PSY 348: Lecture 3
Farmingdale State College
Attitudes
Beliefs
Behaviors
Physiology
Planning
Brain Activity
Thoughts
Performance/Ability
Attenton
Memory
Operationalization
mood
Attitudes
Beliefs
Thoughts
Behaviors
Brain activity
Performance/Abilities
Self-Report or Behavioral Measures
Observation
Implicit Measures
Tests
Surveys
Questionnaires
Polls
Quizzes
Instruments
Most popular method of assessing attitudes
Can obtain large amounts of data
Allows for adaptive testing
Acquiescence
Social desirability
Demand characteristics
Anonymous respondents are less likely to make things up
Allow respondents to answer in private
Don’t make your experiment too obvious/revealing
Add questions that test for respondent awareness
Purposely make some questions opposite
Taking a flier
Signing a petition
Internet Behavior
Moving a chair
Donating money
Live aggression by adult
Videotaped aggression by adult
Cartoon aggression
Time consuming
Different reviewers/observers may score behaviors differently.
Coding scheme
Inter-rater Reliability
Others may choose to measure the phenomena differently from us
Operationalization can be culture-specific
What we measure is based on observable parts of the phenomena, but some parts may not be able to be observed
Observe children for one hour and…
Label them as Aggressive or Non-Aggressive
Rank them from most aggressive to least aggressive
Score them on a 10-point scale.
1 = No Aggression
10 = All of the Aggression
Count number of aggressive behaviors
Labels Nominal
Rank Order
Scale Interval
Count Ratio
IQ scores
Standardized test scores (e.g., SAT, GRE)
Temperature in Fahrenheit or Celsius
GPA
Reaction time (can be 0 ms)
Number of therapy sessions attended
Cortisol level in blood
Hours of sleep last night
Money donated to a cause
Interval may have a 0 on the scale, but it doesn’t mean the absence of something. A 0.0 GPA means an F rather than the absence of a grade.
Ratio has a 0 that. means 0 (the absence of something), like a ‘0’ as a response to the questions, “How many drinks did you have last night?”
Number of siblings (you can’t have 2.5 siblings)
Number of errors on a test
Types of therapy (CBT, psychoanalysis, none)
Diagnostic category (depression, anxiety, bipolar, etc.)
Eye color
Number of cups of coffee consumed today
Reaction time (seconds, milliseconds)
Height (inches, centimeters)
Weight (pounds, kilograms)
Self-esteem score on a 1–10 scale
GPA
Hours of sleep last night (can be 7.25 hours)
Brain activity (EEG voltage, fMRI signal intensity)
Undergraduates report their self-esteem on a scale of 1 to 10. Researchers assess the relationship of undergraduates’ self-esteem and GPA.
Predictor variable: Self-Esteem | Continuous (Interval)
Outcome variable: GPA | Continuous (Interval)
Participants are given a mystery drug or placebo and then asked to complete puzzle tasks. Researchers time how long participants take to complete the puzzles.
Independent variable: Drug vs. Placebo | Categorical (Nominal)
Dependent variable: Reaction Time | Continuous (Ratio)
Undergraduates are shown either pictures of death or pictures of landscapes. They then report their anxiety about aging on a scale of 1 (none) to 10 (lots).
Independent variable: Pictures of death versus landscapes | Categorical (Nominal)
Dependent variable: Anxiety about dying | Continuous (Interval)
Statistical tests can only be run on specific scales of measurement.
t-test & ANOVA: Nominal IV/predictor and Continuous DV/outcome
Correlation & Regression: All continuous
Generally, continuous variables (interval and ratio scales) lend themselves better to statistical analysis.
As a researcher, you should plan out your statistical analysis BEFORE conducting your study, so that you measure your variables in a way that matches the type of analyses you will run.
Data can be coerced into different formats
‘Correct’ or ‘Supporting’ results can come from bogus data
We’ve measured aggression/self-esteem/attention. Now, how do we summarize all these numbers?
\(n\) = the number of people in the sample
\(x\) = one datapoint in a sample
\(\bar{x}\) = Arithmetic mean (average) for a sample, x
\(\Sigma(x)\) = The sum of all values in a sample, x
\(\sum^{n}_{i = 1}= x_1 +x_2+x_3...\)
Esmeralda is taking Statistics for the Psychological Sciences. On her first exam, she scored 98. The average grade among the 26 students on that exam was 97
n:
x:
\(\bar{x}\):
Dave is interested in how many FSC students watch the Netflix Original, Dark. He plans to ask 200 students if they have seen it or not.
N:
n:
Scale of Measurement:
A research study surveys 500 college student across 5 campuses in the United States, asking them to rate their self-esteem on a scale of 1 to 10.
Sample:
Population:
Scale of Measurement:
What are the 4 scales of measurement
What are the only scales of measurement that matter for us?
Nominal
Ordinal
Interval
Ratio
n = ?
x = ?
n = the number of people in the sample
x = one participant’s score
What is a parameter?
What is a statistic?
What is a parameter?
What is a statistic?
Information that describes the sample.
Inferences about the population based on information from the sample.
Think about our data on time spent on Instagram.
How could we describe it?
Average
Min & Max
Central tendency: What most people said
Dispersion: How spread out the data are
Tell us about the mid-point (center) of a distribution.
Tell us about the mid-point (center) of a distribution.
Tell us most participants’ answer.
Mean
the average; obtained by summing all values and dividing by the number of values.
\(\sum^{n}_{i=1} = x_1 + x_2...\)
\(\frac{\sum(x)}{n}\)
Median
Mode
In a normal distribution, the mean, median, and mode are equal to one another.
Mean = Median = Mode
The mean is only helpful in a normal distribution, which you may have heard called a normal curve or bell curve.
We overwhelmingly use the mean, because in the social and behavioral sciences, we nearly always assume the distribution is normal.
A skewed distribution occurs when one side of the date gets cut off due to measurement limitation.
In a skewed distribution, the mean gets pulled out toward the tail, and the mode gets pulled to the cluster.
The median is the best measure of central tendency for skewed distributions because it tells us how most people answered.
The mode is good for bi-or-tri-modal distributions.
When might you have a bi-modal distribution?
Mean
Median
Mode
We use the mean most often in social science, because many of our statistical tests can only be used on normal distributions.
We assume normality of the distribution, and use the mean.
Sum all x values
Divide by the number of x values (n).
. . .
x = 4,2,5,6
. . .
\(\frac{\Sigma(x)}{n}\)
. . .
\(\Sigma(x) = 4 + 2 + 5 + 6 = 17\)
. . .
\(\frac{17}{4} = 4.25\)
Put all x values in order from smallest or largest or largest to smallest.
Find the middle number.
If there are an even number of x values, take the average of the 2 middle numbers.
x = 4,2,5,6,3
x = 2,3,4,5,6
Put all x values in order from smallest or largest or largest to smallest. \(x = [3,2,1,5,3,7,8,3,2,4,1]\)
Find the number or numbers that repeat the most. \(x = [1,1,2,2,3,3,3,4,5,7,8]\)
There can be more than 1 mode.
Cell
Columns
Rows
What does each row in this spreadsheet represent?
What does each column in this spreadsheet represent?
Adding & Summing
Subtracting
Multiplying
Dividing
Squaring
Adding & Sum: SUM(B2:B10)
Subtracting: C9 - C2
Multiplying: B12 * 4
Dividing: A2/B3
Squaring: (B2-C2)^2
Average: AVERAGE(B2:B20)
Open our Class Data in Microsoft Excel.
Calculate the Mean
Using (SUM(x))/n
Using Average(x)
Calculate the Median
Find the Mode(s)
Mean?
Median?
Mode?
Mean?
Median?
Mode?
. . .
\(\sum(x)\)
When are the mean, median, and mode the same (equal to one another)?
Mean → good for symmetric/normal.
Median → good for skewed/outlier-heavy.
Mode → good for categorical or multi-modal
We assume normality a lot because it lets us use the mean.
⬡⬡⬡