Research Methods Exam 3

What is a single-subject design?

A research method that focuses on observing the behavior of a single participant or a very small group over time, often with repeated measures, to evaluate the effect of an intervention or treatment. The participant serves as their own control.

What do we typically think of as the dependent variable, the independent variable, and the research question in single-subject designs?

• Dependent Variable: The behavior being measured.
  
  

• Independent Variable: The treatment or intervention applied.
  
  

• Research Question: Whether the treatment causes a change in the behavior.

What are the three required parts of a good single-subject design?

Baseline phase (measure behavior without intervention).

Intervention phase (introduce treatment).

Repeated measurement to monitor changes over time.

Be able to define the terms phase and treatment.

Phase: A distinct period with a specific set of conditions (e.g., baseline or treatment phase).

Treatment: The intervention or independent variable applied to alter the dependent behavior.

How does notation work in single-subject designs (e.g., what is an ABAC design?)

• Letters represent phases:

  • A = Baseline (no treatment)

  • B = First treatment

  • A = Return to baseline

  • C = Introduction of a new or different treatment
    

ABAC means baseline → treatment 1 → baseline → treatment 2.

How do we determine whether a phase is “stable”?

A phase is stable when the behavior shows minimal variability (the data points are relatively flat or consistent) and no major trends upward or downward.

How are trends defined? (CHECK NOTES)

A trend is a consistent change in behavior in one direction over time, such as steadily increasing, decreasing, or cyclic patterns.

What is a baseline, and what are the two main functions of a baseline?

• A baseline is the measurement of the behavior before any treatment is introduced.
  
  

• Functions:

  1. Describe the natural rate of behavior.

  2. Predict future behavior if no intervention occurs.
     

In general, what is the minimum number of data points that we should collect to establish a proper baseline?

Typically, at least 3–5 data points are recommended, although more is better for establishing stability.

What are drifting baselines and unrecoverable baselines?

• Drifting baseline: Behavior slowly changes over time even without treatment (problematic because it reduces confidence in conclusions).

• Unrecoverable baseline: After treatment, the behavior does not return to its original baseline level.

How can reactivity, observer drift, complexity, and expectancy affect data collection in single-subject research?

• Reactivity: Participants change behavior because they know they are observed.

• Observer Drift: Observers unconsciously change criteria over time.

• Complexity: Hard-to-define or very detailed behaviors are difficult to code reliably.

• Expectancy: Observer’s expectations bias their recording of behavior.

What is a reversal design?

A design (such as ABA or ABAB) where treatment is introduced, withdrawn, and reintroduced to see if behavior changes systematically with the treatment.

What are considerations that the researcher must make when determining whether a reversal design is appropriate?

Behavior must be reversible.

Ethical concerns: it may be unethical to remove a beneficial treatment, especially for harmful behaviors.

Be able to describe changing criterion designs, alternating treatment designs, and multiple baseline designs.

Changing Criterion Design: Behavior is shaped gradually by setting new "criteria" or goals over time (e.g., increasing homework by 5 minutes per week).

Alternating Treatments Design: Different treatments are rapidly alternated to see which one is more effective.

Multiple Baseline Design: Treatment is introduced at different times across behaviors, settings, or participants.

What are the advantages of Changing Criterion design?

Good for showing control over gradual behavior change.

What are the advantages of Alternating Treatments design?

Direct comparison of multiple interventions quickly.

What are the advantages of Multiple Baseline design?

No withdrawal of treatment is needed, and good for irreversible behaviors.

How do we determine the “significance” of results in a single-subject design?

Through visual inspection: look for changes in level, trend, and variability between phases.

Overall, what are the main advantages and disadvantages of a single-subject approach to research?

• Advantages:

  • Detailed information about individuals.

  • Good for rare or ethical cases.

  • Flexible and low-cost.

• Disadvantages:

  • Limited generalizability.

  • Subjective visual analysis.

  • Replication needed to strengthen conclusions.

What is the difference between an experimental and a non-experimental study?

Experimental: Researcher manipulates variables.

Non-experimental: Researcher only observes naturally occurring variables.

What are some circumstances under which we would not want to conduct an experiment?

Ethical concerns (can't manipulate harmful variables).

Practical issues (too costly or impossible to manipulate).

Natural phenomena that cannot be artificially created.

What is correlational research?

Studies relationships between variables without manipulation to determine associations.

What does it mean to test “predictive hypotheses”?

Use existing data to predict future outcomes based on known relationships.

How are correlational studies limited by third variable and directionality problems?

Third-variable problem: An unknown variable influences both studied variables.

Directionality problem: Unsure which variable causes the other.

What is a correlation coefficient? What is its range?

A statistic (r) measuring the strength and direction of a relationship.

Range: -1.00 to +1.00.

What is program evaluation, and why is it used? What are some key ethical considerations in program evaluations?

Program evaluation: Systematically collecting information to assess a program’s effectiveness.

Ethical considerations: Protect participant confidentiality, avoid conflicts of interest, and be honest about results.

What is archival research, and what are some strengths and weaknesses?

Using existing records or documents for research.

Strengths: Saves time and money, large datasets.

Weaknesses: Limited control over how data was collected; missing or biased data.

Weakness of Cross-sectional Design

Cohort effects (groups differ beyond age).

Sociometry- define, strengths, weaknesses

Define- Measures social relationships in groups.

Strengths- Good for mapping social structures.

Weakness- Subject to bias/self-report errors.

Content Analysis- define, strengths, weaknesses

Define- Systematic analysis of media/documents.

Strengths- Large datasets can be analyzed easily.

Weakness- Interpretation bias possible.

How do sequential designs address the limitations of both longitudinal and cross-sectional designs?

Sequential Designs: Combine both, following multiple age groups over time to minimize cohort and time effects.

What are the different types of variables that sociometry can assess?

Popularity, friendships, social rejection, leadership, and group dynamics.

What is an unstructured naturalistic observation?

Observation without predefined categories—just general watching and recording of behaviors.

What are some examples of the way researchers in different fields might use unstructured naturalistic observations?

Developmental psychologists watching children play.

Ethnographers observing social interactions at public events.

Know the advantages and disadvantages of unstructured naturalistic observations.

Advantages: Captures rich, spontaneous behavior.

Disadvantages: Difficult to quantify and analyze; subject to observer bias.

What is a structured observation?

Observers use a specific coding system to record behaviors systematically.

What are key considerations when conducting a structured observational study?

Clear operational definitions, training observers, ensuring reliability.

How can structured observations be conducted in laboratory or naturalistic settings?

Laboratory: Controlled environment.

Naturalistic: Real-world setting but using structured coding.

What is topography in coding?

Exactly what the behavior looks like — the specific physical form or pattern of the behavior.

Imagine you are studying helping behavior.

Instead of just recording "helped: yes/no," you would record how they helped:

Picked up dropped papers

Held the door open

Gave verbal encouragement

Helped carry something heavy

Know the six types of behavioral coding that we discussed in lecture.

Coding Type	Example
Frequency	Number of times child hits peer.
Duration	How long a tantrum lasts.
Interval	Whether aggression occurs during 10-minute intervals.
Latency	Time from instruction to action.
Intensity	How hard a child hits.
Topography	Specific form of behavior (e.g., type of helping behavior).

What is Percent Agreement?

• Definition:
  Percent agreement is simply the percentage of times that two (or more) observers record the same behavior in the same way.

• How it works:

  (number of agreements / total number of ratings) * 100

  Example:

  • Out of 100 observations, if raters agreed on 85,
    → Percent agreement = 85%.

• Major Limitation:
  ✅ Does not account for chance agreement.

  • Even if raters agree randomly (by luck), percent agreement treats it as perfect reliability, which overestimates true agreement.

What is Pearson Product-Moment Correlation (Pearson's r)?

Pearson’s r measures the strength and direction of the linear relationship between the ratings of two observers.

How it works:

You calculate a correlation coefficient (r) between Observer 1's scores and Observer 2's scores.

r ranges from -1.00 (perfect negative relationship) to +1.00 (perfect positive relationship).

Major Limitation:

✅ Measures association, not agreement.

Two raters could score consistently higher or lower than each other but still have a high correlation.

Example:

Observer A always rates behavior at 3, 4, 5.

Observer B always rates the same behaviors at 6, 7, 8.

They correlate perfectly (r = 1.0), but they disagree on actual scores.

What are factors that affect reactivity?

How noticeable the observer is, setting familiarity, and awareness of being studied.

How can reactivity be minimized?

Habituation, unobtrusive observation, concealed observation.

What are partial concealment and total concealment?

Partial Concealment: Participant knows they’re observed but not exactly why.

Total Concealment: Participant doesn’t know they’re being observed at all.

What are contrived observations?

Researcher sets up a situation to elicit a behavior (e.g., dropping papers to see who helps).

nonreactive observations

Example	Why it's Nonreactive
Analyzing public social media posts	People already posted them — they aren't changing behavior for research.
Using police records to study crime trends	Data already collected for other purposes — no influence from researchers.
Observing recycling behavior through hidden cameras	People don’t know they’re being watched — behavior stays natural.
Content analysis of newspaper articles	Newspapers were written for public, not for the researcher.

What are ethical issues in nonreactive observations?

Consent, invasion of privacy, deception without debriefing.

What is observer bias?

When observers' expectations influence how they record data.

Know the basic methods, results, and conclusions of Langer & Abelson (1947).

Observers rated the same person more negatively when told he was a "patient" vs. "job applicant," showing observer expectations heavily influenced their judgments.

What are the various ways of reducing observer bias when conducting research?

Train observers, use blind observers, multiple raters, strict operational definitions.

How do the research processes differ in qualitative vs. quantitative research?

Sample selection: Purposeful vs. random.

What are the basic tenets of qualitative researchers?

Reality is subjective.

Context matters.

Rich descriptions are valued over numerical data.

Be able to explain the beliefs and methodological choices involved in naturalistic inquiry and purposeful sampling.

Intentionally select participants with rich, relevant experiences.

What are the basic methods and objectives of case studies, participant-observer research, and ethnographic methods?

Case Studies: Deep study of one case.

Participant-Observer Research: Researcher actively engages with group.

Ethnographic Methods: In-depth study of cultures.

What are the major advantages and disadvantages of qualitative research?

Advantages: Deep insights, flexible, captures context.

Disadvantages: Time-consuming, hard to generalize, potential bias.