II. Basics of Experimentation

Independent variable

the variable that is manipulated by the researcher to see if the different conditions of the independent variable affects the dependent variable

Dependent variable

the variable that is measured to see if it is caused to change by the different levels of the independent variable.

Extraneous variable

any variable that is not examined in the experiment, but can still affect the results of the experiment (e.g., hunger, lighting, temperature, noise, etc)

Experiment

A research method that examines a certain hypothesis under certain conditions to establish cause-and-effect relationship

Manipulation

The researcher “manipulates” the independent variable by assigning it to different levels.

Experimental group

Control group

treatment groups fall under two types

Experimental group

the group that gets a stimuli related to the independent variable

Control group

the group that is given nothing, a neutral stimulus, or a placebo

Control

Careful and accurate measurement of dependent variable

Random assignment

Manipulation

Four Important Elements of Experiment

Control

The researcher exercises constant control over the situation during the experiment is conducted.

Making sure that every participant undergoes the same conditions uniformly.

Making sure that extraneous variables are kept constant throughout the experiment.

how do researchers exercise control over the experiment

in our experiment, we can safely conclude that only the independent variable is affecting the dependent variable, and not any other variables.

Controlling extraneous variables ensures that?

• Lighting of laboratory/classroom

• Temperature in the laboratory/classroom

• Noise inside and outside the laboratory/classroom

• The materials that are given to participants

• The behaviors of experimenters/researchers

• The internal state of participants (hunger, wakefulness)

External Variables that are Controlled in Experiments

Careful and accurate measurement of dependent variable.

The researcher must be able to operationally define the dependent variable in order to measure it.

Operationalization

the process of giving definitions to your variables so that it is measured the way it is define

Random assignment

The process of assigning subjects to different treatment levels randomly, so that everyone has an equal chance of getting assigned to each treatment.

extraneous variables coming from individual differences (also called as random noise) such as personality, intelligence, ability, motivation, hunger, and wakefulness are controlled

Random assignment ensures that __________ such as personality, intelligence, ability, motivation, hunger, and wakefulness are controlled

• Between-subjects Design

• Within-subjects Design

• Factorial Design

Simple Types of Experimental Design

Between-Subjects Design

Also called as independent measures design.

A design where each treatment level has different sets of participants

Easier to conduct as the experiment is easier to set up and can be conducted with shorter sessions

• Minimizes order effects (mastery of measurement procedure because of repeated testing)

Advantages of Between-Subjects

Random noise

Disadvantages of Between-Subjects

Random noise

, also called as participant variables (e.g., age, intelligence, educational attainment) or extraneous variables coming from individual differences, can exert effect on the dependent variable.

random assignment

Solution to Random Noise

Within-Subjects Design

Also called as repeated measures design

repeated measures design

is technically a quasi-experiment because no random assignment is conducted for this design.

Only needs fewer participants, hence, more economical and financially practical for the researcher.

Increases the chance of finding a treatment effect since it eliminates participant variables (i.e., individual differences).

Advantages of Within-Subjects

• Increases the number of sessions spent with participants

• Fatigue from repeated-testing

Repeated testing may also induce order effects (mastery of the task)

Disadvantages of Within-Subjects

counterbalancing

Solution to the Disadvantages of Within-Subjects

factorial design

two independent variables are examined if they cause changes to a DV.

interaction effect

is the combined effects of two IVs on a DV.

1. The main effect of IV1 on DV

2. The main effect of IV2 on DV

3. The interaction effect of IV1 and IV2 on DV

Factorial Design Has three hypothesis:

Reliability:

Validity:

How can we evaluate if our experiment (including its results) is a good experiment?

Reliability

the extent to which results are consistent when they are tested twice or more than once

through replication of experimental procedures

How do we test reliability

replication crisis

was an observation that results of psychological experiments from previous decades were not replicable, and thus, not reliable. was an observation that results of psychological experiments from previous decades were not replicable, and thus, not reliable.

Validity

the extent to which the claims of the experiment is true and valid, and can be generalized to the population

Internal validity

it is concerned with correctly concluding that an independent variable is, in fact, responsible for causing changes in the dependent variable

It is ensured through controlling for extraneous variables and random assignment

How is internal validity ensured

External validity

: it is concerned with the generalizability of the findings of the experiment to the larger population.

Usually determined by replication to other samples or populations.

It is also determined by making sure that the results did not occur by chance.

How is external validity determined

Extraneous variables

pose threats to the validity of an experiment

• History

• Maturation

• Testing

• Instrumentation

• Statistical Regression

• Selection

• Mortality

Threats to Validity

History

This pertains to prior or present events that occurred outside of the experimentation which might affect the results of the data

Maturity

This pertains to physiological and psychological changes in the participants

longitudinal studies

Maturity is more common in?

Testing / Order Effects

If participants undergo repeated testing, this may result to familiarity or mastery of the test, which may affect the measurement of the dependent variable

test sophistication

Testing / Order Effects in the parlance of psychological testing, this is also called as?

Instrumentation

If the tool used to measure the dependent variable is in itself not reliable and valid, it can affect the results of the data.

Instrumentation

This is very common for self-made tests, because they did not undergo the rigorous process of establishing reliability and validity.

Regression Towards the Mean

This occurs if subjects are assigned to conditions that can artificially produce extreme scores.

Selection

becomes a threat to validity if the experimenters did not use random assignment (they were not selected using randomization procedures), or; the random assignment failed to balance out the extraneous variables.

Mortality

pertains to the loss of subjects over time. This may alter the results of each treatment condition, especially if there is a large loss for the condition`

• Experimenter-Expectancy Effect

• Demand Characteristics

Other Threats to Validity

Experimenter-Expectancy Effect

his happens when experimenters’ expectations and desires affect the outcome/results of their experiments in the way they want it to be.

Rosenthal / Pygmalion Effect

A researcher’s high expectations towards their participants, due to subtle cues of approval on the participants’ behavior, leads to an increase in performance of the participants

Golem Effect

A researcher’s low expectations leads to a decrease in performance of the participants

Demand Characteristics

This happens when participants try to search for subtle cues pertaining to information about the experiment, and act in accordance with that perception.

• Keep the extraneous variable constant for all participants

• Random assignment

General Approaches to Control

• Single-blind experiment

• Double-blind experiment

• Deception

Specific Approaches to Control

Double-blind experiment

happens when neither participants nor the researchers are informed about the nature of the treatment that the subjects receive

Double-blind experiment

This helps minimize experimenter-expectancy effects and demand characteristics

it is difficult to execute given that researchers will have decreased control over the experiment.

Disadvantage of Double-blind experiment

Deception

This occurs when subjects are told that the experiment has one purpose when in fact the purpose is really something else.

Unethical

Disadvantage of Deception