Descriptive Studies: Research Methods in Psychology

Descriptive Studies

Overview

• Descriptive studies aim to characterize a phenomenon by answering questions like:

  • How often does this occur?

  • How likely is something to happen?

  • What are people’s opinions?

  • What is this phenomenon like?

Frequency Claims

• Descriptive studies can test frequency claims, such as:

  • "3-year-olds understand sarcasm."

  • "Homeless adults feel disconnected from society."

  • "People with bipolar disorder have difficulties forming romantic relationships."

Sampling Procedures and External Validity

• Sampling procedures mainly affect external validity, determining how much the findings can be generalized.

• Key questions to consider:

  • What is the population of interest (who do you care about)?

  • How randomly can you sample from that population?

  • The more random the sample, the higher the external validity.

Sampling Challenges

• It's usually not possible to measure everyone in the population, so we rely on taking a sample.

• Illustrative diagram:

  • All people → Getting a sample → People near me → People in So Cal → People in college → People with similar interests → "Convenience sampling"

• Sample bias occurs when a sample is systematically different from the population it’s supposed to represent.

Issues with Sampling

• Psychology often relies on convenience/self-selected sampling.

  • Voluntary participation, motivated participants.

• Psychology participants are often WEIRD:

  • Western

  • Educated

  • Industrialized

  • Rich

  • Democratic

Improving Sampling

• With true random sampling, every member of the population has an equal opportunity to participate, but this is difficult to achieve.

• It’s easy to unintentionally introduce sample bias.

• To make samples better:

  • Have a system, such as approaching every third person.

• Stratify your sample to better represent the population.

• Use clusters to diversify your sample.

• Oversampling ensures small sub-populations are equally represented.

Ecological Validity

• Another component of external validity is ecological validity:

  • The extent to which findings can generalize to actual experience.

• A carefully controlled study environment can be very different from the real world.

• This creates some fundamental tradeoffs with other validities.

Assessing External Validity

• Start simple:

  • Where in the world was the study conducted?

  • When was the study conducted?

  • How specialized is the target population?

• Consider details:

  • How were participants recruited?

  • What might be different about people who decide to participate?

  • Are the measures reflective of what people actually do?

Construct Validity: Getting Accurate Measurements

Self-Report Measures

• Participants tell us about their own behaviors/mental processes.

• They’re good for:

  • Subjective impressions

  • Personal opinions

  • Easily reportable phenomena

  • Fundamentally unobservable phenomena

• However, participants sometimes lack insight.

Maximizing Construct Validity in Self-Reports

• Better when participants are:

  • Not relying on unreliable info (biases, memories, etc.)

  • Not motivated to lie (minimize social desirability)

  • Motivated to answer correctly

• Double checking validity:

  • Participants may fall into a response set

    • Reverse-scored items: ask the opposite

    • Validation items: are you paying attention?

• Questions should be clear and specific.

Improving Question Clarity

• Example of a bad question: "Do you believe that if Republicans were obstructing Obama like Democrats are doing to Trump, the mainstream media would attack Republicans?"

• Reason: Double-barreled questions ask about more than one thing at a time.

• How to make questions clearer:

  • Bad question: “Are you a drug user?” (Focuses on personal identity, not behavior)

  • Less bad question: “Have you ever used an illicit drug for recreational purposes?” (Categorical scale of measurement (yes/no); poor statistical validity)

  • Less bad question: “In the past 3 months, how many times have you used an illicit drug for recreational purposes?” 0 times 1-5 times 6-10 times 11+ times (Interval scale of measurement; better statistical validity)

Observational Measures

• Measure the actual variable.

• Reduce the problems of self-report, but have their own potential problems.

• Observer bias:

  • Observer’s expectations change their interpretation of participants’ behaviors.

• Observer effects:

  • Observer’s expectations change the participants’ behaviors.

• Reactivity:

  • Participants’ awareness of observation changes their behavior.

Group Research

• Decide on one frequency question about the topic.

• Come up with a research strategy.

  • Come up with a way of measuring the topic.

  • This could be either self-report or observational

  • What’s your sampling strategy?

  • What are your materials (ex., questionnaires)?

  • What’s your procedure?

  • How will you code (and share) your data?

• Collect data.

• Share and analyze data.

• Write your research report.

Writing the Results

• For your descriptive study: Provide relevant descriptive statistics ($M$ & $SD$) for your variables and interpret them—what do these numbers say about the variables? You’ll also be providing a graph of your findings