Chapter 2 Research Methods Copyright Allyn Bacon 2007
Chapter 2 Research Methods Copyright © Allyn & Bacon 2007
The Monty Hall Problem I present three envelopes labeled A, B and C. One of them contains a dollar. The contestant, Sally, chooses envelope A. I then tell the contestant that the dollar was NOT in envelope B. Then I offer Sally a chance to switch to envelope C. Should she? Copyright © Allyn & Bacon 2007
Let’s try it together On your paper, record the outcome of our trials. Copyright © Allyn & Bacon 2007
Results After our trials, what is the general rule you should follow: stay or switch? Let’s watch a short video explaining why… Copyright © Allyn & Bacon 2007
The Monty Hall Problem Copyright © Allyn & Bacon 2007
Implications Common knowledge and assumptions are weak. Things that we “know” are true still need to be tested in the sciences. Copyright © Allyn & Bacon 2007
Did We Know It All Along? Hindsight Bias • Hindsight Bias – “I knew it all along”
The Scientific Attitude: Curious, Skeptical and Humble • Three main components – Curious eagerness – Skeptically scrutinize competing ideas – Open-minded humility before nature • Hindsight bias, overconfidence and our tendency to perceive patterns in random events often lead us to overestimate our intuition.
You want patterns… Copyright © Allyn & Bacon 2007
Types of Studies • Case study: one person is studied in depth in the hope of revealing universal principles. • Anecdotes don’t prove anything, but suggest further study • Naturalistic observation: observing and recording behavior in naturally occurring situations • Describes behavior; does not explain behavior Copyright © Allyn & Bacon 2007
Other types of research • Survey [asking questions] • Must make sure questions are worded accurately • 27% approved of “government censorship” of television, but 66% approved of “more restrictions on what is shown on television” • Must make sure population is randomly sampled • People may answer unreliably Copyright © Allyn & Bacon 2007
Observational studies • Longitudinal study: a sample of the population is studied at intervals to examine changes/effects— often many decades (difficult and time consuming) • Longitudinal studies allow scientists to distinguish between short-term and long-term phenomena, such as poverty. If the poverty rate is 20% at a point in time, this may mean that 20% of the population is always poor, or that the whole population experiences poverty for 20% of the time. It is not possible to conclude which of these possibilities is the case using one-off cross-sectional studies. • See the 56 UP trailer here • Read about the amazing Harvard study here Copyright © Allyn & Bacon 2007
Observational studies Cross-sectional: Sampling groups at different points along a development path e. g. Surveying people every 10 years is ______. Surveying 20 year olds, 30 year olds and 40 year olds at the same time is ______. Cohort Sequential: a combination of both longitudinal and cross-sectional Copyright © Allyn & Bacon 2007
Other types of studies Non-experimental methods include: • Correlational studies • How well does A predict B…and vice versa • measures whether factors vary together • Does NOT prove cause and effect **CORRELATION DOES NOT EQUAL CAUSATION Question: there is a correlation between depression and low self-esteem. Does depression cause low self esteem? Copyright © Allyn & Bacon 2007
Copyright © Allyn & Bacon 2007
Correlations: A relationship between Two Variables Correlation– A relationship between two variables, in which changes in one variable are reflected in changes in the other variable Correlation coefficient–(r) A number between -1 and +1 expressing the degree of relationship between two variables. r Copyright © Allyn & Bacon 2007
Correlation
Correlation
Correlation
Correlation
Correlation
Correlation
Correlation
Correlation
Correlation
Illusory correlations Spurious correlations Copyright © Allyn & Bacon 2007
What type of studies are these? 1. 2. 3. 4. 5. 6. Researchers study a teen with schizophrenia to determine what sort of difficulties people like him have. Researchers follow 20 graduates of LC, taking measurements of mental health every five years. Researchers ask 300 randomly selected people how they cope with anxiety. Researchers observe and describe behavior of music fans waiting in line for entrance to a concert. Researchers test to see if body mass index and happiness are connected. Researchers randomly select 150 people to test for anxiety. Using random assignment, researchers teach 50 people relaxation techniques, teach 50 people fake relaxation techniques, and do nothing to 50 people, then test all the subjects to see if relaxation techniques work. Copyright © Allyn & Bacon 2007
How Do Psychologists Develop New Knowledge? Empirical investigation – A reliance on sensory experience and observation as research data Scientific method – A five-step process for empirical investigation of a hypothesis under conditions designed to control biases and subjective judgments Copyright © Allyn & Bacon 2007
Copyright © Allyn & Bacon 2007
Types of Psychological Research Experiment: an investigator manipulates one factor (IV) to observe the effect on another factor (DV). In experiments, the researcher controls all the conditions and manipulates the conditions. Thus, only an experiment can isolate cause and effect. Copyright © Allyn & Bacon 2007
The Five Steps of the Scientific Method Developing a hypothesis Performing a controlled test Hypothesis – A statement predicting the outcome of a scientific study—must be falsifiable Gathering objective data Analyzing the results Publishing, criticizing, and replicating the results Operational definitions – a statement of how a researcher will measure behaviors—makes behavior measurable Copyright © Allyn & Bacon 2007
The Five Steps of the Scientific Method Developing a hypothesis Performing a controlled test Gathering objective data Analyzing the results Publishing, criticizing, and replicating the results Independent variable – The variable manipulated by the experimenter Controlled variables – Variables that are held constant in order to analyze the relationship between other variables without interference. Copyright © Allyn & Bacon 2007
Things a controlled experiment must have Randomization: Random sampling: subjects are chosen at random from the population (population: ALL people in a group being studied) [avoids sampling bias] Random assignment: subjects are assigned to experimental (receives treatment), control (receives no treatment) and placebo (receives inert treatment) group randomly Placebo Effect: experimental results caused by expectation alone; any effect on behavior caused by an inert substance, which the recipient assumes is an active agent. See the placebo effect in action (23: 40) Copyright © Allyn & Bacon 2007
Things a controlled test must have Blind experiment: information is withheld from subject Double blind: information is withheld from subject and experimenter Control of confounding variables: things that have an effect on the dependent variable, but were not taken into account in the experimental design. Researchers MUST make every attempt to control them. Copyright © Allyn & Bacon 2007
The Five Steps of the Scientific Method Developing a hypothesis Performing a controlled test Gathering objective data Analyzing the Dependent variable – results The measured outcome of a Publishing, study criticizing, and replicating the results Copyright © Allyn & Bacon 2007
The Five Steps of the Scientific Method Developing a hypothesis Performing a controlled test Gathering objective data Analyzing the results Publishing, criticizing, and replicating the results Based on statistical analyses of results, the hypothesis is accepted or rejected Copyright © Allyn & Bacon 2007
The Five Steps of the Scientific Method Developing a hypothesis Performing a controlled test Gathering objective data Analyzing the results Publishing, criticizing, and replicating the results Research must withstand the scrutiny of the scientific community Copyright © Allyn & Bacon 2007
Hypothesis: Excessive noise causes acute anxiety How do we design this experiment? What is the independent variable? The dependent variable? Our operational definitions? What are some confounding variables? Copyright © Allyn & Bacon 2007
Experiment time! I will need three volunteers Now that we have done the experiment… Let’s look at the results Copyright © Allyn & Bacon 2007
Review What do you think the hypothesis of the fidget study was? What was the IV? DV? What were possible confounding variables? eyes closed/eyes open; lack of training; imprecision of counting How could I have improved this study? -operational definitions -videotape Copyright © Allyn & Bacon 2007
Let’s analyze experiment design I will give you a handout with three design scenarios. Do the first one with an aisle partner, then let’s discuss. Copyright © Allyn & Bacon 2007
Organizing the Data First results must be arranged in a chart known as a frequency distribution, which simply shows numerically how frequently items occur (put the numbers in order). We can convert the data into a bar graph called a histogram Let’s look at a histogram from the survey you submitted—for instance, scores for chapter 14 test. Copyright © Allyn & Bacon 2007
Normal distribution with histogram Copyright © Allyn & Bacon 2007
Describing the Data With Descriptive Statistics Descriptive statistics include: The range: the difference between the highest and lowest score Measures of Central Tendency • The mean: the average; affected by extreme scores • The median: the middle score; resistant to extreme scores • The mode: the most frequent score Copyright © Allyn & Bacon 2007
Quiz scores: Set A: 1, 1, 1, 7, 10 Set B: 1, 3, 5, 5, 6 what is the mean? The median? The mode? The range? Notice how the mean doesn’t tell the whole story Copyright © Allyn & Bacon 2007
Normal distribution=bell curve Standard deviation: shows the average difference between a score and the mean Copyright © Allyn & Bacon 2007
Calculating Standard Deviation Copyright © Allyn & Bacon 2007
Survey Now let’s make some predictions about the survey. Now let’s look at the data. Copyright © Allyn & Bacon 2007
Looking at statistical results Previously we have looked at descriptive statistics. Now let’s look at an inferential statistic. P value: likelihood that results are due to chance. P<. 05 is good. Copyright © Allyn & Bacon 2007
Statistical Significance Copyright © Allyn & Bacon 2007
Ethics of research on animals Two main issues: 1) Is it right to place the well-being of humans above that of animals? --Ok to kill animals/make animals suffer in order to find cure for human disease? 2) If human life has first priority, what safeguards should be in place to protect animals in research? --Care for them? Housing? Copyright © Allyn & Bacon 2007
Ethics in Research psychologists must: 1. Obtain informed consent a. Outline risks and benefits; must be voluntary b. Sample informed consent form 2. Protect subjects from harm & discomfort 3. Keep subjects’ identity confidential 4. Fully debrief afterward Copyright © Allyn & Bacon 2007
- Slides: 52