Experimental Design: The Basic Building Blocks

Chapter Nine: Designing, Conducting, Analyzing, and Interpreting Experiments with Two Groups

Experimental Design: The Basic Building Blocks

Experimental Design: The Basic Building Blocks
The general plan for selecting participants, assigning participants to experimental conditions, controlling extraneous variables, and gathering data.

The Two-Group Design Principle of parsimony

The Two-Group Design Principle of parsimony (Occam’s razor)
The belief that explanations of phenomena and events should remain simple until the simple explanations are no longer valid.

The Two-Group Design Independent variable

The Two-Group Design Independent variable (IV)
A stimulus or aspect of the environment that the experimenter directly manipulates to determine its influences on behavior.

The Two-Group Design Independent variable (IV) Dependent variable (DV)
A stimulus or aspect of the environment that the experimenter directly manipulated to determine its influences on behavior. Chapters 9 and 10 deal with research designs that have one IV. Chapter 11 deals with research designs that have more than one IV. Dependent variable (DV)

The Two-Group Design Independent variable (IV) Dependent variable (DV)
A stimulus or aspect of the environment that the experimenter directly manipulated to determine its influences on behavior. Dependent variable (DV) A response or behavior that is measured. It is desired that changes in the DV be directly related to manipulation of the IV.

The Two-Group Design How many groups?

The Two-Group Design How many groups?
Although an experiment can have only one IV, it must have at least two groups.

Although an experiment can have only one IV, it must have at least two groups. The simplest way to find out whether our IV caused a change in behavior is to compare some research participants who have received our IV to some others who have not received the IV.

Although an experiment can have only one IV, it must have at least two groups. The simplest way to find out whether our IV caused a change in behavior is to compare some research participants who have received our IV to some others who have not received the IV. If those two groups differ, and we are assured that we controlled potential extraneous variables, then we conclude that the IV caused the participants to differ.

The Two-Group Design Extraneous variables

The Two-Group Design Extraneous variables
Undesired variables that may operate to influence the DV and, thus, invalidate an experiment.

The Two-Group Design Levels

The Two-Group Design Levels
The most common manner of creating two groups with one IV is to present some amount or type of IV to one group and to withhold that IV from the second group.

The most common manner of creating two groups with one IV is to present some amount or type of IV to one group and to withhold that IV from the second group. Thus, the presence of the IV is contrasted with the absence of the IV.

The most common manner of creating two groups with one IV is to present some amount or type of IV to one group and to withhold that IV from the second group. Thus, the presence of the IV is contrasted with the absence of the IV. These differing levels of the IV are referred to as the levels (also known as treatment conditions) of the IV.

The Two-Group Design Experimental group

The Two-Group Design Experimental group
In a two-group design, the group of participants that receives the IV.

The Two-Group Design Experimental group Control group
In a two-group design, the group of participants that receives the IV. Control group

The Two-Group Design Experimental group Control group
In a two-group design, the group of participants that receives the IV. Control group In a two-group design, the group of participants that does not receive the IV.

The Two-Group Design Assigning Participants to Groups

The Two-Group Design Assigning Participants to Groups
Random Assignment

Random Assignment A method of assigning research participants to groups so that each participants to groups so that each participant has an equal chance of being in any group.

Random Assignment A method of assigning research participants to groups so that each participants to groups so that each participant has an equal chance of being in any group. Random assignment is not the same as random selection.

Random Assignment A method of assigning research participants to groups so that each participants to groups so that each participant has an equal chance of being in any group. Random assignment is not the same as random selection. When we randomly assign participants to groups, we have created what are known as independent groups.

Random Assignment A method of assigning research participants to groups so that each participants to groups so that each participant has an equal chance of being in any group. Random assignment is not the same as random selection. When we randomly assign participants to groups, we have created what are known as independent groups. When we wish to compare the performance of participants in these two groups, we are making what is known as a between-subjects comparison.

The Two-Group Design Independent groups

The Two-Group Design Independent groups
The participants in one group have absolutely no ties or links to the participants in the other group.

The Two-Group Design Independent groups Between-subjects comparison
The participants in one group have absolutely no ties or links to the participants in the other group. Between-subjects comparison

The Two-Group Design Independent groups Between-subjects comparison
The participants in one group have absolutely no ties or links to the participants in the other group. Between-subjects comparison Refers to a contrast between groups of participants who were randomly assigned to groups.

The Two-Group Design Confounded experiment

The Two-Group Design Confounded experiment
An experiment in which an extraneous variable varies systematically with the IV.

An experiment in which an extraneous variable varies systematically with the IV. Confounding makes drawing a cause-and-effect relation impossible.

An experiment in which an extraneous variable varies systematically with the IV. Confounding makes drawing a cause-and-effect relation impossible. Confounding may occur if participants are not equal before the start of the experiment.

The Two-Group Design Nonrandom Assignment to Groups.

The Two-Group Design Nonrandom Assignment to Groups.
Random assignment tends to create equal groups in the long run. As groups get larger, we can place more confidence in random assignment achieving what we want it to.

The Two-Group Design Nonrandom Assignment to Groups.
Random assignment tends to create equal groups in the long run. As groups get larger, we can place more confidence in random assignment achieving what we want it to. If we are faced with a situation in which we have few potential research participants and we are worried that random assignment may not create equal groups. What can we do?

The Two-Group Design Correlated assignment

The Two-Group Design Correlated assignment
A method of assigning research participants to groups so that there is a relationship between small numbers of participants.

The Two-Group Design Correlated assignment
A method of assigning research participants to groups so that there is a relationship between small numbers of participants. These small groups are than randomly assigned to treatment conditions (also known as paired or matched assignment).

The Two-Group Design Correlated assignment Matched pairs

The Two-Group Design Correlated assignment Matched pairs
Research participants in a two-group design who are measured and equated on some variable before the experiment.

Research participants in a two-group design who are measured and equated on some variable before the experiment. Typically we measure a variable that could result in confounding if not controlled.

Research participants in a two-group design who are measured and equated on some variable before the experiment. Typically we measure a variable that could result in confounding if not controlled. After we have measured this variable, we create pairs of participants that are equal on this variable.

Research participants in a two-group design who are measured and equated on some variable before the experiment. Typically we measure a variable that could result in confounding if not controlled. After we have measured this variable, we create pairs of participants that are equal on this variable. After we have created our matched pairs, we then randomly assign participants from these pairs to the different treatment conditions.

Repeated measures

Repeated measures The same participants are tested in both treatment conditions of our experiment.

Repeated measures The same participants are tested in both treatment conditions of our experiment. The matched pairs are perfectly equal because they consist of the same people or animals tested across the entire experiment.

Repeated measures The same participants are tested in both treatment conditions of our experiment. The matched pairs are perfectly equal because they consist of the same people or animals tested across the entire experiment. No extraneous variables should be able to confound this situation because any difference between the participants’ performance in the two treatment conditions is due to the IV.

Repeated measures The same participants are tested in both treatment conditions of our experiment. The matched pairs are perfectly equal because they consist of the same people or animals tested across the entire experiment. No extraneous variables should be able to confound this situation because any difference between the participants’ performance in the two treatment conditions is due to the IV. In this type of experiment, participants serve as their own controls.

Repeated measures Natural pairs

Repeated measures Natural pairs Pairs of participants are created from naturally occurring pairs (e.g. biologically or socially related).

Repeated measures Natural pairs Pairs of participants are created from naturally occurring pairs (e.g. biologically or socially related). For example, psychologists who study intelligence often use twins as their research participants.

The Two-Group Design Within-subjects comparison

The Two-Group Design Within-subjects comparison
Refers to a contrast between groups of participants who were assigned to groups through matched pairs, natural pairs, or repeated measures.

Refers to a contrast between groups of participants who were assigned to groups through matched pairs, natural pairs, or repeated measures. We are essentially comparing scores within the same participants (subjects).

Refers to a contrast between groups of participants who were assigned to groups through matched pairs, natural pairs, or repeated measures. We are essentially comparing scores within the same participants (subjects). Although this direct comparison is literally true only for repeated-measures designs, participants in matched or natural pairs are the same with regard to the matching variable.

Comparing Two-Group Designs
Choosing a two-group design

Choosing a two-group design Random assignment should should equate your groups adequately (assuming that you have large groups).

Choosing a two-group design Random assignment should should equate your groups adequately (assuming that you have large groups). If you are using 20 or more participants per group, you can feel fairly safe that randomization will create equal groups.

Choosing a two-group design Random assignment should should equate your groups adequately (assuming that you have large groups). If you are using 20 or more participants per group, you can feel fairly safe that randomization will create equal groups. If you are using 5 or fewer participants in a group, randomization may not work.

Choosing a two-group design Advantages of correlated groups designs

Choosing a two-group design Advantages of correlated groups designs Control issues

Choosing a two-group design Advantages of correlated groups designs Control issues The three methods for creating correlated-groups designs give us greater certainty of group equality.

Choosing a two-group design Advantages of correlated groups designs Control issues Statistical issues

Choosing a two-group design Advantages of correlated groups designs Control issues Statistical issues Correlated-groups designs can help reduce error variation.

Error variability

Error variability Variability in DV scores that is due to factors other than the IV – individual differences, measurement error, and extraneous variation (also known as within-groups variability).

Choosing a two-group design Advantages of correlated groups designs Advantages of independent-groups designs

Choosing a two-group design Advantages of correlated groups designs Advantages of independent-groups designs Simplicity

Choosing a two-group design Advantages of correlated groups designs Advantages of independent-groups designs Simplicity Use of correlated-groups designs is impossible in some situations.

True experiment

True experiment An experiment in which the experimenter directly manipulates the IV.

True experiment An experiment in which the experimenter directly manipulates the IV. Ex post facto research

True experiment An experiment in which the experimenter directly manipulates the IV. Ex post facto research A research approach in which the experimenter cannot directly manipulate the IV but can only classify, categorize, or measure the IV because it is predetermined in the participants (e.g. IV = sex).

Statistical Analysis: What Do Your Data Show?
The Relation Between Experimental Design and Statistics

The Relation Between Experimental Design and Statistics Selecting the appropriate experimental design determines the particular statistical test you will use to analyze your data.

The Relation Between Experimental Design and Statistics Selecting the appropriate experimental design determines the particular statistical test you will use to analyze your data. You should determine your experimental design before you begin collecting data to ensure there will be an appropriate statistical test you can use to analyze your data.

Analyzing two-group designs

Analyzing two-group designs For a two-independent groups design, you should use a t test for independent samples to analyzed your data.

Analyzing two-group designs For a two-independent groups design, use a t test for independent samples to analyzed your data. For a two-correlated-groups design, analyze your data with a t test for correlated samples (a.k.a dependent t test, a within-groups t test, or a paired t test).

Interpretation: Making Sense of Your Statistics

Interpretation: Making Sense of Your Statistics Statistics are a tool to help you understand the data from your experiment.

Interpretation: Making Sense of Your Statistics Statistics are a tool to help you understand the data from your experiment. Statistics are useless if you don’t know how to interpret them.

Interpreting Computer Statistical Output

Interpreting Computer Statistical Output The t test for independent samples

Interpreting Computer Statistical Output The t test for independent samples Homogeneity of variance

Interpreting Computer Statistical Output The t test for independent samples Homogeneity of variance The assumption that the variances are equal for two (or more) groups you plan to compare statistically.

Interpreting Computer Statistical Output The t test for independent samples Homogeneity of variance The assumption that the variances are equal for two (or more) groups you plan to compare statistically. Heterogeneity of variance

Interpreting Computer Statistical Output The t test for independent samples Homogeneity of variance The assumption that the variances are equal for two (or more) groups you plan to compare statistically. Heterogeneity of variance Occurs when we do not have homogeneity of variance.

Interpreting Computer Statistical Output The t test for independent samples Homogeneity of variance The assumption that the variances are equal for two (or more) groups you plan to compare statistically. Heterogeneity of variance Occurs when we do not have homogeneity of variance. This means that our two (or more) groups’ variances are not equivalent.

Interpreting Computer Statistical Output The t test for independent samples Generally speaking, t tests are robust with regard to the assumption of homogeneity (Kirk, 1968).

Interpreting Computer Statistical Output The t test for independent samples Generally speaking, t tests are robust with regard to the assumption of homogeneity (Kirk, 1968). A robust test is one that can tolerate violations of its assumptions and still provide accurate answers.

Interpreting Computer Statistical Output The t test for independent samples Translating statistics into words

Interpreting Computer Statistical Output The t test for independent samples Translating statistics into words If two equal groups began the experiment and they are now unequal, to what can we attribute the difference?

Interpreting Computer Statistical Output The t test for independent samples Translating statistics into words If two equal groups began the experiment and they are now unequal, to what can we attribute the difference? If our controls have been adequate, our only choice is to assume that the difference between the groups is due to the IV.

Interpreting Computer Statistical Output The t test for independent samples Translating statistics into words For example, if you were writing an interpretation of the results from the sample experiment in your text, you might write something like the following:

Interpreting Computer Statistical Output The t test for independent samples Translating statistics into words For example, if you were writing an interpretation of the results from the sample experiment in your text, you might write something like the following: Salesclerks who waited on well-dressed customers (M = 43.38, SD = 10.11) took significantly less time t(14) = 2.61, p = .021, to respond to customers than salespeople who waited on customers dressed in sloppy clothing (M = 63.25, SD = 12.54). The effect size, estimated with Cohen’s d, was .92.

Interpreting Computer Statistical Output The t test for correlated samples

Translating statistics into words Here is the example from your text:

Translating statistics into words Here is the example from your text: Salespeople who waited on well-dressed customers (M = 48.38, SD = 10.11) took significantly less time, t(7) = 5.47, p = .001, to respond to the customers than when they waited on customers dressed in sloppy clothes (M = 63.25, SD = 12.54). The effect size, estimated with Cohen’s d, was 1.93.

The Continuing Research Problem
It is rare for a psychologist to conduct a single research project and stop at that point because that one research project had answered all the questions about the particular topic.

Let’s review the steps involved in choosing a research design (using the hypothetical example from the text):

After reviewing relevant research literature, we chose our IV (style of dress) and our DV (salesclerk response time).

After reviewing relevant research literature, we chose our IV (style of dress) and our DV (salesclerk response time). Because we were conducting a preliminary investigation into the effects of clothing on salesclerks’ reactions, we decided to test only one IV (the style of dress).

After reviewing relevant research literature, we chose our IV (style of dress) and our DV (salesclerk response time). Because we were conducting a preliminary investigation into the effects of clothing on salesclerks’ reactions, we decided to test only one IV (the style of dress). Because we wanted to determine only whether clothing style can affect the performance of salespeople, we chose to use only two levels of the IV (dressy clothing vs. sloppy clothing).

After reviewing relevant research literature, we chose our IV (style of dress) and our DV (salesclerk response time). Because we were conducting a preliminary investigation into the effects of clothing on salesclerks’ reactions, we decided to test only one IV (the style of dress). Because we wanted to determine only whether clothing style can affect the performance of salespeople, we chose to use only two levels of the IV (dressy clothing vs. sloppy clothing). If we have a large number of participants available, then we can use random assignment, which yields independent groups. In this case, we would use the two-independent-groups design and analyze the data with a t test for independent groups.

Because we wanted to determine only whether clothing style can affect the performance of salespeople, we chose to use only two levels of the IV (dressy clothing vs. sloppy clothing). If we have a large number of participants available, then we can use random assignment, which yields independent groups. In this case, we would use the two-independent-groups design and analyze the data with a t test for independent groups. If we expect to have a small number of participants and need to exert the maximum degree of control, we choose to use a design with repeated measures or matched groups, thus resulting in correlated groups. Therefore, we would use a two-correlated-groups design for the experiment and analyze the data with a t test for correlated groups.

If we have a large number of participants available, then we can use random assignment, which yields independent groups. In this case, we would use the two-independent-groups design and analyze the data with a t test for independent groups. If we expect to have a small number of participants and need to exert the maximum degree of control, we choose to use a design with repeated measures or matched groups, thus resulting in correlated groups. Therefore, we would use a two-correlated-groups design for the experiment and analyze the data with a t test for correlated groups. We concluded that salespeople responded more quickly to customers in dressy clothes than customers dressed in sloppy clothes.

Experimental Design: The Basic Building Blocks

Similar presentations

Presentation on theme: "Experimental Design: The Basic Building Blocks"— Presentation transcript:

Similar presentations

About project

Feedback

Log in

Auth with social network:

Experimental Design: The Basic Building Blocks

Similar presentations

Presentation on theme: "Experimental Design: The Basic Building Blocks"— Presentation transcript:

Similar presentations

About project

Feedback