## Some Definitions, Notes and Descriptions of Methods of Psychological Research

### Null hypothesis

1. A prediction is a specific hypothesis
2. Question a relationship between 2 or more variables
3. A suggested answer to a problem
4. It should be testable
5. Meta - analysis -- Hypothesis based on past research.
6 . Null hypothesis H0: No difference between groups
g. Alternate Hypothesis H1: There is a difference

+ H0: X1 = X2 No difference
+ H1: X1 X2 non directional difference
+ H1: X1 > X2 Directional difference

### Statistical significance, inferential and descriptive statistics

Statistical significance -- A statistically significant event is a rare event by chance, and it must be as rare or rarer than the most extreme 5%.

Differential statistics -- To describe something using a statistic, such as mode, mean or standard deviation.

A Normal distribution

#### Central tendencies

The tendency of a group of individual's scores to concentrate in one place.

1. Mean: Arithmetic average of all scores

a. Used with a normal distribution.
b. Used with interval and ratio data

Median: Point at which you can account for 50% of the cases.

a. Can be used with rank order data
b. Used with a skewed distribution

Mean and median both use interval and ratio data

Mode: The most popular category.
Most meaningful when used with nominal data.

1. Range: High - low

a. Technically a single value, (High minus Low)
b. Often reported as the high and low values.
c. Report range with mean and median.

2. Standard deviation

a. The average deviation of the scores around the mean.
b. On average, how people (scores) deviate from the mean.

#### Inferential statistics

1. Difference between something, to infer from the result of a test, to generalize from a sample to a specific population.
2. A process of inferring information from a sample to a population.

a. Use information to express a degree of confidence
b. If group is statistically significant, it must fall above the 95% range.

### Descriptive research

#### Exploratory research

Research which describes something. The descriptive research approach is an attempt to provide an accurate description or picture of a particular situation or phenomenon.

1. The descriptive research approach can be used at the initial stage of investigation to identify the variables that exist in a given situation and to describe the relationship that exist between these variables.

2. The descriptive research can also be used at the final stage of an investigation to describe the effectiveness of a proposed solution and to formulate additional hypotheses about how a more effective solution could be reached.

3. The descriptive research approach cannot be used to determine cause-and-effect relationships. The knowledge acquired through the use of the descriptive research approach can only be used to formulate hypotheses which are subject to experimental investigation.

#### The five different types of descriptive approaches

1. Naturalistic observation: A technique which enables an investigator to unobtrusively observe and record naturally occurring behavior.
2. Secondary record: A technique which enables an investigator to describe a given situation by using data that are gathered currently or previously.
3. Field study: A technique which allows an investigator to generate a description by collecting data through the use of participant observations, surveys, correlational studies, or longitudinal or cross-sectional studies.
4. Ex post facto studies: A technique which allows an investigator to describe a relationship between a given behavior and a variable on which groups of subjects naturally differ, such as skin color or gender.
5. Meta analysis: A technique which involves the quantitative analysis of the results of many studies in order to integrate and describe the overall picture presented by these studies.

Back to index

#### Correlational research

1. Involves the collection and examination of 2 or more sets of data, analyzed to determine if there is a relationship. Such as if X1 varies, X2 also varies.
2. Correlation does not imply causation, some other variable or outside influence could be the cause of the variation.
3. Useful as a first step in causal research.
4. Establishing relationships
5. Correlational

a. Describe relationship between variables
b. Hypotheses are formed to describe difference and relationships

#### Causal comparative research

1. Comparison

a. Covariation - a condition of improbable difference. Covariation - between variables such as IV (Independent Variable) and DV (Dependent Variable)
b. As the levels of the IV vary, the DV should also vary
c. Apparent difference in absolute numbers does not mean that there is a statistically significant difference.
d. The difference is an improbable difference by chance.
e. The difference is large enough to be improbable by chance.

2. Control

a. Eliminate spurious relationships or confounding factors.
b. Achieve control by manipulation
c. Random assignment and matching are used to establish control.

3. Manipulation

a. Time order
b. The IV comes before the DV
c. Manipulate the order of the events.
d. Quasi - Experimental

+ Establish relationships between variables
+ Possible causal relationships

++ Confounding variables
++Intervening variables

e. Hypotheses of difference and hypotheses of relationships

+ Usually hypotheses of difference may be directional

f. Control

+ Pre/post matched group design
+ Matching to control for moderator or confounding variables

2. Experimental design

a. Applies control through randomization

+ random selection and random assignment.

b. Random Assignment -- all subjects in the experiment have an equal chance of participating in any experimental or control group.
c. Causal relationships are being established
d. IV is manipulated

### Experimental research

1. Quasi - Experimental

a. Establish relationships between variables
b. Possible causal relationships

+ Confounding variables
+ Intervening variables

c. Hypotheses of difference and hypotheses of relationships

+ Usually hypotheses of difference may be directional

d. Control

+ Pre/post matched group design
+ Matching to control for moderator or confounding variables

2. Experimental design

a. Applies control through randomization

+ random selection and random assignment.

b. Random Assignment -- all subjects in the experiment have an equal chance of participating in any experimental or control group.
c. Causal relationships are being established
d. IV is manipulated

### Threats to internal validity

1. History
2. Maturation
3. Experimental mortality -- loss of subjects
4. Instrumentation

a. Reliability and validity of instrument or tool being used

+ Reliability -- consistency with which something is being measured over time Test - retest consistency over time
+ Internal reliability -- split half

b. Validity -- does it measure what it is supposed to measure? The concept of validity is never finally established, it is always being examined.

+ Differentiate groups it is supposed to differentiate
+ Establish relationships

- Discriminate -- different from other variables
- Convergent - similar to other variables

c. Criterion group validity -- differentiates different groups

5. Testing sensitization - the effect of pre-testing

a. Remembering particular items
b. increase due to item content importance
c. Increase due to the testing environment
d. Experience of being tested, just good at taking tests.

6. Regression toward the middle range

a. Statistical regression
b. Regression toward the mean
c. High and low scores, ceiling and floor effect
d. Watch for overall high scores
e. Natural tendency for low scores to get better
f. natural tendency for high scores to get worse

7. Selection bias

a. Self selection
b. Random assignment.

### Threats to external validity

1. Ability to generalize a specific instance to a larger population. A result or indicator of external validity.
2. Factors affecting Generalizability

a. Reactive effects of testing
b. Effects of pretesting effecting the results of the experiment, learning
c. Subjects try to "help" the experimenter by trying to do "good" or what they think the experimenter wants
d. Interaction effects of selection bias
e. If samples are not representative, cannot generalize findings
f. Reactive effects of experimental arrangements
g. The arrangements of the experiment or the experience of the participants may create sufficient artificiality to limit Generalizability.

Back to index

Experimental and control groups

The experimental group receives the treatment

The control group examines the possibility that the same thing would have happened even without the treatment.

Control group experience is identical to the experimental group except, the control group doesn't receive the treatment or intervention.

Back to index

Random sampling

Every person has an equal chance to participate in the experiment.

Direct implications for the representativeness of the sample.

Random sample increases representativeness

A scientific representative sample

1. Entire population must be accessible.
2. Each must have an equal chance of being in the sample

Back to index

Types of non-random sampling

Self assignment
Use of an intact group

Back to index

Independent variable

Independent variable -

a. Stimulus, cause, implies more than one level , necessary for comparison. Must have a label and levels.
b. May be manipulated or unmanipulated.

+ Unmanipulated - observed as is, no modification - the levels of the variable are identified in nature.
+ Manipulated - the observer determines what subjects receive what level of the independent variable.

c. Moderator variables, control variables and confounding variables are all forms of independent variables.

Moderator variable -

Variables other than those intended

+ Usually exist when the variable is unmanipulated.
+ Other factors are observable, such as SES, teachers degree, conceptual understanding can be assessed and measured.
+ If a variable may be accounted for it is a moderator variable, if not accounted for it is an intervening variable
+ A type of IV.

Back to index

Dependent variable

Dependent variable - response, consequence, effect, outcome.

Back to index

Control variable

Control variable - keep a variable constant, eliminate the effect of a moderator variable.

+ Random assignment - distribute the effects equally
+ Cancel out or neutralize factors that could affect the DV

Back to index

Intervening variable

Intervening variable - Variables which may not be accounted for

+ may be identified, but not accounted for, may not be measured.
+ possible effect which cannot be identified.
+ If a variable may be accounted for it is a moderator variable, if not accounted for it is an intervening variable
+ It's effect may be inferred from the effects of the IV (Independent Variable) and moderator variables on the DV (Dependent Variable).

Back to index

Confounding variable

a. Possible effect which cannot be identified
b. May be identified but not measured
c. If they can be accounted for, they become moderator variables.

Back to index

Operational definition

1. An operational definition is a definition based on the observable characteristics
2. Allows communication of a definition of concept, variable or observation. Should be clear and as concise as possible to enable maximum understanding.
3. Measurable assessment of observable behavior.
4 Refers to the fact that the terms used in the experiment must be defined by the steps or operations used to measure them
5. Operational definition is necessary to eliminate confusion in meaning and communication.
6. Allows communication of terms and procedures.
7. Allows replication of the experiment.

Back to index

Replication

1. Ability of an event or study to be repeated with certainty

a. Be able to confirm or disconfirm results of a previous study.
b. Test the same hypothesis using the same strategies and/or material

2. The scientific observations that are made must be able to be repeated. This means that the data obtained in an experiment must be reliable. The same conclusions should be repeated if the experiment is repeated.

Back to index

Steps in conducting research

1. Identify a problem.

a. Need a researchable question
b. Identify a researchable question, there must be sufficient information available.
c. Question should be specific, not to broad, not to vague, it should be testable.

2. Review the literature.
3. Form a hypothesis.
4. Identifying and labeling variables.
5. Constructing operational definitions.
6. Manipulating and controlling the variables.
7. Constructing a research design.
8. Identifying and constructing devices for observation and measurement.
9. Constructing questionnaires and interview schedules.
10. Conduct the experiment.

a. What are the subjects?
b. How many subjects are required?
c. What instructions are necessary to the subjects?
d. What experimental material are needed?
e. Follow the required procedure precisely.
f. Collect data.

11. Carrying out statistical analysis

a. Hypothesis testing

+ analyze the data.
+ decide what statistic to use to analyze the data.

12. Writing a research report.

Back to index

Generalizability

1. Ability to generalize a specific instance to a larger population. A result or indicator of external validity.
2. Factors affecting Generalizability

a. Reactive effects of testing
b. Effects of pretesting effecting the results of the experiment, learning
c. Subjects try to "help" the experimenter by trying to do "good" or what they think the experimenter wants
d. Interaction effects of selection bias
e. If samples are not representative, cannot generalize findings
f. Reactive effects of experimental arrangements
g. The arrangements of the experiment or the experience of the participants may create sufficient artificiality to limit Generalizability.

Back to index

Identify different parts of a research study

a. Hypothesis
b. Research question

2. Population studied, the sample

a. geographic / demographic
b. characteristics
b. How many, male / female
c Age group

3. Design

a. nature of study
b. procedures
c. instruments

4. Thrust of the finding, presented in a formal manner of standard format consisting of,

a. Abstract
b. Introduction / literature review
c. Method
d. Results
e. Discussion

+ findings.
+ weakness of results
+ Future research

Back to index

Evaluation research

Formative and summative evaluation

1. Formative is the internal evaluation of a program, usually as part of the development process. It is an attempt to debug the program.

2. Summative is the attempt to determine whether the program is fully meeting it's objectives more successfully than alternate or no program.

Back to index

Qualitative research

1. Natural setting is the data source and the researcher is the key data collection instrument.
2. It attempts to primarily describe and secondarily to analyze.
3. The concern is with the process, what has transpired, as much as the product or outcome.
4. The data are analyzed inductively as putting together the parts of a puzzle.
5. It is essentially concerned with what things mean, the "Why?" as well as the "What?"

Back to index

Ethics in research

1. Right to privacy - The right to keep from the public certain information about themselves.
2. Right to nonparticipation - The right to not participate or discontinue participation at any time.
3. Right to remain anonymous - Right to insist that individual identities remain anonymous.
4. Right to confidentiality - Right to insist that all data be treated responsibly. Right to know who will have access to the data.
5. Right to expect experimenter responsibility - The experimenter will treat each subject with human dignity, participants should not be hurt by their participation in the experiment.

a. Minimize risk and harm to subjects
b. Subjects should leave in a positive mode, they should feel good about themselves and what they have done.
c. Avoid distress to subjects.
d. State that if certain issues have caused a problem they can contact a mental health agency, give them the names/numbers of at least 2 that will not charge them for services.
e. Debriefing: Spending time with subject to explain and clarify the study and what was done, make sure that the subjects leave feeling good about themselves and the process. During the debriefing, the experimenter should; Clarify, Explain and Support.

#### To the institution conducting the research

1. Proper compensation to helpers
2. Explain who is going to get credit for project
3. Help must be allowed to quit at any time.
4. Authorship must be clarified.
5. Institution needs to be protected

#### To scientific community

1. What you do to your subjects
2. Use of data, honesty of data reporting.
3. Keep your raw data for inspection by APA and others with a right to inspect.
4. Openness, honesty, integrity of data
5. How data was obtained
6. Consequences of study should be for betterment of "humanity"

Back to index

Survey techniques

1. Mail survey

a. Lots of subjects
b. Lots of questions

+ Too many questions, people will respond less
+ Too many open ended questions, people will respond less

2. Interview

a. More expensive than mail
c. Possible threat to confidentiality
d. Smaller number of subjects

3. Telephone

a. More subjects than interview
b. Cheaper than interview
c. Cost more than mail in
d. Limited number of questions
e. Don't have chance to probe

#### Response Rate: Factors that contribute to response rate.

1. Length of survey -- shorter, better
2. Open ended questions / close ended questions -- fewer open ended, better

a. Greater response if government sponsor, people recognize that something may be made better by their response.

4. Diversity of the population, the more the diverse the population, the less the response rate.

a. More focused, better response.
b. More educated, better response.

5. Stamped self addressed envelope will increase response of mail in survey.
6. Follow up letter increases response.
7. Ease with which the receiver perceives the survey will take place increases response rate.
8. Cover letter and gimmicks, or incentives such as money, tea bag, lottery

a. Give receiver a reason to participate.
b. Better condition of humanity
c. Let them know that they are something special, they have been chosen through the scientific method to represent many other people.

Back to index

Field studies

Field studies allow us to generate a description by collecting data through the use of participant observations, correlational studies, or longitudinal, or
cross-sectional studies.

The similarity that exists between the approach of naturalistic observation and the field study approach:

1. Both types of studies are conducted in the real world.
2. Both approaches are free from the criticism of artificiality of environments in which data are collected.

Differences that exist between the approach of naturalistic observation and the approach of field study

1. When a field study is conducted, the investigator actively interacts with the subjects in the course of data collection; when naturalistic observation is conducted, the investigator remains unobtrusive.
2. The observations of interest to the investigator in a field study are typically focused on a more specific aspect of behavior than are those in a naturalistic observation.
3. Field studies use a variety of diverse approaches; Naturalistic observations use only one general procedure..

Five different approaches that fall under the classification of Field Studies are:

1. Participant observation.
2. Surveys.
3. Correlational studies.
4. Longitudinal and cross-sectional studies.
5. Field experiments.

Back to index

#### The purpose of this information.

This information was first published as an experiment in publishing on the WWW. While attending a local university, I assembled this information as a study guide for a final exam in a graduate Methods of Research class. I make no representation as to the accuracy of this information, use it with caution, however I did receive an A in the class, so obviously the information was valid and useful to me.

[To Andy's Psychology Menu Page] [To Welcome Page]

How you can help... I hope you enjoyed these pages and found them useful in some way. Now, here's how you can help. Some of you have expressed a desire to help support these pages financially, I thought about it, and here it is... All I'm asking is a small donation (\$3.00 - \$5.00). Use of these pages is still free, no charge, you may continue to use them free of charge, as much as you want, as often as you want, anytime you want... But you can also use the links below to donate and help keep these pages here, and help the site expand. You may use your PayPal account, or most any credit card. Thank you for your continued support.

 Select this button to donate \$3.00 Select this button to donate \$5.00