Independent Samples T-test
Use it for comparing the means of a DV across two unrelated groups (determined by a single IV)
Example: Is mean anxiety different between men and women?
Dependent (Paired) Samples T-test
Use it for (1) comparing the means of a DV in the same group, but before and after and intervention or (2) comparing the means of a variable in two "paired up" groups.
Example (1): Did symptoms change from before the treatment to after the treatment?
Example (2): Do wives and husbands in two-parent working households spend different amounts of time reading to their children (wives and husbands paired up in the analysis)
One-Way ANOVA
Like the independent samples t-test (see above), but for 3+ groups
Example: Are there differences in level of concern for the environment among Democrats, Republicans, and Independents? Note: use post-hoc analysis to determine exactly which groups differ, if any.
Two-Way ANOVA
Compare the mean of a single DV, across the groups determined by two different IVs
Example: Do household incomes differ between households that rent or own their house (IV1) and / or between households in suburban and urban areas (IV2). Note: the two-way ANOVA will test for an effect of each IV individually, as well as the interaction of the two IVs.
Chi-Square Test of Independence
Tests if the grouping of people (or other entities) across levels of one IV depends on the level of another IV
Example: Does choice of car color (red, green, blue, yellow, other) depend on state of residence (PA, NJ, NY)? Note: the variables are categorical, and what is being compared is whether people distribute differently in terms of car color choice, depending on which state they live in.
Correlation
Tests for a linear relationship between two variables, determines its strength and direction
Example: Do children's IQs correlate with their parent's total income?
Regression
Tests if the value of one continuous (or nearly) variable can be predicted from the value of another continuous (or nearly) variable. Regression also provides an equation for making the prediction.
Example: Can the amount of dust in the air in a city be predicted by the number of cars?
Independent Samples T-test
Analyze→ Compare means→ Independent Samples T Test
The IV is the “Grouping Variable”
The levels of the IV must be defines (click on “Define Groups”) as it was coded (For example, Male =1, Female = 2)
The DV goes in the “Test Variables” box
Dependent (Paired) Samples T-test
Analyze→ Compare means→ Paired Samples T Test
Put both variables in as a pair - click “OK”
One-Way ANOVA
Analyze→ Compare means→ One-way ANOVA
The IV is the “Factor”
The DV goes in the “Dependent List” box
If result is significant, a Post-Hoc test (e.g., Tukey) is run to find where differences lies
Two-Way ANOVA
Analyze→ General Linear Model→ Univariate
The IVs are “Fixed Factors”
Click on “Options” to get means of the DV at all levels of the two IVS [CHECK]
Click on "Post Hoc" to run Tukey's HSD (select any IVs with 3+ levels)
Chi-Square Test of Independence
Analyze → Descriptive → Crosstabs
One categorical variable goes in “Column” and other in “Row”
Click on “Statistics”, then check “Chi Square”
Correlation
Analyze→ Correlate→ Bivariate
Select all of the variables for which you would like to calculate correlations
Linear Regression
Analyze→ Regression→ Linear
The DV (the to-be-predicted variable) goes in “Dependent”
The IV (predictor variable) goes in “Independents”
Select Cases
Data→ Select cases→
Use this if you wish to only run analysis on a subset of your data. For example, you only wish to analyze male participants.
Highlight “If” then click Blue button “If”
You can select cases based on specific values or ranges of values of one or more variables
Split Files [This needs better explaining]
Data → Split files→
Use this if you want to run the same analysis separately for people in different groups. For example, you may want to compare men and women on their anxiety scores (independent samples t-test), but you want to do this same analysis three separate times: once for men and women living in cities, once for men and women living in the suburbs, and once for men and women living in the country.
Click “Organize output by groups” then use the arrows to put the variable you wish to split by (e.g., Gender) into box on the right, then click “OK”
Computing New Variable
Transform → Compute variable
Create new name for variable in box on left
Provide mathematical formula for computing that variable in box on right (e.g., (HappinessScore1 + HappinessScore2)/2), then click “OK”
How to report data in the correct format and to the correct number of decimals.
Writing Numbers:
Generally, you should spell out numbers below 10 in words (seven, three), and use numerals for anything 10 and higher (10, 42).
Decimals:
APA is somewhat vague on exactly how to report decimals, but they suggest the following general principle: "Round as much as possible while keeping prospective use and statistical precision in mind."
So, fewer decimal digits are easier to comprehend, meaning it is better to round to two decimal places when possible. You should report correlations, proportions (percentages), and inferential statistics such as t, F, and X2 to two decimals.
However, when the measure requires 3 or more decimals (e.g., human reaction time), report to 3 or more decimals.
P-Values:
When reporting p values greater than .001, always report exact p values (e.g., p=.22 instead of p>.05).
Report p-values .05 and greater to two decimals (e.g., p=.08).
Report p-values <.05 to three decimals (e.g., p = .031).
Report p-values less than .001 as p.< .001.
When SPSS provides a p-value of .000, report as p<.001.
Note: For all tests, report exact p value to 3 decimals. If SPSS indicates a p-value = .000, then report p <.001.
Independent t-test
Significant
Students who took a quiz while listening to classical music (M = 8.20) were found to score higher than students who took a quiz while listening to rock music (M = 6.00), t(8) = 2.75, p = .025.
Not significant
There was no significant difference between students who took a quiz while listening to classical music (M = 8.20) and students who took a quiz while listening to rock music (M = 6.00), t(8) = 1.60, p = 0.120.
Paired sample t-test
Participants were less anxious after taking medication (M = 3.20) compared to their pre-medication baseline (M = 8.00), t(4) = 8.23, p < .001.
One-Way ANOVA With Tukey Post-Hoc Analysis
There was a significant difference in happiness scores between those who ate fruits (M = 8.00), vegetables (M = 6.00), and donuts (M = 4.0), F(2,6) = 7.19, p =.025. Post-hoc analyses utilizing a Tukey HSD indicated that people who ate fruits were significantly happier than those who ate donuts (p < 0.05).