages.

Operational definitions

Define an abstract concept in a way that can be observed or measured consistently.

Define “engagement” before comparing how audiences respond.

Interpersonal and mass communication

Compare the number of participants, feedback speed, channel and audience control. A public social-media post can combine interpersonal interaction with mass distribution.

Classify the communication event, rather than assuming that every use of one platform has the same audience.

Reliability and validity

Reliability concerns consistency of a measure. Validity concerns whether the evidence supports the interpretation you want to make from that measure.

A consistently counted click is reliable as a click count, but may be a poor measure of understanding.

Sampling an audience

A sampling frame determines who can be selected. Volunteer responses can over-represent people with strong opinions, even when the total number of replies is large.

Report how participants were recruited before treating survey findings as representative.

Coding messages

Content analysis needs clear categories, a defined unit of analysis and a reproducible coding procedure. Independent coders can reveal ambiguous category definitions.

Decide whether one unit is a sentence, a post or an entire conversation before counting themes.

Worked example: response rate

A survey sends 250 eligible invitations and receives 75 completed responses. Its simple completion proportion is 75 ÷ 250 = 30%; non-response may still bias the findings.

A 30% completion proportion is not a claim that 30% of the population agrees.

Common mistake: medium determines meaning

A channel’s features influence communication but do not determine interpretation on their own. Language, relationships, prior knowledge and cultural context also matter.

Use evidence about the message and audience before claiming that a format caused an effect.

Check your understanding

Why can a very large online poll still be biased?

Large size reduces some random error but does not correct a non-representative sampling process.

Curriculum connections

Sources