Research Methods (1) : Variables and Research Design

parsimonious

Stated in the simplest possible terms.

theory

An explanation of a particular behavior or phenomenon typically based on scientific research.

hypothesis

A specific prediction about a behavior or phenomenon that can be tested in a scientific research project.

variables

Anything that can be measured or observed and that can change.

operational definitions

Refers to how you will define and measure a specific variable in your study.

construct validity

The degree to which a test measures what it is supposed to be measuring.

reification

Treating a construct as though it is something concrete.

spurious correlation

A relationship between 2 variables that occurs purely by chance.

scientific approach

Emphasis on empirical rather than intuitive processes, verifiable by observations, testable hypotheses, valid and reliable measurements, and objective reporting of results.

initial observations > theory > hypothesis > observations and measurements

(observations and measurements should challenge and support the theory)

good theory

Testable and falsifiable, predicts novel events, and is parsimonious.

falsifiability principle

* A way of demarcating science from non-science.

* For a theory to be considered scientific, it must be able to be tested and falsified.

* scientific knowledge is provisional - theres always a possibility that a future observation could refute a previously confirmed theory

scientific knowledge

Provisional, with the possibility that a future observation could refute a previously confirmed theory.

open science movement

Movement to make scientific research open and accessible to everyone, transparent, reproducible, with pre-registration, preprints, preliminary feedback, collaborations, follow-up studies, and open data and code.

disadvantage of open science movement

Misuse or misinterpretation of data.

IV

The variable that is manipulated by the experimenter (can be naturally occurring).

DV

Variable that is measured (outcome variable that changes as a result of IV).

control variable

Variables that are kept controlled and constant so they don't confound with dependent variable.

EV/Confounding variables

Other variables that might have an effect on the relationship between IV and DV (tried to control).

measurement of variables

Data obtained in research by measuring dependent variables. (The way we measure a variable depends on its properties.)

reaction times

The time it takes for a person to respond to a stimulus.

IQ

A measure of intelligence.

personality scores

Quantitative assessments of an individual's personality traits.

body fat percentage

The proportion of fat in the body compared to total body weight.

lung function

The capacity of the lungs to exchange oxygen and carbon dioxide with the blood.

nominal variables

Categories with no order. Variables that represent categories with no numerical properties.

hypothetical constructs

variables that cannot be directly observed e.g intelligence, depression, agression motivation

problems with operational definitions

- may be different definitions for one hypothetical construct (construct validity)

-operational definitions may not be meaningful simply just descriptive

-problem of reification: treating a construct although its concrete

most to least powerful in the scale of measurement

(least) nominal, ordinal, interval, ration (most)

problem with nominal variables

there may be disagreement about the number of categories neccassary e.g gender

ordinal variables (ranked data)

Rank order of cases based on value. Categorical variables that place objects/individuals in order according to the characteristic of interest. differences (intervals) between points of scale aren't necessarily equivalent. e.g race results

interval variables

Ordered scale with equal intervals but no absolute zero. Ordered scale in which the intervals between units of measurements are all equal.

(negative values are possible e.g temperature)

ratio variables

A scale with equal intervals and an absolute zero. Interval scale with an absolute zero.

- negative values are not possible

-gold standard of measurement e.g speed/time/distance

why do scales of measurement matter?

- accuracy of measurement is very important in research

- we can only use certain type of statistic with certain types of measurement

- higher levels of measurement allows us to perform more tests

- lower level can restrict our statistical methods measurement

experimental design

-researcher manipulates IV and Measures DV

- T-Test, analysis of variance

correlational design

- measures relationship between variables

-correlation, regression

categorical design

- measures nominal variables (frequency)

-chi square

what does a correlational design do?

- examines how variables are related to each other

- measures 2+ variables

- no DV's

limits of correlational design

-Correlation is not causation and we cannot infer this

-An EV may cause changes in both measured variables

-Spurious correlation: a relationship between 2 variables that occurs purely by chance.

experimental design

manipulates one of the variables systematically to see if it has a casual effect on the other variables

independant measures/groups

participants are randomly assigned to different conditions or groups

within subjects (repeated measures)

each participant takes part in all conditions

quasi experimental design

between groups design participants are not randomly allocated to different conditions of the IV

- cannot directly manipulate e.g gender, religion

ethics

Informed consent and debriefing, deception, protection from potential harm, right to withdraw, anonymity, and confidentiality.

validity

Does your study measure what you want it to.

reliability

Are your measurements consistent?

selecting participants

Sampling methods.