Research Methods and Statistics
W1.1
CHAPTER 3 – THREE CLAIMS, FOUR VALIDITIES: INTERROGATION TOOLS FOR
CONSUMERS OF RESEARCH
Variable: Something that can be changed or altered, such as a characteristic or value – Must
have two levels
Constant: Something that could potentially vary, but has only one level in the study of
question
↓
E.g.: ‘15% of Americans smoke’ → ‘Nationality = Constant; ‘Smoking’ = Variable; ‘Smoker
and nonsmoker’ = Levels of the variable
Measured variable: A variable whose measures are simply observed and reported
Manipulated variable: A variable that is controlled by researchers, usually by assigning study
participants to the different levels of that variable – Some variables can only be measured and
not be manipulated because of ethical reasons, or because it is impossible to manipulate them
(such as IQ of age)
Conceptual variable/construct: Abstract concepts that must be carefully defined at the
theoretical level; the conceptual definition (E.g.: spending time socialising; school
achievement)
Operational variable: A conceptual variable that is turned into a measured or manipulated
variable (E.g.: researchers ask people how often they spend an evening alone, socialise with
friends or see relatives in a week)
Claim: An argument that someone is trying to make
- Frequency claim: Describe a particular degree of a single variable; how frequent or
common something is (E.g.: 15% of Americans smoke) – Accompanied by a margin
of error of the estimates – Observational research
- Association claim: Argues that one level of a variable is likely to be associated with a
particular level of another variable – Variables are said to correlate or to be related –
When one variable changes, the other variable tends to change too (E.g.: people with
higher incomes spend less time socialising) – Type of study involved with association
claims is a correlational study – Uses words like ‘link’, ‘correlate’, ‘predict’, ‘tie to’,
correlate’ and ‘being at risk for’ - Type I error: False positives – Type II error: False
negatives
1. Positive association
2. Negative association
3. Zero association
- Causal claim: Argues that one variable is responsible for changing the other (E.g.:
music lessons enhance IQ) – Uses words like ‘affect’, ‘change’, ‘enhance’ and ‘cause’
– Can be supported by experiments in which the independent variable is manipulated
, and the dependent variable is measured – Outcome is best ensured through random
assignment
1. Covariance: Two variables are correlated
2. Temporal precedence: The causal variable came first and the outcome variable
came later
3. Internal validity: No other explanation can exist for the relationship
Validity: Refers to the appropriateness of a conclusion – Does the measurement measure what
it is supposed to measure?
- Construct validity: How well is the conceptual variable operationalised? – Entails
reliability (= does the measurement yield similar scores on repeated tests?)
- Statistical validity: The extent to which a study’s statistical conclusions are accurate
and reasonable
- Generalisability/external validity: How well do the results of a study generalise to, or
represent, people besides those in the original study? – Important in frequency claims!
CHAPTER 6 – SURVEYS AND OBSERVATIONS: DESCRIBING WHAT PEOPLE DO
Survey: Used when people are asked about a consumer product Method of posing
questions to people
Poll: Used when people are asked about their social or political opinions
Question formats:
- Open-ended questions
- Forced-choice questions
- Likert scale (E.g.: ‘strongly agree’, ‘disagree’ etc.) → Semantic differential formats
Wording of questions:
- Leading questions: Questions that potentially could lead people to a particular
response
- Double-barrelled questions: Questions that ask two different things in one
- Negative wording: Using double denial makes questions unnecessarily difficult
- Question order
Response sets/nondifferentiation: Shortcuts respondents may take when answering a survey
- Acquiescence/yea-saying
- Fence sitting
- Socially desirable responding – Faking good / Faking bad
Observational research – Can be better than self-report measures
- Observer bias: When observers’ expectations influence their interpretation of the
participants’ behaviour or the outcome of the study
- Observer effects (E.g.: Clever Hans)
- Reactivity: When people know they are being observed, they behave differently
Can be prevented by masked design/blind
design, unobtrusive observations or
habituation
W1.1
CHAPTER 3 – THREE CLAIMS, FOUR VALIDITIES: INTERROGATION TOOLS FOR
CONSUMERS OF RESEARCH
Variable: Something that can be changed or altered, such as a characteristic or value – Must
have two levels
Constant: Something that could potentially vary, but has only one level in the study of
question
↓
E.g.: ‘15% of Americans smoke’ → ‘Nationality = Constant; ‘Smoking’ = Variable; ‘Smoker
and nonsmoker’ = Levels of the variable
Measured variable: A variable whose measures are simply observed and reported
Manipulated variable: A variable that is controlled by researchers, usually by assigning study
participants to the different levels of that variable – Some variables can only be measured and
not be manipulated because of ethical reasons, or because it is impossible to manipulate them
(such as IQ of age)
Conceptual variable/construct: Abstract concepts that must be carefully defined at the
theoretical level; the conceptual definition (E.g.: spending time socialising; school
achievement)
Operational variable: A conceptual variable that is turned into a measured or manipulated
variable (E.g.: researchers ask people how often they spend an evening alone, socialise with
friends or see relatives in a week)
Claim: An argument that someone is trying to make
- Frequency claim: Describe a particular degree of a single variable; how frequent or
common something is (E.g.: 15% of Americans smoke) – Accompanied by a margin
of error of the estimates – Observational research
- Association claim: Argues that one level of a variable is likely to be associated with a
particular level of another variable – Variables are said to correlate or to be related –
When one variable changes, the other variable tends to change too (E.g.: people with
higher incomes spend less time socialising) – Type of study involved with association
claims is a correlational study – Uses words like ‘link’, ‘correlate’, ‘predict’, ‘tie to’,
correlate’ and ‘being at risk for’ - Type I error: False positives – Type II error: False
negatives
1. Positive association
2. Negative association
3. Zero association
- Causal claim: Argues that one variable is responsible for changing the other (E.g.:
music lessons enhance IQ) – Uses words like ‘affect’, ‘change’, ‘enhance’ and ‘cause’
– Can be supported by experiments in which the independent variable is manipulated
, and the dependent variable is measured – Outcome is best ensured through random
assignment
1. Covariance: Two variables are correlated
2. Temporal precedence: The causal variable came first and the outcome variable
came later
3. Internal validity: No other explanation can exist for the relationship
Validity: Refers to the appropriateness of a conclusion – Does the measurement measure what
it is supposed to measure?
- Construct validity: How well is the conceptual variable operationalised? – Entails
reliability (= does the measurement yield similar scores on repeated tests?)
- Statistical validity: The extent to which a study’s statistical conclusions are accurate
and reasonable
- Generalisability/external validity: How well do the results of a study generalise to, or
represent, people besides those in the original study? – Important in frequency claims!
CHAPTER 6 – SURVEYS AND OBSERVATIONS: DESCRIBING WHAT PEOPLE DO
Survey: Used when people are asked about a consumer product Method of posing
questions to people
Poll: Used when people are asked about their social or political opinions
Question formats:
- Open-ended questions
- Forced-choice questions
- Likert scale (E.g.: ‘strongly agree’, ‘disagree’ etc.) → Semantic differential formats
Wording of questions:
- Leading questions: Questions that potentially could lead people to a particular
response
- Double-barrelled questions: Questions that ask two different things in one
- Negative wording: Using double denial makes questions unnecessarily difficult
- Question order
Response sets/nondifferentiation: Shortcuts respondents may take when answering a survey
- Acquiescence/yea-saying
- Fence sitting
- Socially desirable responding – Faking good / Faking bad
Observational research – Can be better than self-report measures
- Observer bias: When observers’ expectations influence their interpretation of the
participants’ behaviour or the outcome of the study
- Observer effects (E.g.: Clever Hans)
- Reactivity: When people know they are being observed, they behave differently
Can be prevented by masked design/blind
design, unobtrusive observations or
habituation
