Você está na página 1de 5

Psychology Research Methods and Statistics

Learning outcomes

By the end of this session you should have the ability to:

Lecture 4: Variables Reliability and Validity Scales of Measurement Dr Ed Walford Room 509c
e.walford@aston.ac.uk

Define constructs and variables and describe their differences Define and identify independent and dependent variables Define error and identify potential sources of error Define and describe reliability and validity

Constructs
Theoretical concepts formulated (i.e. constructed) to serve as causal or descriptive explanations e.g. Psychosis: a mental condition characterised by a misapprehension and misinterpretation of the nature of reality (Gregory, 1987, p. 657)

Variables

Any characteristic that can assume multiple values or can vary


e.g. age, gender, body weight, alcohol consumption, occupation, test score etc, etc.

An event or condition the researcher observes or measures Variables must be operational


i.e. explicitly stated

Constructs vs. Variables


Constructs defined by theoretical definitions e.g.: Intelligence: The capacity to acquire and apply knowledge; the faculty of thought and reason Variables defined by operational definitions e.g.: score on a standardised test of intelligence (for example, the Wechsler Intelligence Scale)

Examples of Constructs/Variables
T: Schizophrenia is genetically determined
C: disintegration of personality expressed as disorder of feeling, thought and conduct V: A score of 30 or lower on MMPI personality scale with identified disorientation of thought and speech processes

T: Lecturing improves student knowledge


C: acquaintance with or understanding of course material V: retention of material as assessed by an exam etc.

More examples 1
Dave Gorman questioned the proposal that by following our horoscopes we can lead happier lives. He spent 40 days following the precise instructions of astrologers, while his twin brother ignored the instructions of astrologers. The happiness of each twin was measured. Constructs? Variables?

More Examples 2
The biscuit study from last week

Constructs? Variables?

Independent variables: Levels


Independent vs. Dependent Variables Independent IV:
the variable that is manipulated or in some other way is hypothesised to bring out the change in the dependent variable. Also called conditions (the grouping variable) Independent variables each have at least 2 levels that are typically compared with one another in a given study

Quasi-independent variables

Dependent DV:
the variable upon which the study depends. Usually the variable that is measured (the outcome variable)

True experimental variables

e.g. gender, handedness, age group: these characteristics cant be randomly assigned! e.g. drug treatment versus counselling, before versus after training: you can randomly assign people to groups

Independent vs. Dependent Variables

Independent vs. Dependent Variables

IV: Number of lectures attended in the course (high lecture attendees/low lecture attendees) DV: Final mark in the course

H: Following the instructions of astrologers will lead to greater happiness


IV: Following/ignoring the instructions of astrologers DV: wealth, health and love measures

H: The presence of an audience will improve performance IV? Levels? DV? Quantitative? Qualitative?

Identify the IV and DV


A psychologist is interested in the effect of sleep deprivation on reaction time. She assigned each of 60 psychology students to one of four conditions for which the students were deprived of sleep for: 0 days, 1 day, 2 days, 3 days. She then measured the average reaction time of students in each condition. IV? Levels? DV?

Identify the IV and DV


A developmental psychologist was interested in studying the amount of verbal behaviour displayed by young children, depending on whether or not an adult is present. Children who were either 2, 3, or 4, years old were selected, and observed in a lab for 30 minutes. Half had an adult present, the other half were accompanied by another child. The psychologist measured the number, duration and complexity of the utterances made by each child. IV? Levels? DV?

How good are my measures?


The data gathered to test a hypothesis are only as good as the measures that were used to obtain them. Accuracy vs. Precision Reliability vs. Validity

Accuracy vs. Precision


Accuracy: the difference between what an instrument says is true vs. what is actually true Precision: the quality or exactness of the measure.

Error
There are two kinds of error associated with measurement: Random Constant or systematic (bias) Confounding variables Variables which act on the relationship between the IV and DV and introduce constant error

What is the confound?


You set out to determine the effect of mobile phone use on hazard awareness while driving. In the morning you ask participants to identify potential hazards during a driving simulation with no distracter. In the afternoon you ask the same participants to repeat the task while talking on a mobile phone.

Reliability
The stability of our measures.
Do you get the same value across a series of measurements? For example, consider the reliability of: Measures of height Measures of reaction time Measures of attitude

Evaluating Reliability 1
Test-retest reliability: measures fluctuations from one time to another give measure to same group of people on two different occasions Beware of practise effects! Split-half reliability give half the questions to the group and the other half on another occasion

Evaluating Reliability 2
Alternate forms reliability give different versions of the same test to the group on different occasions Inter-rater reliability: measures fluctuations between observers Ask two different observers to measure the variable of interest and compare their findings

But..reliable measures are not necessarily valid ones!

The duration a person can keep a ball in the air may be measured reliably; but this does not mean that it is a good measure of intelligence.

Validity
Refers to the truthfulness of the measure.
does the test or instrument measure what it claims to measure?

Forms of Validity 1

Psychological phenomena (e.g. attitudes) do not allow objective measurement in the way that physical phenomena do
e.g. we cant gather direct evidence of intelligence, we need to imply intelligence on the basis of indirect measures

Face Validity: Does it look like a good test? Criterion Validity: Does the measure give results which are in agreement with other measures of the same thing?
Concurrent: comparison of new test with established test Predictive: does the test predict some future attempt

Construct Validity: How well does the test tap into the underlying construct?

Our measures may not be valid

Forms of Validity 2
Ecological Validity: does the study reflect naturally occurring behaviour? Population Validity: is our sample adequate for the claims we make about the population?
e.g. if we measured attitudes to recreational drug use from a sample in this lecture, could we generalise the results to conclude the attitudes of everyone/young people/students/Aston students/ psychology students

Issues of reliability and validity


The music study! There will be questions on this video in Week 4 Quiz.

Housekeeping
Week 4 quiz deadline: Friday 6th November @ 6pm Week 2 quiz feedback available now
Recommended Reading (Langdridge, 2004) Chapter 2: pp21-26 Chapter 3: pp31-37
Reliability Validity Variables Error

covering material from weeks 3+4

Você também pode gostar