SPSS – 9/25/14
• Every line represents a person.
• Every column represents a variable (aka a question).
• String = both alphabets and numbers.
o Fine for names.
• Numeric = only numbers.
o Fine for age.
• Label = the question asked.
• Data Coding: assigning numbers to responses.
o For Gender: male is 1, female is 2
o Write explanation in values.
SPSS – 10/2/14
• Data Cleaning: making sure there aren’t any errors in data before you analyze.
o Out of range
▪ EX: 3.5 on a 1-7 scale.
▪ EX: 3 on age when there’s only 2 variables.
▪ Delete the whole individual (row) from the dataset.
o FINDING OUTLIERS (mean – 3sd, mean + 3sd)
▪ Analyze Descriptive Stats Descriptive
▪ Not for gender and age or demographics
▪ Only for scalable variables
▪ If we have an outlier, delete the individual from the dataset.
o Missing Value: blank
▪ Do not delete the individual from the dataset.
▪ Replace the missing value with the mean.
▪ REPLACING MISSING VALUES:
• Transform Replace Missing Value
• Get rid of the old variable column.
o NORMAL CURVE: done for scaled variables only.
▪ Analyze descriptive stats frequencies
▪ Charts Histograms show normal curves
SPSS – 10/16/14
1. What’s the final sample size?
2. What’s the final mean for performance?
3. What percentage of the sample is male? Female?
• FINDING MEANS:
o Analyze Descriptive Stats Descriptives • FINDING PERCENTS:
o Analyze Descriptive Stats Frequencies
4. Frequency distribution for class, MBA program
5. Sort performance in ascending order.
• SORTING THE DATA:
o Data Sort Cases Pick Ascending or Descending Okay
6. Recode gender into: 0-male; 1-female.
• RECODING THE DATA:
o Transform Recode into the Same Variable Old and New Values Add
Still Missing Remains Still Missing Okay
o Transform Recode into Different Variables Name the Variables Old and
New Variables Range and New Values
7. Split individuals into two groups
a. Who has a higher performance mean, males or females?
• SPLIT FILE:
o Data Split File Organize Output by Groups Groups Based On … Choose
Variable … Define Groups Okay
o To analyze the variable per group: Analyze Descriptives Okay
o Get the original file back: Data Split File Analyze All Cases Okay
8. Insert Variables
• INSERT VARIABLES:
o Click a variable to insert the specified one above it Edit Insert Variable
Name the Variable Data View Add Data
SPSS – 10/23/14
We’ll be tested on…
• Data Cleaning
• Testing Hypotheses
What Tests do we Use for Different Types of Hypotheses?
• Group Differences: comparing groups.
o EX: what are appropriate age groups for different types of restaurants?
o EX: what is the difference between business professionals from MBA program
and from BS program? • IV leads to DV: prove an experiment (cause-effect)
o REGRESSION! When IV (cause) is a scaled variable.
o ANOVA! When IV is a not scaled - categorical variable or a treatment
o EX: lack of brand awareness led to decline in Blackberry sales.
EX: Are males more satisfied than females with the steakhouse restaurant?
• Split the data by “familiar restaurant” and then by “gender.”
o Data Split File Organize Outputs by Groups; Groups Based On Gender
• Run a frequency on “satisfaction” and check the “mean” box in “statistics.” Output
shows satisfaction means and frequencies for males and females, for each restaurant
o Analyze Descriptive Statistics Descriptives Satisfaction
• Unsplit the data.
• We have to prove that the difference in means is statistically significant.
o Use T-Statistic.
• Independent Sample T-Test (compare two groups that are independent, not overlapping).
o Split the data, run a frequency, unsplit the data???
o Analyze Compare Means Independent Sample T-Test
▪ Grouping Variable (two groups you’re comparing): gender.
• Define groups (use specified values -> enter coding for each
▪ Test Variable (what you are comparing among the two groups):
o Read the first line of the independent sample t-test: equal variances assumed.
▪ The t-value is 3.197.
▪ We want to be at least 95% confident of the results.
▪ *If the absolute value of “t” is greater than or equal to 1.96, the groups
are statistically different in terms of the test variable.*
10/30/14 – T-Test
H1: the pepsi attitude mean is 2.1.
• One-sample t-test (is a predicted value correct or incorrect?)
o Analyze -> compare means -> one-sample t-test
o Test value – predicted value
o Test variable – the subject you’re predicting (attitude)
o One sample statistics chart shows the actual mean of the data. o The one sample test chart shows the t-statistic, which is 19.97 in this case (and
because the absolute value of t is greater than or equal to 1.96, there is indeed a
significant statistical difference.
o REJECT H1.
▪ I ran a one-sample t-test, and my t-value is 19.97. Because absolute value
of the t-value is greater than