"Experimental and Quasi-Experimental Designs for Research" (1963), which classified designs using X, O and R, was written by:

Norman Denzin and Yvonna Lincoln

Roger Wimmer and Joseph Dominick

Assertion (A): Quasi-experimental designs give weaker grounds for causal inference than true experiments. Reason (R): Quasi-experiments do not randomly assign participants to groups. Choose the correct option.

Both A and R are true, and R is the correct explanation of A

Both A and R are true, but R is not the correct explanation of A

Which of the following are qualitative research methods? Select all that apply.

A structured survey with closed-ended questions

Communication Research I: Definition, Concepts, Constructs and Approaches to the Research Process; Research Designs — Classical, Experimental and Quasi-Experimental, Variables and Hypotheses; Types and Methods of Research; and Research in Journalism, PR, Advertising, Cinema, Television, the Internet and Children's Media, and in India

The tenth unit is communication research. This chapter takes its first three items: the definition, concepts, constructs and approaches of the research process; research designs, their types, structure and components, from classical and experimental to quasi-experimental, with variables and hypotheses and the main types and methods of research; and research in journalism, public relations, advertising, cinema, animation and graphics, television, the internet, social media, magazines and children's media, with communication, journalism and media research in India. Research methods are also taught on this platform in Paper-1 and in several subjects; here they are taught as the methods of media research, with media examples throughout and a worked calculation of an experimental effect. Chapter 23 continues with measurement, sampling, data collection, statistics, ethics and report writing.

1. Communication research: definition, concepts, constructs and approaches

Fred Kerlinger defined scientific research as systematic, controlled, empirical and critical investigation of natural phenomena, guided by theory and hypotheses about the presumed relations among them. Communication research applies this to the processes, institutions, content, audiences and effects of communication. Its building blocks are concepts and constructs. A concept is an abstract idea that summarises observations, such as credibility or exposure. A construct is a concept deliberately created for a research purpose and usually built from several dimensions, such as "source credibility" (expertise and trustworthiness) or "media literacy". To be studied, each must be given an operational definition that says how it will be observed or measured: exposure to television news, for example, as the number of minutes a respondent reports watching news bulletins on an average day.

Three broad approaches coexist. The quantitative approach, rooted in positivism, measures variables and tests hypotheses statistically through surveys, experiments and content analysis, aiming at generalisation. The qualitative approach, rooted in interpretivism, seeks to understand meanings from participants' points of view through ethnography, in-depth interviews, focus groups and textual analysis, aiming at depth. The critical approach, from the traditions described in Chapter 2, asks about power, ideology and ownership. Mixed methods combine them, and Norman Denzin's idea of triangulation (1970), using more than one data source, investigator, theory or method to study the same problem, strengthens confidence in findings. The research process usually runs: identifying the problem, reviewing the literature, building a theoretical framework, stating research questions or hypotheses, choosing a design, sampling, collecting data, analysing and interpreting, and reporting.

⚠️ A concept is not yet a variable

"Media violence" is a concept; it becomes a variable only when operationally defined, for instance as the number of violent acts per hour of programming. Examiners test whether candidates know that operational definition is the step that makes a concept measurable.

2. Research designs: types, structure and components

A research design is the plan that links the research questions to the data and the analysis. By purpose, designs are exploratory (to gain familiarity with a little-known problem), descriptive (to describe characteristics, such as a survey of news consumption) or explanatory or causal (to test why something happens). Every design has the same components: the problem and objectives, the variables and hypotheses, the population and sample, the tools of data collection, the procedure, and the plan of analysis. Donald Campbell and Julian Stanley's Experimental and Quasi-Experimental Designs for Research (1963) classified designs, using X for a treatment, O for an observation and R for random assignment.

Campbell and Stanley's designs
Family	Design	Notation and note
Pre-experimental	One-shot case study; one-group pre-test post-test; static-group comparison	X O; O X O. No random control group, so causal claims are weak
True (classical) experimental	Pre-test post-test control group	R O X O / R O O. Random assignment plus a control group
True experimental	Post-test-only control group	R X O / R O. Avoids the effect of a pre-test
True experimental	Solomon four-group	Combines the two above to measure the effect of pre-testing itself
Quasi-experimental	Non-equivalent control group; time series	Comparison groups or repeated measures, but no random assignment, as when two existing classes or towns are compared

The value of a control group is shown by a simple calculation. Suppose an experiment tests whether a media-literacy programme improves students' ability to spot misinformation. The experimental group's mean score rises from 40 before the programme to 55 after it, a gain of 15; the control group, which did not receive it, rises from 40 to 45, a gain of 5 caused by practice, maturation or outside events. The effect attributable to the programme is 15 − 5 = 10 points. Internal validity is the confidence that the treatment, not something else, caused the change; Campbell and Stanley listed its threats: history (outside events), maturation (natural change in participants), testing (the effect of a pre-test), instrumentation (changes in the measuring tool), statistical regression to the mean, selection (non-equivalent groups), and experimental mortality (drop-outs). External validity is the extent to which results generalise to other people, settings and times.

3. Variables and hypotheses

A variable is a characteristic that takes different values. The independent variable is the presumed cause (hours of television watched), the dependent variable the presumed effect (fear of crime). An intervening or mediating variable explains how the cause works (television increases perceived risk, which increases fear). A moderating variable changes the strength or direction of the relation (the effect may be stronger among viewers who live in high-crime areas). Extraneous or confounding variables are other influences that could produce the effect and must be controlled, for example by random assignment or by measuring and holding them constant. Control variables are those deliberately held constant.

A hypothesis is a testable statement of the expected relation between variables. The null hypothesis (H0) states that there is no relation or difference ("exposure to television news is not related to political knowledge"); the alternative or research hypothesis (H1) states that there is one. A directional hypothesis predicts the direction ("heavier viewers know more"); a non-directional one predicts only a difference. A good hypothesis is grounded in theory or earlier findings, specific, testable with available methods, and stated in terms of operationally defined variables. Where a study is exploratory, research questions take the place of hypotheses.

🧠 Mediator explains how, moderator says when

If the question asks through what process A affects B, the answer is a mediating (intervening) variable. If it asks for whom or under what conditions the effect is stronger, the answer is a moderating variable.

4. Types and methods of research

Types of research
Type	Aim	Media example
Basic (pure)	To build theory and knowledge	Testing agenda-setting across media systems
Applied	To solve a practical problem	Testing which health message increases vaccination
Descriptive	To describe what exists	A survey of news sources used by young people
Analytical	To analyse facts already available and explain relations	Analysing ratings data to explain channel choice
Historical	To reconstruct the past from primary and secondary sources, testing them by external criticism (authenticity) and internal criticism (credibility)	The history of the nationalist press
Case study	To study one case in depth, using several sources; suited to "how" and "why" questions about contemporary events (Robert Yin)	How one newsroom adopted digital-first working
Longitudinal	To study change over time	Trend study (different samples from a population over time), cohort study (different samples from a group sharing an experience, such as those born in one decade), panel study (the same people repeatedly)
Cross-sectional	To study a population at one point in time	A one-time survey of media use

⚠️ Panel is the same people

Only the panel study measures the same individuals repeatedly, which is why it alone can show individual change; trend and cohort studies draw new samples each time. Panels suffer from attrition as members drop out.

5. Research across media fields, and communication research in India

Each field of media practice has its own research traditions. Journalism research includes newsroom ethnographies such as Herbert Gans's Deciding What's News (1979) and Gaye Tuchman's Making News (1978), content analyses of coverage, and surveys of journalists. Public relations research measures outputs, outtakes and outcomes and audits communication (Chapter 11). Advertising research pre-tests and post-tests campaigns (Chapter 9). Cinema research combines textual analysis, industry studies and audience research; animation and graphics research tests usability and attention with eye-tracking. Television research ranges from ratings to cultivation analysis; internet and social-media research now uses digital ethnography (Robert Kozinets's netnography) and computational analysis of large data sets. Magazine research studies readership and content. Children's media research has asked about the effects of television violence since Bandura's experiments, and the research programme behind Sesame Street (from 1969), with formative research during production and summative evaluation afterwards, became a model for educational media.

Communication research in India grew out of development communication. The evaluations of SITE (1975–76), carried out by social scientists at ISRO, among them Binod C. Agrawal, and the Kheda Communication Project made India a laboratory for studying television's social effects, and Shingi and Mody's "communication effects gap" (Chapter 4) came from this work. University departments and the Indian Institute of Mass Communication, whose journal Communicator has published Indian media research for decades, broadened the field; audience measurement bodies such as BARC and the Indian Readership Survey supply industry data; and the Asian Media Information and Communication Centre (AMIC, founded in Singapore in 1971), with its journal Media Asia, connected Indian scholars to regional research. Recent Indian research has turned to digital media, misinformation, regional-language media, caste and gender in media, and political communication, while scholars continue to call for more work in Indian languages and more theory grounded in Indian experience.

Key takeaways

Research is systematic, controlled, empirical and critical (Kerlinger); concepts and constructs become variables through operational definitions; approaches are quantitative, qualitative, critical and mixed, with Denzin's triangulation (1970); the process runs from problem to report.
Designs are exploratory, descriptive or explanatory; Campbell and Stanley (1963) classify pre-experimental, true experimental (pre-test post-test control group, post-test-only, Solomon four-group) and quasi-experimental designs; effect = experimental gain − control gain.
Internal validity threats: history, maturation, testing, instrumentation, regression, selection, mortality; external validity is generalisability. Variables: independent, dependent, mediating (how), moderating (when), extraneous and control; hypotheses: null and alternative, directional or not.
Types: basic and applied, descriptive and analytical, historical (external and internal criticism), case study (Yin), longitudinal (trend, cohort, panel, the last with the same people) and cross-sectional.
Field research: newsroom ethnography (Gans 1979, Tuchman 1978), PR and advertising evaluation, cinema and television studies, netnography and computational methods, children's media and Sesame Street's formative and summative research; in India SITE and Kheda evaluations, IIMC's Communicator, BARC and IRS, and AMIC (1971).

Practice questions (10)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

The Solomon four-group design belongs to which family of research designs?
1. Quasi-experimental
2. True experimental
3. Pre-experimental
4. Historical
Show answer
Answer: B — True experimental
The Solomon four-group design uses random assignment and control groups, combining the pre-test post-test and post-test-only designs to measure the effect of pre-testing itself, so it is a true experimental design. Pre-experimental designs lack random control groups, and quasi-experiments lack random assignment.
A longitudinal study that measures the same group of individuals repeatedly over time is a:
1. Cross-sectional study
2. Cohort study
3. Trend study
4. Panel study
Show answer
Answer: D — Panel study
A panel study follows the same individuals. Trend studies draw different samples from a population at each point, cohort studies draw different samples from a group sharing an experience, and a cross-sectional study measures once.
Which of the following are threats to internal validity identified by Campbell and Stanley? Select all that apply.
1. Population validity
2. Maturation
3. Ecological validity
4. History
Show answer
Answer: B — Maturation; D — History
History (outside events) and maturation (natural change in participants) threaten internal validity, the confidence that the treatment caused the effect. Population and ecological validity are aspects of external validity, whether results generalise to other people and settings.
"Experimental and Quasi-Experimental Designs for Research" (1963), which classified designs using X, O and R, was written by:
1. Donald Campbell and Julian Stanley
2. Norman Denzin and Yvonna Lincoln
3. Fred Kerlinger and Howard Lee
4. Roger Wimmer and Joseph Dominick
Show answer
Answer: A — Donald Campbell and Julian Stanley
Campbell and Stanley's short book set the standard classification of experimental designs and threats to validity. Wimmer and Dominick wrote the textbook Mass Media Research, Kerlinger wrote on the foundations of behavioural research, and Denzin and Lincoln edited handbooks of qualitative research.
Assertion (A): Quasi-experimental designs give weaker grounds for causal inference than true experiments. Reason (R): Quasi-experiments do not randomly assign participants to groups. Choose the correct option.
1. Both A and R are true, and R is the correct explanation of A
2. Both A and R are true, but R is not the correct explanation of A
3. A is true, but R is false
4. A is false, but R is true
Show answer
Answer: A — Both A and R are true, and R is the correct explanation of A
Both are true and R explains A: without random assignment the groups may differ before the treatment, so selection threatens internal validity and a difference in outcomes cannot be confidently attributed to the treatment. That is the defining weakness of quasi-experiments.
Using both a survey and in-depth interviews to study the same research problem, so that each checks the other, is an example of:
1. Experimental mortality
2. Methodological triangulation
3. Operational definition
4. Statistical regression
Show answer
Answer: B — Methodological triangulation
Denzin called the use of more than one method methodological triangulation; he also described data, investigator and theory triangulation. An operational definition specifies how a concept is measured, while mortality and regression are threats to internal validity.
In a pre-test post-test control group experiment on a media-literacy programme, the experimental group's mean score rises from 40 to 55 and the control group's from 40 to 45. What is the effect attributable to the programme, in points? Type the number.
Numerical answer — type the value.
Show answer
Answer: 10
Experimental gain = 55 − 40 = 15; control gain = 45 − 40 = 5, caused by practice, maturation or outside events. Effect = 15 − 5 = 10 points. Reporting 15 would credit the programme with change that happened anyway.
A variable that changes the strength or direction of the relationship between an independent and a dependent variable is a:
1. Control variable
2. Dependent variable
3. Moderating variable
4. Mediating variable
Show answer
Answer: C — Moderating variable
A moderator specifies when or for whom an effect is stronger or weaker. A mediator explains how the effect occurs, the dependent variable is the outcome, and a control variable is one held constant.
Which of the following are qualitative research methods? Select all that apply.
1. Ethnography
2. A structured survey with closed-ended questions
3. Focus group discussion
4. In-depth interview
Show answer
Answer: A — Ethnography; C — Focus group discussion; D — In-depth interview
Ethnography, focus groups and in-depth interviews seek meanings in participants' own terms. A structured survey with closed-ended questions produces numbers for statistical analysis and is quantitative.
"Deciding What's News" (1979), a landmark ethnography of American newsrooms, was written by:
1. Philip Meyer
2. Herbert Gans
3. Gaye Tuchman
4. David Manning White
Show answer
Answer: B — Herbert Gans
Herbert Gans studied news selection at CBS, NBC, Newsweek and Time. Tuchman's Making News (1978) is the companion classic; Meyer wrote Precision Journalism and White did the 1950 gatekeeping study.