L2: Research Methods

Question 1

what is a variable?

Answer 1

• An attribute that assumes different values across people, places and timepoints
• variation in people: individual difference
• variables of interest: thoughts, feelings, and behaviours

Question 2

what are conceptual variables

Answer 2

• variables are ideas that do not exist physically – are subject to interpretation
• cannot prove that these variables are real

Question 3

how do you measure conceptual variables

Answer 3

Look for behaviors associated with conceptual variables and measure them instead ie with operational variables

Question 4

give an example on how we would measure the conceptual variable of intelligence

Answer 4

• Intelligence is defined as ‘the capacity to acquire and apply knowledge and skills’ (conceptual variable)
• Intelligence might be related to other quantities we can measure, such as cognitive abilities
• ie. IQ (intelligence quotient) test
• by taking an IQ test we can get a score of ones intelligence by measuring associated operational definitions like vocab, arithmetic, picture completion, symbol search etc.

Question 5

How do you define conceptual variables?

Answer 5

Through standardized tests or operational definitions.

Standardized tests: get operational definitions from criteria validated by the academic community

Operational definition: In studies we pick one way to operationalize a conceptual variable

• Defining conceptual variables is a compromise between validity, cost and convenience; there are no perfect measures

Question 6

what are the cons of self-report questionnaires? (4)

define positive impression management, malingering

Answer 6

1. Positive impression management (exaggerating positive traits) and malingering (exaggerating/manufacturing problems)
2. Framing (wording of the question matters)
   - would you save 200 people or kill 600
3. The accuracy
   a. Accuracy varies by trait and individual
   * High for extroversion, low for anxiety
   * Very low in some cases (Narcissistic personality disorder)
   * we may say we are always anxious, when we actually have no anxiety etc.

b. Accuracy greater for rating others (but not perfect)
* Halo and horns effect have been observed
* halo = positive traits associated if they are a good person, opposite for horns
* we are better at judging other than ourself

Question 7

what are the three things we need for a test to be useful?

define Test-retest, inter-rater reliability, construct validity

Answer 7

Test-retest reliability:
Every time we do the test, we get a similar result

Inter-rater reliability:
No matter who is scoring the test, we get a similar result

Construct validity:
The degree to which a construct (i.e. our test) measures what it claims to be measuring. The test should measure what we are looking for.

Question 8

what is the key distinction between operational definition and construct validity?

Answer 8

Operational definition = what specific test you are applying to measure the conceptual variable

Construct validity = does your test result actually predict the real-world behaviors linked to the conceptual variable (not a given) does it measure what it claims to measure?

Question 9

what are the measures of central tendency and variability?

Answer 9

Measures of central tendency: mean, median and mode

Measures of variability: range and standard deviation (SD)

Question 10

what is mean?

Answer 10

the average. perhaps the most useful psychological measure

found by computing the sum of all the scores by the number of cases.

for IQ mean = 100 (in randomly selected pop)

Question 11

what is the standard deviation

Answer 11

A measure of how much a score in the population typically deviates (±) from the mean
* High SD means a lot of spread around the mean
* Low SD means little spread around the mean

for iq SD= 15
68% are within one SD

Question 12

what are outliers and how do they relate to the mean?

Answer 12

Outliers: scores very far from the mean

The mean is sensitive to extreme scores (outliers) – particularly if the sample is small
- can change the mean drastically
- Outliers lead to nonrepresentative means and limit the usefulness of means

Identifying SD: varies; generally a score must be at least 2 absolute SD units away from the mean

For IQ (M = 100, SD = 15), outliers might have:
* IQ ≤ 70 (2SD below 100) or IQ ≥ 130 (2SD above 100)

Question 13

what is the median?

Answer 13

the value separating the higher half of a population from the lower half

• the median is not influenced by extreme scores/outliers
• lowes score could be 0.001 and the highest may be 100000 but the median would not change
• in analysis of salary- the median is often used
• mean is rarely equal to median

Question 14

what is the mode?

Answer 14

the most frequent occurring value
- no formal analysis of the mode

Question 15

what is the range?

Answer 15

• the distance between highest and lowest score
• large ranges are possible but rare
• grades could vary 0-100, but 0s and 100s are rare
• thus use standard deviation
• large ranges are misleading, outliers effect
• most scores might be between 40-70, but if one kid gets a 99, the range is 40-99 which is not representative

Question 16

explain normal distribution curves in psychology

Answer 16

For any variable in a population there will be a distribution.

For each IQ score (x axis), there will be a number of people with that score (y axis)
- a normal distribution (bell curve) can only be found with a randomly selected population

• Symmetrical, bell-shaped (also called Bell Curve)
• 68% of cases between ± 1SD, 95% between ± 2SD
• No skewness
• Limited kurtosis
• Mean = Median = Mode
• most preferred stats are based upon the assumption of a normal distribution

Question 17

explain skewed distributions

Answer 17

Negatively skewed: vertex shifted to the right, outliers on the left. mean is smaller than median and mode

normal distribution: bell curve. mean, median, and mode align

positively skewed: vertex shifted to the left, outliers on right. mean is larger than the median and mode.

Question 18

what is an evidence-based theory

Answer 18

an evidence based theory is where in a study, we collect data on multiple variables so that we may determine: 1) if the variables are related and 2) the nature of their relationship.

we collect evidence to support or not support a theory

Question 19

what is a hypothesis

Answer 19

• an explanation of a phenomenon made from preliminary evidence
• A good hypothesis must be simple, clear and testable. falsifiable

Question 20

How do we test a hypothesis?

Answer 20

• control group against variations and experimental groups
• compare the scores

most studies go through revision because the hypothesis is not supported – thats okay

Question 21

what are the three research approaches

Answer 21

1. descriptive research
2. correlational analysis
3. experimental research

Question 22

explain descriptive research

and three subtypes

Answer 22

assessment through systemic observation

cannot infer causation – no manipulation just observation

There are three main subtypes of descriptive research:
* Case studies
* Surveys
* Naturalistic observation

Question 23

explain what we find through a descriptive study subtype: case studies

what is a major case study?

Answer 23

• An intensive examination of one individual
• Can give valuable insight into rare phenomena, providing proof of existence (that something can happen, even if improbable)
• Can inspire new hypothesis
• Common in medicine; basis of Freud’s theories
• Though useful, difficult to generalize to large populations

famous case study:
- H.M: role of hippocampus in memory

Question 24

explain what we find through a descriptive study subtype: surveys

example?

Answer 24

• Record data on a variable (or many variables) in a large population via questionnaires or interviews
• Can be highly generalizable
• Surveys are particularly valuable in social psychology and sociology
• Many famous examples; Kinsey’s survey on sexual behavior (1948, 1953) among them (taboo subject to talk about verbally at the time, so it was done via a questionaire)

Question 25

explain what we find through a descriptive study subtype: naturalistic observation

example?

Answer 25

• Observation of an animal in its natural setting without direct intervention
• Highly generalizable (external/ecological validity); avoids concerns about the observer effect (if covert)

Several flaws because you cannot intervene:
- Poorly controlled
- Limited range of variables can be assessed
- Difficult to study infrequent behaviors and thoughts

example: darwin and the origin of species

Question 26

research subtype 2: explain correlational analysis

Answer 26

• measure of the relationship strength between two variables (X, Y)
• Technically an analytic technique, not a research design - mathematical

The statistical measure of correlation is termed the
correlation coefficient
* Pearson r is the coefficient for continuous variables
* r values range from 0 – 1. The higher the absolute r value (positive or negative), the stronger the relationship!
* Suitable for linear relationships

• You can do a correlation between any two variables in one population (e.g. alcohol and grades)

Question 27

what is one con about correlations

Answer 27

correlation does not equal causation
- show that they are associated but cannot prove any cause

Question 28

why are strong correlations misleading? explain three reasons

Answer 28

1. non-linearity (very correlated but not linear – correlation equation won’t work)
2. third factor effects (variable C causes A + B)
   - associating two variables together, through a third
   - example: ice cream consumption causes criminal activity
   false…but…
   - greater temp = inc. ice cream consumption
   - greater temp = inc. criminal activity
   - thus, False assumption to equate ice cream consumption and criminal activity
3. spurious (random) associations (raised by chance)

Question 29

Research subtype 3: Experimental research

what is the methodology of an experiment and what is unique about it + explain the three types of variables

Answer 29

• An experiment is a controlled environment wherein we study the relationship between a defined set of variables by controlling all the rest
• Allows for us to infer causal relationships between variables (only method to show us casual relationships!)

Variables:
Independent variable/IV: Manipulate at least one variable
Dependent variable/DV: Measure at least one variable
Controlled variables/CV: All other variables are kept constant if possible

Question 30

why is random assignment important in an experimental method

Answer 30

Random assignment will eliminate pre-exisiting differences in a group of people.

• if we want to test how stress effects grades, we want to eliminate all other variables beside stress (like study time, IQ, background in subject etc.)
• to do this we must use random assignment

Question 31

what is a quasi experiment

Answer 31

the experimental and controlled group occur naturally. for example: studying how MDD in stressful environments can have an effect on cognition.
The experimental and controlled group are people with and without MDD – this occurs naturally and cannot be randomly assigned. then the stress and controlled environments are created.

in a sense there are two IVs.

Question 32

what are the three types of study designs (will be tested - like 8 questions)

Answer 32

between subject designs: 2+ groups, each given a different treatment (group 1 - control, and group 2 - given drug X)

Within-subject Design: One group observed 2+ times (e.g. before given drug Z and after given drug Z)
- problem: attrition - loose some of your participants over time
- the control group is the group before treatment

Mixed Design (Between and Within Factors):
2+ groups, each observed 2+ times (group 1 @ T1 without training and then group 1 @ T2 also without training. then group 2 @ T1 with meditation training and then group 2 @ T2 also with meditation training)

Question 33

what is the pro of these within-subject designs and mixed designs

Answer 33

control for confounding variables (as subjects are compared to themselves)

Question 34

list the cons in experimental designs (6)

Answer 34

• Sampling
• Confounding variables
• Demand Characteristics (acting for what the study wants)
• Observer Effect (the act of the researching looking)
• The Placebo Effect (act differently because you thought you got the drug)
• The Experimenter Effect (acting for what the experimenter wants)

Question 35

explain con in experimental designs: why are small samples bad?

Answer 35

• not generalizable
• To generalize results of any one study to other contexts, we need to make sure our study’s sample is representative of that context. typically the lab sample should represent the world.

Question 36

what is the W.E.I.R.D problem and why does it highlight representative sampling

Answer 36

Studies focus on Western, Educated individuals living in Industrialized, Rich + Democratic countries

• Non-WEIRD communities may differ in many ways
• Morals, reasoning, fairness, visual perception + more
• While the WEIRD problem is a fault which should be acknowledged, this does not mean all WEIRD studies are wrong or useless
• If we threw out every imperfect study, we’d have none left

representative sampling is crucial because if we study people in a similar group (like a WEIRD group) we can get varying or similar answers. must evaluate all communities.

Question 37

explain con in experimental designs: confounding variables

Answer 37

confounding variable: An uncontrolled variable that is related to the DV and/or IVs

• An experiment is a carefully regulated environment where most variables are controlled
• It is the control of other variables that allows researchers to infer causal relationships
• Only IV changed, so the IV caused the difference in the DV
• Variables that a researcher fails to control can limit the usefulness of a study

Question 38

explain con in experimental designs: the act of observation

when/how can this con be prevented

demand characteristics and observer/hawthrone effect

Answer 38

demand characteristics: can be limited by opacity and/or deception by the researcher. you act based on what the experimenter wants

observer/hawthorne effect:
difficult to prevent in most cases, through less of an issue for naturalistic observation studies (you act differently when being observed)

Question 39

explain con in experimental designs: the power of expectation (explain placebo effect relation)

Answer 39

• People generally have pre-existing beliefs about how treatments might affect them
• You might believe certain treatments are effective and some are ineffective and harmful
• The results of a study might be affected (or in some cases determined) by the subject’s belief in this treatment (subject bias comes in the way)

essentially, the placebo effect
- If someone responds to a placebo, it is likely because of expectation/beliefs (because there are no active ingredients in the placebo)
- the placebos can work even if you dont know they are placebos
- associated with neurophysiological changes

Question 40

what is the pure treatment and the pure placebo effect.

Answer 40

• Pure treatment effect: the placebo and control treatment should give the same results if participants don’t have expectation
• data:
  drug> control
  drug> placebo
  placebo = control
• Pure placebo effect: the placebo equates to the effect of the drug
• data:
  placebo > control
  drug > control
  placebo = drug (this is shows the power of expectation)

Question 41

what is single blinding or participant blinding. and what is double blinding

Answer 41

single blinding = Groups do not know which treatment they are getting (placebo or treatment)

double blinding = groups and researchers do not know which treatment anyone is getting – the computer records the groups

Question 42

explain con in experimental designs: researchers

Answer 42

• Researchers, just like participants, can affect the results of a study without intending to
• If a researcher wants a certain result, they might unconsciously behave in a way that gets that result (experimenter effect - bias)
• For this reason, it is preferred that researchers and participants are blinded (double-blinding)
• Requisite for clinical studies