Unit 2 Research methods

Research Methods (Modules 4-5)

Learning Goal 2.1

• Discuss the need for psychological research.

• Identify, compare, and contrast different research methods.

Personality Profile #1

• Great need for others to like and admire you, but tendency to be self-critical.

• Strive to appear disciplined and self-controlled but are often worrisome and insecure.

• At times have serious doubts about decisions made.

• Found it unwise to be too frank.

• Sometimes extroverted, affable, and sociable; other times introverted, wary, and reserved.

• Prefer change and variety, become dissatisfied with restrictions.

• Pride self as an independent thinker, require proof.

• Some aspirations may be unrealistic.

• Have unused capacity not fully turned to advantage.

• Generally able to compensate for personality weaknesses.

• Security is a major goal.

The Barnum Effect (Forer Effect)

• People tend to believe descriptions of themselves are true and insightful, even when the description is general and vague.

Math in Scientific Research

• Lily pad problem:

  • A patch of lily pads doubles in size each day.

  • It takes 48 days to cover the entire lake.

  • How long to cover half the lake?

  • Answer: 47 days

• Bat and Ball Problem

  • A bat and a ball cost 1.10 in total.

  • The bat costs 1.00 more than the ball.

  • How much does the ball cost?

  • Answer: The ball costs 0.05. The bat costs 1.05. Together, they total 1.10.

Why We Need Psychological Science

• Hindsight bias: Believing you could have foreseen something after it happened ("I knew it all along").

• Overconfidence: Being more confident than accuracy merit.. How quickly can you unscramble a 5 letter word?: WREAT TREYN GRABE OCHSA

• Finding patterns in random events: Trying to make sense of the world by seeing predictable patterns where there are none

• Confirmation bias: (Remember from unit 5?)

Table 4.1 True or False?

• Psychological research will confirm or refute these statements:

  1. If you want to teach a habit that persists, reward the desired behavior every time, not just intermittently (see Module 27).

  2. Patients whose brains are surgically split down the middle survive and function much as they did before the surgery (see Module 13).

  3. Traumatic experiences are typically "repressed" from memory (see Module 33).

  4. Most abused children do not become abusive adults (see Module 50).

  5. Most infants recognize their own reflection in a mirror by the end of their first year (see Module 47).

  6. Adopted siblings usually do not develop similar personalities (see Module 14).

  7. Fears of harmless objects are just as easy to acquire as fears of dangerous objects (see Module 15).

  8. Lie detection tests often lie (see Module 41).

  9. The brain remains active during sleep (see Modules 22-23).

An Interesting Finding

• Marriages with firstborn daughters are more likely to result in divorce than those with firstborn sons.

• What does this mean?

Correlation vs. Causation

• Correlation = when two factors vary together.

• Causation = when one factor CAUSES another thing to change.

• Correlation does NOT equal causation!

  • Just because A and B tend to happen together, it doesn’t guarantee that A caused B.

    • Maybe B caused A.

    • Maybe C caused both A and B.

    • Maybe there’s some other random connection.

• Example:

  • People who wear scarves also drink more hot cocoa.

    • (a) scarves cause a desire for hot cocoa,

    • (b) hot cocoa causes people to want to wear scarves, or

    • (c) both scarf-wearing and hot cocoa cravings are caused by cold weather.

  • Scarves and hot cocoa are correlated, but not necessarily causal.

• Revisiting the finding

  • Divorce Daughters Marital Stress

  • Divorce Daughters (female embryos are more likely than male embryos to survive maternal stress)

Spurious Correlations

• Number of people who drowned by falling into a pool correlates with Films Nicolas Cage appeared in (tylervigen.com).

• Letters in Winning Word of Scripps National Spelling Bee correlates with Number of people killed by venomous spiders (tylervigen.com).

7.1 Descriptive Methods (continued)

• Please get out your 7.1 HW (you will have some blanks because we’re finishing in-class notes today.)

• Retakes begin today! If you were unable to secure a spot, let me know.

• HW due next class can be used on a HW quiz (Wed)

Three Major Types of Studies

• Descriptive (reports and describes behaviors).

• Correlational (looks for relationships between 2+ factors or variables).

• Experimental (investigates cause and effect, manipulate variables to discover their effects).

The Four Major Research Methods

• HOW the study is conducted

  • Survey (descriptive, correlational).

  • Naturalistic Observation (descriptive, correlational).

  • Case Study (descriptive).

  • Experiment (experimental).

Survey

• Strengths:

  • Can collect lots of data in little time.

  • Inexpensive.

  • Low risk to participants.

• Limitations:

  • Participants may misinterpret questions.

  • Participants may not self-report accurately.

  • Provides limited data.

  • Cannot establish causation.

Naturalistic Observation

• Strengths:

  • Authentic – translates well to real-world scenarios.

  • Usually low-risk to the participants.

• Limitations:

  • Lack of control over environment.

  • Danger of confirmation bias – researcher pays more attention to behaviors that support his/her expectations.

  • Danger of participant bias – participants act differently when watched.

  • Cannot establish causation.

Case Study

• Strengths:

  • May be the only ethically viable option (as in studying abuse).

  • Very holistic – can consider a variety of factors at once.

• Limitations:

  • Danger of confirmation bias.

  • Requires more time/effort.

  • Most difficult to generalize to the general population.

  • Cannot establish causation.

Experiment

• Strengths:

  • Allows the researcher to closely control the environment and inputs.

  • The only method that can establish cause-and-effect relationships.

  • Note: if you are ever asked if it’s possible that a study can conclude a cause and effect relationship between variable, your answer may be as simple as “yes, because it’s an experiment.”

• Limitations:

  • Often poses greater risk to participants.

  • Requires more time, effort, and money.

  • Can be difficult to generalize to more authentic situations.

• Method: Manipulate and control certain variables to see how people react under different conditions

Operational Definition

• A carefully worded statement of the exact procedure (operations) used to measure a variable in a research study.

• For example, human intelligence may be operationally defined as score out of 100 on an intelligence test.

Two Data Collection Techniques

• Cross-sectional Study

  • many people at one point in time.

    • Gathers lots of data efficiently and captures a wider sector of the population

    • Have to compare different people to each other

• Longitudinal Study

  • fewer people over an extended period of time

    • Allows researchers to measure changes over time

    • Takes longer; external conditions may change over time

Ethical Flaws v. Design Flaws

• Ethical Flaws

  • Ex: violates confidentiality, violates protection from harm, failed to debrief

• Design Flaws

  • Ex: IV does not have construct validity, there is no control group, the researcher did not use a random sample, DV does not have inter-rater reliability, the researcher did not use a representative sample

Proposal 1—accepted

• Accepted of Rejected? ○ Accepted (but why might we hesitate?)

• Population ○ Students

• Sample ○ 40 students who responded to the ad

• IV: ○ Stress ○ operationally defined as being told they “ failed the test and that it is surprising that they were able to do well enough in high school to even get into college”

• DV: ○ IQ Test Scores ○ Operationally defined as their score on the McCord Intelligence Test, Form B

• Design flaw: ○ There is no control group

Statistical Analysis

• Learning Goal 2.3: Apply basic statistical concepts to analyze data.

Bar Graphs

• The dependent variable, which is the responding variable, goes on y axis

• The manipulated variable, known as the independent, variable goes on x axis

• DRY MIX

• Be sure to label your axes!

DRY MIX

• D = dependent variable R = responding variable Y = graph information on the vertical or y-axis

• M = manipulated variable I = independent variable X = graph information on the horizontal or x-axis

Measures of Central Tendency

• Mean, median, mode

• All the same for normal distributions

• Mean most heavily affected by skew

Measures of Variation

• Tells us how diverse the data set is

• Range tells us the difference between the highest and lowest data points AKA score; measure of DV

  • If data points are 15, 12, 10, 20, 25, 30; the range is 20

• Standard Deviation tells us approximately how much the data points vary from the mean…

Standard Deviation

• Standard deviation measures how spread out the scores are from the average, or mean. If the standard deviation is small, it means most scores are close to the average. But if it's large, it means the scores are more spread out.

• \sigma (sigma) stands in for the symbol for standard deviation

Normal Distribution/Bell Curve

• 68% score within one SD of the mean

• 95% score within two SDs of the mean

Normal Distribution Example

• A group of 625 students has a mean age of 15.8 years with a standard deviation of 0.6 years. The ages are normally distributed.

  • a. How many students are between 15.2 and 16.4 years old?

    • 15.2 = one SD below the mean; 16.4 = one SD above

    • 68% of 625 students = 0.68 * 625 = 425 students

  • b. How many students are younger than 17 years?

    • 17 = two SDs above the mean

    • 95% are within two SDs; another 2.5% are even younger

    • 97.5% in total are younger; 0.975 * 625 = 609 students

Regression to the Mean

• Regression to the Mean is when extreme scores tend to get closer to the average when measured again. This happens because extreme scores are often due to temporary factors, not lasting characteristics.

Positive Skew

• The majority of the data clusters on the left side, with the tail extending towards the right, suggesting the presence of outliers or unusually high values.

• A positive skew has a tail that goes to the right.

Negative Skew

• The bulk of the data clusters on the right side, with the tail extending towards the left, suggesting the presence of outliers or exceptionally low values.

• A negative skew has a tail that goes to the left.

Percentile Rank

• Percentile Rank a statistical measure that indicates the percentage of scores in a distribution that are equal to or below a particular value.

• It is commonly used to compare an individual's score with those of a larger group or population.

Bimodal Distribution

• Bimodal Distribution type of data distribution in which there are two distinct peaks or high points on a histogram or frequency distribution graph.

• This indicates that the data has two modes, or values that occur most frequently.

• It suggests that the data is not evenly distributed and may represent two distinct groups or phenomena within the dataset.

• Bimodal distribution helps us understand when data doesn't just have one main group, but two distinct ones, showing us interesting patterns in the information we're looking at.

Descriptive vs. Inferential Statistics

What if there is no p value given?

• You will never be asked to calculate the p value.

• Can we determine results are statistically significant by looking at a graph?

• A general rule of thumb: If the average results of your CC and EC are 2 SDs or more apart, then the difference is statistically significant!

Participants’ Rights

• Since psychologists frequently conduct their research using human subjects, we carry an extra responsibility to ensure that our investigations honor the rights and dignity of our participants.

Institutional Review Boards (IRBs)

• Institutional Review Boards (IRBs) are committees responsible for reviewing and approving research proposals to ensure that they meet ethical standards and protect the rights and welfare of research participants.

• Federal law and American Psychology Association (APA) standards require IRB review of all human participant research projects.

• Proposal Submission: Researchers submit their research proposal to the IRB for review.

• Initial Review: The IRB conducts an initial review to ensure the proposal meets ethical guidelines and regulations.

• Ethical Assessment: The IRB evaluates the proposed research's potential risks and benefits to participants.

• Feedback and Revision: If necessary, researchers may need to revise their proposal based on feedback from the IRB.

• Final Approval: Once satisfied with the revisions, the IRB grants final approval for the research to proceed.

• Ongoing Oversight: The IRB continues to monitor the research to ensure ongoing compliance with ethical standards.

#1 Informed Consent

• Researchers are required to obtain informed consent from participants before their involvement in the study, and participants have the right to withdraw from the study at any time without penalty.

• The Tuskegee Syphilis Study is a stark example of the need for informed consent in research.

• Informed consent serves to protect the rights and well-being of research participants and is essential for ethical research conduct.

Right to Informed Consent

• Participants must know the following before consenting to participate:

  • What the study is investigating

  • What participants will be asked to do

  • The risks and benefits of participating

  • What data will be collected and how it will be used

  • Contact information for the researchers

  • That they may withdraw from the study at any time

#1 (continued) Informed Assent

• Informed Assent similar to informed consent but applies to research involving children or individuals who may have limited decision-making capacity, such as those with cognitive impairments or developmental disabilities.

• Provides these individuals with age-appropriate or understandable information about the research study, including its purpose, procedures, risks, and benefits.

Right to Informed Consent

• Why might we limit it?

  • If participants’ awareness of the study goals would significantly interfere with results AND study is minimally risky or invasive

• What should we do if we must bypass this right?

  • Ensure that risk of harm is minimal

  • Debrief after the study and get consent at that time (if participants refuse, their data MUST be thrown out)

Deception

• In psychology, deception may involve misleading participants about the true purpose or nature of the study, the procedures involved, or the expected outcomes.

• While deception is sometimes necessary to achieve research objectives, it must be used sparingly and justified by compelling scientific or ethical reasons.

• In the Milgram study, participants were led to believe that they were administering electric shocks to another person as part of a learning experiment.

• This example highlights the ethical dilemma of using deception in research and underscores the importance of carefully considering its necessity and potential impact on participants.

Confederates

• Confederates are individuals who are part of a research study but are actually working in collaboration with the researcher and are aware of the true purpose of the study.

#2 Protect Participants from Harm

• emphasizes the obligation of researchers to minimize risks and ensure the well-being of research participants throughout the study process.

• Researchers must carefully assess potential risks associated with the study procedures and take appropriate steps to mitigate or minimize these risks.

• The Stanford Prison Experiment conducted by Philip Zimbardo in 1971.

• The study, which aimed to investigate the psychological effects of perceived power and authority in a simulated prison environment, quickly devolved into a situation of extreme psychological distress and abuse.

• This study serves as a stark reminder of the ethical imperative to safeguard participants' well-being in research settings, emphasizing the critical need for ethical oversight and adherence to guidelines to prevent harm.

Right to Protection from Harm or Discomfort

• Participants must not be harmed except under very specific circumstances.

• Why might we limit it?

  • If some level of harm is crucial to the research topic

• What should we do if we must bypass this right?

  • Participants MUST give informed consent.

  • Provide resources for securing treatment as needed.

#3 Confidentiality

• It requires researchers to handle sensitive data and personal information obtained during the study with the utmost discretion.

• This means that researchers must take measures to ensure that participants' identities and responses are kept confidential and protected from unauthorized access or disclosure.

• The importance of protecting individuals involved in case studies is the case of "Genie," a pseudonym given to a girl who suffered extreme neglect and abuse during her childhood.

• However, to protect her privacy and well-being, researchers took extensive measures to maintain her anonymity, using pseudonyms and ensuring that her identity was not disclosed to the public.

Right to Confidentiality

• Participants’ individual data must be kept secret. When results are published, no identifying information about the participants may be shared. Context may be shared but should be vague (e.g. “at a public school in SoCal”).

• Why might we limit it?

  • If we have reason to believe a person is in danger.

• What should we do if we must bypass this right?

  • Get informed consent beforehand.

  • Minimize the amount of information shared and the scope of the audience.

#4 Debriefing

• providing participants with comprehensive information about the true nature, purpose, and procedures of a research study immediately after their participation.

• Debriefing allows participants to reflect on their experience, ask questions, and provide feedback to researchers.

• The Little Albert experiment conducted by John B. Watson and Rosalie Rayner in 1920.

• This study lacked proper informed consent and debriefing procedures, as Albert's mother was not fully informed of the nature of the study, and Albert was not debriefed or treated for his fear after the experiment.

• This example underscores the importance of ethical considerations such as informed consent and debriefing in psychological research to prevent harm to participants.

Right to Debriefing

• Participants must receive a full explanation of the research when done, especially if deception is used.

• Why might we limit it?

  • If the study is straightforward, low-risk, and the consent form fully disclosed the nature and purpose of the study.

• What should we do if we must bypass this right?

  • Ensure that participants have contact info for researchers.

Observer Error

• Confirmation bias: The observer pays more attention to the details that support their prior theory.

  • If I expect that dog owners are happier than cat owners, I might interpret a “Mona Lisa smile” as a smile if a person is walking a dog but not if the person is holding a cat.

• Observer-expectancy effect: The observer influences the participants’ behavior.

  • If I expect that dog owners are happier than cat owners, I might smile more at dog owners than at cat owners, thereby influencing their behavior in response.

Participant Error

• Demand characteristics: The participant tries to give “good” data.

  • If participants know I’m studying the effect of pet ownership on happiness, they might try to smile more so that I get the results I want.

• Social desirability bias: The participant tries to “look good.”

  • Participants may overstate their love for their pets because they don’t want to come off as animal haters.

Correlational Research

• examines the relationship between two or more variables.

• Correlational studies do not involve manipulation of variables; instead, they focus on observing and measuring naturally occurring associations between variables.

• A study that examines the relationship between hours of sleep and academic performance among high school students.

• This correlational research suggests a relationship between sleep duration and academic performance but does not establish causation.

Third Variable Problem

• the possibility that a third, unmeasured variable may be influencing the relationship between the two variables of interest.

• The presence of a correlation does not imply a cause-and-effect, as there could be other factors at play.

• A student who is heavily involved in extracurricular activities might have less time to sleep and also less time to study, which could negatively impact their academic performance.

• Therefore, when interpreting the correlation between hours of sleep and academic performance, researchers need to consider the influence of potential third variables like extracurricular activities to ensure accurate conclusions.

Scatterplot

• a visual representation used in correlational research to display the relationship between two variables.

• Scatterplots allow researchers to visually assess the strength and direction of the correlation between the variables and identify any outliers or unusual patterns in the data.

Correlation Coefficient

• statistical measure used in correlational research to quantify the strength and direction of the relationship between two variables.

• It is represented by the symbol "r" and ranges from -1 to +1. The correlation coefficient is usually represented by the letter r.

Positive Correlation

• A correlation coefficient of +1 indicates a perfect positive relationship, one variable tends to decrease as the other variable decreases, or one variable tends to increase when the other increases.

• Variables moving in the same direction.

Negative Correlation

• A correlation coefficient of -1 indicates a perfect negative relationship, meaning that as one variable increases, the other variable decreases.

• Variables moving in opposite directions.

Independent Variable

• the variable that the researcher deliberately changes or manipulates in an experiment. The factor that researchers are testing to see if it causes changes in behavior or outcomes.

Dependent Variable

• the variable that is observed and measured for changes in an experiment. The outcome or response that may be affected by the independent variable.

Confounding Variables

• It's a variable that wasn't accounted for or controlled in the study but still affects the results.

• Confounding variables can distort the true effect of the independent variable on the dependent variable, making it difficult to determine causality.

Random Assignment

• research method used to assign participants to different groups in an experiment randomly.

• Ensures that each participant has an equal chance of being assigned to any condition, helping to minimize bias and ensure that the groups are comparable at the start of the study.

Placebo Condition

• administering the placebo to one group of participants while the other group receives the actual treatment being tested.

• In a research study, the placebo is typically given to the control group.

Single-Blind Study

• research design where participants are unaware of whether they belong to the experimental or control group, but the researchers conducting the study are aware of this information.

Double-Blind Study

• research design where both the participants and the researchers conducting the study are unaware of who belongs to the experimental or control group.

Representative Sample

• a subset of individuals selected from a larger population in such a way that it accurately reflects the demographics, characteristics, and diversity of that population.

• By including a diverse range of participants that mirrors the population's composition, researchers can minimize biases and increase the likelihood that their results are applicable to the broader group they represent.

• Convenience sampling, on the other hand, involves selecting participants based on their availability and accessibility to the researcher.

• While convenient, this method may introduce bias because it doesn't ensure that the sample is representative of the entire student population, potentially limiting the generalizability of the study's findings.

Random Sample

• Random sampling ensures that each participant in the study has an equal opportunity to be included, which helps to minimize the influence of researcher bias and increase the generalizability of the findings to the population as a whole.

Validity

• A measurement is valid if it actually measures what it claims to measure. (Examples given are for an IQ test.)

  • Construct validity: How accurately the measurement reflects the desired trait or concept.

    • Do my questions ACTUALLY measure intelligence, or do they measure education, culture, height, etc.?

  • Content validity: How thoroughly the measurement addresses all elements of the desired construct.

    • Do my questions measure ALL elements of intelligence, or just a small subset (e.g. just math or just analogies)?

  • Predictive validity: How well the measurement can predict future abilities or scores on related tests.

    • Do high scores on my test correlate with good grades in college?

Reliability

• A measurement is reliable if it gets consistent results

  • Test-retest reliability: A participant who completes the task multiple times provides similar results each time.

  • Inter-rater reliability: Multiple evaluators would all measure the outcome the same way.

  • Split-half reliability: Different halves of the same test yield similar results