Neuro 2001 Lecture 3-4

Front 1

Factor

Back 1

Differentiates between a set of groups being compared in an experiment

Front 2

Levels

Back 2

different values of the independent variable that are selected to create the treatment condition

Front 3

Condition

Back 3

How is the group treated in an experiment.

Front 4

Variables

Back 4

Conditions that change or have different values for different individuals

Front 5

Characteristics of a variable

Back 5

Needs to be observable. Can be measured directly or indirectly. Needs to be replicable. Can be consistently measured. Must have 2 levels/values.

Front 6

Operational Definitions

Back 6

Procedure for indirectly measuring and defining variables that cannot be observed or measured directly. Good operational definitions are clear, precisely articulated.

Front 7

Validity

Back 7

Accurate measurement

Front 8

Reliability

Back 8

Consistent measurement

Front 9

Types of variables

Back 9

Situational, Response, Participant, Mediating

Front 10

Situational Variables

Back 10

Describes characteristics of a situation or environment. What aspect of the environment changes for the subjects. Most experimentally manipulated variables are in this category.

Front 11

Response variables

Back 11

Responses or behaviours of subjects/ participants. Typically the variable that you are measuring when you manipulate the situational variable.

Front 12

Participant variables

Back 12

Differences between individuals. Constant within individuals, variable between individuals. Gender, height, genetic composition

Front 13

Mediating variables

Back 13

Help explain how and why a relationship exists between two other variables. Independent variable causes a mediating variable that then causes a dependent variable. Typically found in psychological theories, and prevention and treatment research.

Front 14

Continuous variable

Back 14

Measured along a continuum. As a whole or fractional unit. Can have infinite decimal places if desires. weight, temperature, duration of drug abuse.

Front 15

Discrete Variable

Back 15

Measured in whole units or categories that are not distributed on a continuum. MCAT score, gender

Front 16

Quantitative variables

Back 16

Varies by amount, measured in numeric units. Both discrete and continuous variables can be quantitative.

Front 17

Qualitative variables

Back 17

Varies by type or class, often a category or label for a group of events. Non-numeric. Depression, kinds of drug use, health condition.

Front 18

Independent variable

Back 18

Plotted on the x axis. Manipulated variable. Determined by the experimental design/ researcher. Known in advance. Determined by treatment conditions.

Front 19

Dependent variable

Back 19

plotted on the y axis. This is where we observer the change that was effected by the independent variable. Essentially this variable is dependent on the independent variable.

Front 20

control variable

Back 20

All other variables. that should be controlled. Potential independent variables that are held constant during an experiment.

Front 21

Confounding variable.

Back 21

Ideally none. Potentially independent variables that are not held constant because they are over looked. Confounding variables result in inability to determine what the experimental outcomes indicate.

Front 22

Reducing Error

Back 22

Error is reduced, never eliminated when the testing procedure is standardized. Sources of error are experimenter, environment, participant, and instrument.

Front 23

Sources of Error: experimenter

Back 23

Ensure experimenter behaves similarly with each subject. Animal habituation and handling. Same clothing/experimenter/script per participant.

Front 24

Sources of Error: Environmental

Back 24

Choose best conditions for testing, temperature, time of day, noise level.

Front 25

Sources of Error: Participants/ subject

Back 25

Inclusion/ exclusion criteria decided before starting the study. Animal body weight, overall health.

Front 26

Sources of Error: Instrument

Back 26

Using same instruments. Calibration.

Front 27

Scatterplot

Back 27

Shows correlation, Two events are related. Line of fit, line that closely approximates data. Correlation Coefficient, measures how well data is modelled by a linear equation.

Front 28

Nominal Scale

Back 28

Label/category. Qualitative differences in levels of a variable. Used for categorization, Measurements where numbers represent something/ someone. Resulting data indicated the proportion of subjects/ participants in each category. Resulting data indicates the proportion of subjects/ participants in each category: Gender, fur coat colour, genotype.

Front 29

Ordinal Scale

Back 29

Rank order. Measurements that convey order or rank. the numbers are meaningful but distance between numbers are not necessarily equal. Limitation is the difference between consecutive places within the rank, can be highly variable: grades A-B not the same as C-D. Rank order of athletes during a race, University ranking.

Front 30

Interval scale

Back 30

Rank order at equal intervals. Similar to ordinal scale, but intervals between adjacent values are constant. No true zero, temperature. 0 temperature is not no temperature; arbitrary zero point. Rating scale in behavioural sciences.

Front 31

Ratio Scale

Back 31

Rank order at equal intervals with zero. Similar to the interval scale but it has a zero. Here zero means the absence of a phenomenon. Most informative scale of measurement.

Front 32

Ratio vs. Interval

Back 32

The way you conceptualize your variable dictates whether or not its a ratio or an interval scale.

Front 33

Different types of validity of measurement

Back 33

Face, concurrent, predictive, construct, convergent, divergent.

Front 34

Face validity

Back 34

Looks like a good measure. Sometime useful for disguising true purpose of measurement.

Front 35

Predictive validity

Back 35

Scores obtained from a measure predict future behaviour. But can only be assessed after the experiment is done.

Front 36

Construct validity

Back 36

Scores obtained from a measure behaves exactly as everything that is known about your construct.

Front 37

Concurrent Validity

Back 37

Scores obtained using a new measure correlates with previously established measure. Previous measure is usually the gold standard. New measure is done at the same time as the gold standard. Strong positive correlation between new and old measurement.

Front 38

Convergent validity

Back 38

two measurements give strongly related scores and converge on the same construct. No emphasis on timing or gold standard. Two different methods of testing show positively related scores.

Front 39

Divergent Validity

Back 39

Same method to measure two different constructs to discriminate between the constructs. demonstrate that your measurement does not correlate with the two constructs.

Front 40

Types of reliability

Back 40

Test - retest. Inter rater reliability

Front 41

Test-retest

Back 41

determine the correlation between the scores from measurements taken at two time points.

Front 42

Inter-rater reliability

Back 42

Determine the correlation between scores from two independent scores.