PSY101 – Intro to Psych B Week 2 Notes

Chapter 2: Research Methods

Learning Objectives
 Identify two modes of thinking and their applications for scientific reasoning.
 Describe the advantages and disadvantages of using naturalistic observation,
case studies, self-report measures and surveys.
 Describe the role of correlational designs and distinguish correlation from
causation.
 Identify the components of an experiment, the potential pitfalls that can lead to
faulty conclusions and how psychologists’ control for these pitfalls.
 Explain the ethical obligations of researchers towards their research
participants.
 Describe both sides of the debate on the use of animals as research subjects.
 Identify uses of various measures of central tendency and variability.
 Explain how inferential statistics can help us to determine whether we can
generalise from our sample to the full population.
 Show how statistics can be misused for purposes of persuasion.
 Describe the differences between qualitative and quantitative data analysis.
 Identify flaws in research designs and how to correct for them.
 Identify why it is important for researchers (including students) to publish their
findings.

How we can be fooled: Two modes of thinking

System 1 thinking - 'intuitive thinking'
- First impressions, especially of other people, can sometimes be surprisingly
accurate.
Quick, reflexive thinking that requires relatively little mental effort.
'Gut feelings', 'hunches'
Often relies on heuristics

System 2 thinking - 'analytical thinking'

- Reasoning through a problem, or figuring out new concepts
- In some cases, allows us to override system 1 thinking and reject our
automatic when they seem to be wrong
Slow and reflective, takes mental effort

Heuristics: Mental shortcuts or rules of thumb

- Heuristics can cause problems when we use them too often or in

inappropriate situations leading us to oversimplify reality
QUESTION AND RESPONSE
Describe intuitive thinking and analytical thinking in your own words. Explain why
analytical thinking is essential in psychological research.

Intuitive thinking, or 'System 1 thinking' often relies on the subconscious processing

of the brain that exists when we create our 'first impressions' of people, places or
things. This quick thinking method, requires little mental effort and is not always
accurate. On the other hand, analytical thinking (or system 2 thinking) is a slow and
reflective thinking process that requires mental effort. It is often used when trying to
navigate new concepts or problems and can sometime allow us to override out
system 1 thinking when it appears to be wrong. Analytical thinking is essential in
psychological research as much of the workings of the brain remain 'clouded' and we
cannot continue to navigate new theories of why the brain works the way it does
without first considering that what we initially think or assume may be wrong.

The scientific method: a toolbox of skills

Naturalistic observation: watching behaviour in real-world settings. Tend to have

low internal validity.

External validity: the extent to which findings can be generalised to real-world

settings (applicability)
Internal validity: the extent to which we can draw cause-and-effect inferences.

Case study: where researchers examine either one person, or a small number of
people, often over an extended period of time.
Case studies can be helpful in providing existence proofs.
 Case studies rarely allow for systematic testing of alternative hypotheses
about why a given phenomenon has occurred
 This, case studies can be helpful for generating hypotheses but limited for
testing them

Existence proofs: demonstrations that a given psychological phenomenon can

occur.
 E.g. if someone claimed that 'all zebras have stripes' you would only need
ONE case of a non-striped zebra to prove the claim wrong

Time-series design: First, a well-defined series of observations is systematically

gathered and taken repeatedly over time to establish a pre-treatment baseline. Then,
the series of observation is extended across time to include the period when the
treatment is introduced and administered, as well as the post-treatment follow up.

Self-report measures and surveys: asking people about themselves and others
Self-report measures, or questionnaires can be used to assess characteristics such
as personality traits, mental illnesses and interests. Surveys are closely related and
used to measure people's opinions and attitudes.

Random selection: the key to generalisability

In random selection, every person in the population has an equal chance of being
chosen to participate and is crucial if we want our results to generalise the broader
population
 Obtaining a random sample is almost always more crucial than obtaining a
large sample if you want to generalise results to most people

Random assignment: deals with how you assign your participants after you have
already chosen them.

Evaluating measures
Reliability: consistency of measurement. A reliable questionnaire yields similar
scores over time, or test-retest reliability. Strong correlation coefficient is needed.

Inter-rater reliability: the extent to which different people who conduct an interview or
make behavioural observation, agree on the characteristics they are measuring.

Validity: the extent to which a measure assesses what it claims to measure. "truth in
advertising".

Reliability is necessary for validity, however reliability does not guarantee validity.
We need to measure something consistently before we can measure it well.

Correlation design
Research design that examines the extent to which two variables are associated.
 Allows us to generate predictions about the future

Correlations can be positive, zero or negative. A positive correlation means that as

the value of one variable goes up, the other also goes up. A zero correlation means
that the variables do not go together. A negative correlation means that when the
value of one variable goes up, the other goes down and vice versa.

Correlation coefficients range in value from -1.0 to 1.0. A correlation of -1.0 is a

perfect negative correlation and 1.0 is a perfect positive correlation.

Exceptions do not invalidate the existence of correlations!

The correlation statistic represents only the linear relationship between the two
variables.

Unless a correlation is perfect (1.0 or -1.0) there will always be exceptions to the
general trend.

Psychology is a science of exceptions

Illusory correlation: a statistical association between two variables where none
exists.

Experimental designs
When performed correctly, they permit cause-and-effect inferences.
 In correlational designs, researchers are measuring pre-existing differences in
participants (i.e. Age, gender, IQ and extraversion). These are differences
over which researchers have no control. However, in experimental designs
researchers are manipulating variables to see whether these produce
differences in participants behaviour.
For a study to be an experiment, it must have a random assignment of participants to
conditions, and the manipulation of an independent variable.

Random assignment: experimenter randomly sorts participants into one of two

groups. Often, one of these groups is called the experimental group (the group of
which the variable is manipulated) and the other is called the control group (the
baseline).

Independent variable: the variable being manipulated

Dependent variable: the variable being measured to determine whether there is an
effect

Defining independent and dependent variables for the purposes of a study, providers
an operational definition.

Operational definition: a working definition of what they are measuring.

Confounds/confounding variable: any difference between the experimental and

control groups other than the independent variable.

The placebo effect

Improvement resulting from the mere expectation of improvement.
In order for researchers to avoid being misled by the placebo effect, it is critical that
participants do not know whether they are receiving the real medication or placebo.
That is, participants must remain blind.

The nocebo effect

Harm resulting from the mere expectation of harm

The experimenter expectancy effect

When researchers' hypotheses lead them to unintentionally bias the outcome of a
study. To counter this, experiments should (where possible) be double-blind (both
participants and researchers are blind to what is being administered)

Demand characteristics
Cues that allow them to generate guesses regarding a researchers hypotheses.

Ethical Issues in research design

Debriefing
The process whereby researchers inform participants what the study was about.

The National Health and Medical Research Council (NHMRC), which oversees
ethical research standards in Australia, has developed a set of guidelines (NHRMC,
2003) for ethical research in Aboriginal and Torres Strait Islander communities.

To guide animal researchers in Australia, the NHRMC has developed the Australian
Code for the Care and Use of Animals for Scientific Purposes, which aims to
minimise the pain and distress caused by animal experimentation.

Three R's of ethical animal experimentation

 Replacement: investigators must actively develop alternative methods to
replace the use of animals wherever viable
 Reduction: experiments must use no more than the minimum number of
animals necessary to ensure statistically valid results
 Refinement: experimental procedures must be continually refined to minimise
the pain and distress caused to experimental animals

Descriptive statistics
Numerical characterisations that describe data

Mean - Average
Median - middle score in a data set ordered from smallest to largest
Mode - most frequent
Variability
Also called dispersion, which gives a sense of how loosely or tightly bunched the
scores are. The simplest measure or variability is the range (difference between
highest and lowest scores).

Standard deviation depicts dispersion, taking into account how far each data point
is from the mean rather than looking at how widely scattered the most extreme
scores are.

Qualitative analysis methods might also be considered as methods of analysing and

understanding various forms of text. Qualitative analysis methods can be applied to
any form of written text, including journal articles, newspapers, government archives,
television and radio broadcasts or the content of webpages on the internet.

Method of agreement
Comparing similar cases and identifying common concepts or themes

Method of difference
Comparing similar cases with different outcomes and seeking to identify key
differences between them.
When evaluating psychological research, apply the six principles of scientific thinking
1. Extraordinary claims - is the evidence as strong as the claim?
2. Testing predictions - can the claim be tested?
3. Occam's razor - does a simpler explanation fit the data just as well?
4. Replicability - can the results be duplicated in other studies?
5. Ruling out rival hypotheses - have important alternative explanations for the
findings been excluded?
6. Correlation vs causation - Can we be sure that A causes B?
Meta-analysis
Investigation of the consistency of patterns of results across large numbers of
studies conducted in different laboratories.

File drawer problem

Tendency for negative outcomes not to be publishes