1/48
Looks like no tags are added yet.
Whats is a Variable?
Is something that varies
⢠Because it varies, it must have at least two levels (or at least two values) ex. Gender, GPA, Age, Selfāesteem, IQ, Lifeāsatisfaction
Conceptual definition (construct)
A researcherās definition of a variable at an abstract/theoretical level
⢠Example: personality, infant temperament, anxiety
Operational definition
ā the specific way in which a concept of interest is measured or manipulated as a variable in a study.
Construct example
Childās physical agression
Operationalization of a childās physical aggression
Teacher Questionnaire and Observe Childās Behavior at School
Scales of Measurement- Categorical
assign numbers to categories (nominal), arbitrary. ex. Gender and Hair color
Scales of Measure-Quantitative
meaningful numbers. ex.Score on a depression inventory, Number of times a child hits or kicks another child, and GPA
Measured variable
A variable that is observed and recorded
⢠Examples: height, IQ, gender, hair color, depression, stress
Manipulated variable
A variable where levels are controlled by assigning participants to various levels of that variable
⢠Example: Testing environment ā randomly assigning participants to take a test in either a quiet or chaotic environment
Variables
Some variables can only be measuredānot manipulated
⢠Examples: Age, quality of a particular school
Some variables can be either manipulated or measured
⢠Example: quality of the testing environment
Variables in an Experiment
An Experiment is a research method in which the researcher systematically
manipulates one variable and measures the other variable, while holding
other variables constant.
**Good experiments are the only way we can make causal claims between variables**
Independent variable
The variable that is systematically manipulated by the researcher to produce changes in the outcome (dependent variable).
⢠Random assignment must be used when assigning to groups
Dependent variable
The outcome or response that is being measured
Does the quality of the testing environment affect test performance?
What is the independent variable?
The testing environment
⢠How would we manipulate this? Two levels/conditions: quiet or chaotic
Does the quality of the testing environment affect test performance?
The dependent variable?
Test performance
⢠How would we measure this? The participantās score on the test. The scores
would be on a scale from the lowest to highest possible score.
Participants-descriptive statistics
The sample (participant demographics)
Materials
⢠Describe how each variable was operationalized and measured, including the scale
Procedure
⢠Describe the procedure (what were participants asked to do)
SPSS stands for
āStatistical Package for the Social Sciencesā
abbreviations for statistics (italicized)
N=sample size (total sample for the study)
n=all other participant numbers (e.g., number of participants in a group)
M=mean, SD=standard deviation
Greek letters are not italicized
Categorical data
⢠Report the n (number) and % for each category (e.g., for males, females, intersex)
⢠For presenting percentages, use āValid Percentā in SPSS output
Quantitative data
ā¢Report the M and SD
⢠Always include units! Example: Age (years)
⢠Always include SD with M; always include n with %
Hypothesis testing
is the process of determining an answer to our research question through probability.
Null hypothesis statistical testing (NHST)
Is a commonly accepted, standard way of testing psychological hypotheses.
⢠āNullā means 0 or nothing
⢠The null hypothesis therefore means that there is no relationship between the 2 variables in the population and that the controlling variable has no effect on the other.
NHST
What is the likelihood (or probability) that we find an association between variables in our sample if H0 is true in the population?
allāimportant pāvalue*
The purpose of statistical analyses is to determine the likelihood of getting a specific result IF H0 is true
⢠In this lab, we will focus on the pāvalue. The pāvalue tells us the likelihood that an association between variables occurred due to andom chance (i.e., IF H0 is true )
⢠In behavioral sciences, a commonly accepted standard cutāoff point for the pāvalue is .05.
If the pāvalue is less than .05,
We reject the null hypothesis.
⢠This means the likelihood of getting a specific result IF H0 is true is very small.
⢠Therefore, we reject the assumption that H0 is true. We reject H0.
⢠We say that the association between the variables was probably not due to random chance and is therefore statistically significant.
⢠Note: The pāvalue can never be 0. If you see p=0 in your SPSS output, report p < .001
Dr. Yās research hypothesis is that disengaged
parents are more likely to have anxious kids.
⢠The results: kids of disengaged parents score 0.5
points higher on average on an anxiety scale and pāvalue is .10.
⢠Reject or fail to reject the null?
⢠Fail to reject
⢠Were the results statistically significant?
⢠No
⢠Was the research hypothesis supported?
⢠No
Dr. Zās research hypothesis is that disengaged
parents are more likely to have anxious kids.
⢠The results: kids of disengaged parents score 0.5
points lower on average on an anxiety scale and pāvalue is <.001.
Reject or fail to reject the null?
⢠Reject
⢠Were the results statistically significant?
⢠Yes, the pāvalue was less than .05
⢠Was the research hypothesis supported?
⢠No, the results were in the opposite
Dr. Xās research hypothesis is that the
amount of time watching TV is positively associated with obesity.
⢠The results: r is +0.35 and pāvalue is .002.
Reject or fail to reject the null?
⢠Reject
⢠Were the results statistically significant?
⢠Yes
⢠Was the research hypothesis supported?
⢠Yes
Dr. Yās research hypothesis is that disengaged
parents are more likely to have anxious kids.
⢠The results: kids of disengaged parents score 0.5
points higher on average on an anxiety scale and pāvalue is .10.
Reject or fail to reject the null?
⢠Fail to reject
⢠Were the results statistically significant?
⢠No
⢠Was the research hypothesis supported?
⢠No
my hypothesis is that people who were born in the south are more likely to prefer sweet food,
compared to people who were born in the north.
What is the null hypothesis? What is the research hypothesis?
⢠People are equally likely to prefer sweet food whether from the north or south
⢠People who were born in the south are more likely to prefer sweet food than people born
I conducted a betweenāgroups study in which I asked people
where they were born and then had them pick out their favorite food from a list.
⢠What are two variables? ⢠Is this an experiment? Why or Why not?
⢠Place of birth (North or South)
⢠Preference for sweet food (measured by choosing favorite food)
⢠NOT an experiment
⢠You canāt randomly assign where a person was
I conducted an ANOVA (youāll learn this term soon)
and obtained a Pāvalue of .07.
⢠Do we reject or fail to reject the null hypothesis?
⢠Fail to reject.
⢠Was our research hypothesis supported?
The pāvalue is greater than .05, which means that any
relationship between place of birth and preference for
sweet food occurred by random chance and likely does
not actually exist in the population.
⢠No, there is no statistical significant difference in sweet
preference between people born in the
BetweenāGroups ANOVA*(Analysis of Variance) and Tātest
1 categorical variable
1 quantitative variable
Mean difference Is there a difference in the means across groups?
Chiāsquare (X2 )
2 categorical variables: Pattern of relation
Does membership in one
category predict membership of another category?
Correlation
2 quantitative variables Linear
If scores on one variable increase (or decrease), does the other variable increase
contingency table.
Pearsonās Correlation (r)
reported as r
The Pearson correlation coefficient (r) tells us the direction and strength of the linear relationship
⢠Direction: Correlations can be positive or negative.
⢠A positive correlation between two variables means that, as one variable increases, so does the other
⢠A negative correlation between two variables means that, as one variable
increases, the other decrease
Pearsonās Correlation (r)
Possible r values range from ā1.00 to 1.00
Values closer to 1.00 or ā1.00 are stronger. Values closer to 0 are weaker.

Dr. Xās hypothesis is that the number of hours
watching TV is positively correlated with a
personās weight.
⢠Results: r = .65, p= .02.
⢠What are the two variables?
⢠Is the correlation positive or negative?
⢠Weak or strong?
⢠Significant or not?
⢠Is Dr. Xās hypothesis supported?
Variables: Number of hours watching TV and a personās weight.
Correlation is positive
Strong or moderate
Is significant
Yes Dr. Xās hypothesis is supported
Dr. Zās hypothesis is that the number of
friends someone has is not correlated with
happiness.
⢠Results: r = ā.06, p = .79.
⢠What are the two variables?
⢠Is the correlation positive or negative?
⢠Weak or strong?
⢠Significant or not?
⢠Is Dr. Zās hypothesis supported?
Variables: Number of friends and happiness.
Negative correlation
Weak
Not significant
Yes Dr. Zās hypothesis is supported
Dr. Yās hypothesis is that parentsā rating of
disengagement is positively correlated with
their kidsā anxiety score.
⢠Results: r = .17, p = .10.
⢠What are the two variables?
⢠Is the correlation positive or negative?
⢠Weak or strong?
⢠Significant or not?
⢠Is Dr. Yās hypothesis supported?
Variables: Parentsā rating of disengagement and the kidās anxiety score
Positive correlation
weak
Not significant
No the hypthesis is not supported
ANOVA results
ā Bar Graph
⢠A bar graph is commonly used to present the results of ANOVAs/tātests
⢠In a bar graph, we present the mean for group 1 and the mean for group 2
Correlations results
Scatterplot
⢠A scatterplot is commonly used to present the results of correlations
⢠In a scatterplot graph, each point represents one participant.
X^2 Results
ā Contingency Table
⢠A contingency table is commonly used to present the results of X^2 analyses
⢠Particularly useful when there are multiple X^2 analyses to report
⢠In the contingency table, the pattern of results are presented.
Strengths, limitations, future directions
Construct validity
Quality of the measures and manipulations
Statistical validity:
Statistical conclusions are appropriate and
reasonable.
Internal validity
No alternative causal explanations for the outcome
External Validity
To whom, what, or where can we generalize?