# Manipulation and Measurement Operational definitions Two types

The Scientific Method Circle Form Hypothesis (from theory, model, or observations) What do your findings mean to the big picture?

What does it predict in various situations?

Design study to test hypothesis: H1vs H0. Eliminate confounding variables.

What conclusion(s) do your data compel?

Conduct study Data collection is done.

Manipulation and Measurement Clearly, in order to carry out an experiment, you will need to manipulate the variable in question to see if it affects the values you are measuring. The variable that you manipulate is called the independent variable. The variable that you measure is called the dependent variable. These are your data.

Operational definitions The second type: • Defining the Dependent Variable: Measured operational definitions, which describe what we do to measure the variables. They include exact descriptions of the specific behaviors or responses recorded. – For example, if we were trying to measure hunger (a hypothetical construct), we would need to show precisely how much work an individual would produce for a given amount of food.

Form Hypothesis • There are really two hypotheses: • H1: the Research (or “alternative”) Hypothesis -There is a difference between conditions (groups) • H0: The Null Hypothesis – there is no difference between conditions (groups). • We do not actually test the research hypothesis of an experiment directly. Instead, we test the null hypothesis. We try to reject the null hypothesis by demonstrating that a difference between conditions actually does exist, thus supporting the research hypothesis.

Operational definitions Two types: • Defining the Independent Variable: experimental operational definitions. They define exactly what was done to create the various treatment conditions. – Example: if the procedure required making the animal hungry, then we might define it by specifying that the subject was maintained at 90% of its freefeeding body weight by supplemental feedings one hour after sessions were completed. Or, we might define hunger as 24 hours of food deprivation.

The Scientific Method Circle Form Hypothesis (from theory, model, or observations) What do your findings mean to the big picture?

What does it predict in various situations?

Design study to test hypothesis: H1vs H0. Eliminate confounding variables.

What conclusion(s) do your data compel?

Conduct study Data collection is done.

1

What does it predict?

Derive Predictions

• Your hypothesis should make explicit predictions in a variety of situations, some (one) of which you can test by rejecting (falsifying) the Null Hypothesis (H0). • If it doesn’t make precise predictions, either the hypothesis is not very good or your situation is not applicable. • Your goal is to discover an explicit prediction in a situation that you can observe, eliminating extraneous influences.

• Rationale: If your research hypothesis is true and you have designed your experiment appropriately, then the conditions necessarily, absolutely, logically, must produce data that are different from one another. • If you do not find any differences, then your research hypothesis necessarily must be wrong in some way (or your experimental design sucks so badly that it is meaningless!). • This is where the principle of falsifiability fits in. If not falsifiable, then no conclusions can be drawn.

The Scientific Method Circle Form Hypothesis (from theory, model, or observations) What do your findings mean to the big picture?

What does it predict in various situations?

Design study to test hypothesis: H1vs H0. Eliminate confounding variables.

What conclusion(s) do your data compel?

Conduct study Data collection is done.

Design study to test hypothesis • The experiment is designed to measure differences between two (or more) conditions. • Extraneous variables (variables that affect all conditions randomly and equally) decrease our ability to demonstrate that our conditions are different. • Confounding variables (variables that are systematically correlated with a condition) prevent you from concluding that the conditions differ because of your hypothesis. These variables may have produced the differences you observed. Thus, there will always be an alternative explanation for your results. Usually, you can just toss your study into the trash.

The Scientific Method Circle Form Hypothesis (from theory, model, or observations) What do your findings mean to the big picture?

What does it predict in various situations?

Design study to test hypothesis: H1vs H0. Eliminate confounding variables.

What conclusion(s) do your data compel?

Conduct study Data collection is done.

What conclusions do your data compel? • If you have been measuring several things, then you are likely to discover that your results imply things that you had had not even considered. This is called serendipity. It is a wonderful side effect of doing research and usually leads to new ideas to test. • Of course, you hope that your results provide a clean test (no confounding variables and alternative interpretations sneaked in) of your null hypothesis (H0).

2

What do your data imply about your hypothesis? inferential statistics

What do your findings mean to the big picture?

• After completing the study, you have some amount of quantitative evidence for differences between the conditions you set up. • How much evidence do you need? How much difference is necessary for you to conclude that the conditions are “definitely” different? • This is where you use inferential statistics (not descriptive statistics). Inferential statistics calculate how likely the observed differences were due to random chance. (Descriptive stats just summarize data within one condition).

• You have now tested your hypothesis, but this hypothesis may be associated with several major theories. If you have rejected H0, then what does that mean for these theories? • Was your results important enough to alter these theories? • What experiment can you do now to clarify the issue? Can you design an experiment in which the different theories make different predictions?

Inductive vs. deductive logic

The Scientific Method Circle

• Deductive logic: inference from the general rule to the specific predictions. If you know a general principle, such as the rule for gravity, then you can make predictions about how it must work in a variety of different situations. • Inductive logic: inference from the specific facts to the general rule. Responsible for creating general principles. Sherlock Holmes combined little bits of evidence to produce a general principle that explained them all: The evil butler did it!

Inductive logic

Form Hypothesis (from theory, model, or observations)

Deductive logic

What do your findings mean to the big picture?

What does it predict in various situations?

Design study to test hypothesis: H1vs H0. Eliminate confounding variables.

What conclusion(s) do your data compel?

Conduct study Data collection is done.

3