91 terms

Descriptive statistics

Provide an accurate summary of the important aspects of your data using as few numbers as possible

Inferential statistics

Uses to make inferences or generalizations from data collected from observing a small sample

Mean, Median, Mode

Measures of central tendency

Standard deviation, variance, range

Measures of dispersion

Person's product moment correlation coefficient

Measures of relationship

Mode

Applicable to data measured on any scales. Which one is the most common?

Median

Is not applicable to nominal scale data. Which one is in the middle

Mean

not applicable to nominal or ordinal scale data. Is known as the average. Formula looks like this M=Ex/n

M

In mean formula; meaning the value of the X's

E

In mean formula; Meaning "Sigma" = Add together all subsequent values

x

In mean formula; meaning the area where numbers are insert

n

In mean formula; meaning the number of observations

Measures of variance

How much the score in a distribution differ from the measure of central tendency

Range

Difference between biggest and smallest score

Variance

Add all deviation scores and divide by the number of deviation scores

Deviation scores

The difference between each score (X) and the mean of all scores (M)

Variance formula

s^2=∑(X-M)^2 ÷n-1; Sum of the squares of the deviations about the mean, divided by the number of scores minus 1

Standard deviation

Square root if the variance; =s=√(∑(X-M)^2÷n-1

BEDMAS

Bracket, Exponent, Division, Multiplication, Addition, Subtraction

PODMAS

Parentheses, power Of, Division, Multiplication, Addition, Subtraction

My Dear Aunt Sally

Multiplication, Division, Addition, Subtraction

Descriptive, correlational, nonexperimental research strategy

Three types of nonexperimental research

Descriptive research strategy

Not looking at the relationship between two variables but looking at the state of a variable or variables.

Data analysis for Descriptive RS

Use descriptive statistics to summarize a variable for a group of individual

Correlational research strategy

Correlation between two variables, Measurement of two or more interval or ratio scale variables across many levels

Data analysis for Correlational RS

Data are analyzed by correlation analysis

Nonexperimental research strategy

Examining a relationship between two variable by looking for a difference between two treatment condition

Data analysis for Nonexperimental RS

Measure of a single dependent variable in two or moregroups

Measurement validity

Is the measurement measuring what you wanted to measure

Experimental validity

Is the experiment answering the question it was intending to answer

Internal and External Validity

Two types of experimental validity

Internal Validity

The cause and effect is caused only by the variables that are measured

External Validity

Can you generalize this relationship between variables beyond the experiment

experimental design

high internal validity, low external validity

nonexperimental design

low internal validity, high external validity

Generalization

The process by which we translate specific results into broad and general principles

Rule one of generalization

Controlled comparison rule; Do we have internal validity

Rule two of generalization

Sampling rule; Do we have external validity

Rule three of generalization

Operational definitions rule; Precisely what principle do we generalize

Extraneous variables

Additional variables (other than IV & DV) that can influence the research study, but are not under direct investigation

Confounding variables

Uncontrolled variables that systematically change with the IV, and which could therefore systematically affect the DV

Obscuring variables

Factor which make changes in the DV hard to observe

Ineffective manipulation

Insufficient manipulation of the ID, resulting in no detectable change in the DV

Measurement error

Low measurement reliability can result in making of the effects of the IV manipulation. Poor instruments and poor training can create low reliability

Ceiling effect

DV can't go any higher

Floor effect

DV can't go any lower

observational research

The researcher observe how people normally behave

Naturalistic observation

Use a variety of approaches to collect data on range of different behaviours in different times and places

Systematic observation

Researcher makes observation of one or more specific behaviours in a particular setting

Descriptive observational research

Each individual scores on a given measure of behaviour

Correlational observation research

Each individual scored on at least two measures of behaviour, to determine if the variables are correlated

Nonexperimental observation research

A least two groups of individuals are scored on one or more behaviours, to determine if there are differences between groups

Disguised participant observation

Research interacts with the participants, but the participants do no know that they are being observed

Ethology

Branch of biology that deals with the study of behaviour

Archival research

Research using data which was not collecting by the researcher

Type of archival research data

Statistical records, survey achieves, Written and mass communication record

Statistical records

Collected by a range of public and private organizations

Survey archives

Data from previous survey are often available to researchers

Written and mass communication records

Anything else for which there are records

Content analysis

Make recording of data both objective and systematic

Case study

Examine individual instances or cases

Four stages of conducting survey research

Develop questions, organize the questions into the full survey, select participants to be targeted in the survey, administer the survey

Develop questions

Open-ended questions, restricted questions, rating scale questions

Open-ended questions

A question where the respondent can give any answer they like

Restricted questions

A question where there is a limited number of potential responses

Rating scale question

A question asking participants to respond by selecting a specific numerical value on a predetermined scale

Linker scale

A scale measuring the agreeableness/disagreeableness within a statement (1-don't agree - 5-agree)

Rating scale questions

A question where the answer is no limited to a 5 point

Nonverbal scales

A scale good for children (happy face to sad face)

Graphic rating scale

A question where the answer is answered with locating on a line

Cross-sectional design

Data is collected at one time only

Successive independent sample design

A series of cross-selection survey

Longitudinal/panel design

Same set of people are surveyed multiple times

Population

Everyone in which you are interested

Sample

Individuals who take part in the study

Sample size

Larger samples will generalize more accurately to the population

Probability (random) sample

Minimize systematic bias in the sample by randomly selecting the participants

Simple random sampling

Every time you select an individual for a sample, each person must have an equal and constant probability of selection

Systematic sampling

A type of random sampling that is similar to simple random sampling

Stratified random sampling

The sampling avoids random bias in the sample by making subgroup and doing a simple random sampling within the subgroups

Nonprobability sampling

Sample is not selected at random

Convenience sampling

A nonprobability sampling that samples the first person they can find

Quota sampling

Type of nonprobability sample that hand-pick participants to represent certain characteristic

Basic experimental design

Experiment with only two variables

Between-subject design

Different participants are assigned randomly to each of the two groups

Posttest experiment design

Get sample - Manipulate the IV - Measure the DV

Pretest-posttest experimental design

Get sample - Measure the DV - Manipulate the IV - Measure the DV

Solomon four-group design

Get sample - random assign to posttest experiment or pretest -posttest design - follow each step for the designs

Withing-subject design

All individuals experience all treatment conditions, and are measured multiple times

Counterbalancing

Split the sample randomly to two groups - both group experience the manipulation but on the opposite time

Latin square

The box with all the samples experience everything