Summary of SPSS Menu Bar:

Summary of SPSS Menu Bar/SPSS analyses:

New à from here you can open a new data window or syntax window.
Like every other program you’ve ever used, you can open, save, and print files. Nothing new here!

Per usual, you can undo and redo changes, copy and paste, as well as insert variables and cases of particular variables.
Option à From this menu, you can alter the appearance of the data view window, but there is really nothing you should ever need to change in here.

Useless. Just use the “Data View”/“Variable View” tab on the bottom left-hand side of the page.

Define Variable Properties à This dialog window is really another window to do the same thing as the variable view. Variable view is what you should use to work with to add labels, change the measurement level (eg: nominal v. scale), and column widths.
The rest of the menu might be used in advanced stats courses (with transposing or weighing cases), but is overkill for an intro methods course.

Compute variable à If you ever need to do any sort of transformation (like a log transformation or translate z-scores into t-scores, this is a handy window. It works much like Calculator on a PC. You simply type in the expression you want (For example, if you have a z-score you want transformed into a t-score, I might input “z_boredom*10+50” in the “Numeric Expression” window and set the “Target Variable” a new variable called t_boredom.). We will work more with this later in the course.
Recode into Same/Different Variable à If for some reason you want to change the way you coded a variable, congratulations for doing it wrong the first time. Now you can fix it.

Case Summaries à This will allow you to display a table with the raw data grouped by subject or any other categorical grouping you wish.
Report Summaries (in Rows/Columns) à This window will display the mean, median, standard deviation, skewness, maximum, minimum and many more statistics in the output window depending on how you want it presented. You choose which statistics you want by choosing “Summary…” under the “Report” option.

Frequencies à Find out how many people answered a particular question a certain way. (How many males & females are in the sample? How many people answered 7 out of 7 for the “boredom in class” question?)
Descriptives à get the N, Minimum, Maximum, Mean and Std. Deviation with basic descriptives (you can add more in “Options”). Also, by checking the box “Save standardized values as variables,” you are creating z-scores (standard scores) for those variables.
P-P Plots à A probability-probability plot can be used to see if your data follows some specified distribution. It should be approximately linear if the specified distribution is the correct model. If not, then a linear transformation might be in order.
Q-Q Plots à A quantile-quantile plot is another method for seeing if your data follows some specified distribution. Again, if it’s not linear, you’ll need to transform or use a nonparametric test.

Means à Use this option if you just want to compare means without any statistical mumbo jumbo. But since you love statistical mumbo jumbo and need to know the statistical mumbo jumbo anyway to do well in the course, you will instinctively use a real statistical test like a t-test or ANOVA.
One-Sample T-Test à Use this test when you want to see if the mean of your normally distributed sample is significantly different from a known mean (usually a population parameter). You enter this known mean in the “Test Value” field. For example, if you know the average salary of a grade school teacher is $48,788, use this test to see if the mean salary of a Northfield teacher is significantly different from the national average.
Independent Samples T-Test à Use this test when you want to compare the means for one or more variables between two independent groups. You will have to define what your two groups are by clicking on “Define groups.”
Paired Samples T-Test à Use this test when you want to compare the means for one or more variables between two groups that are not independent. For example, if you wanted to compare the mean weight loss between males and females from a sample of couples you could not assume independence of observations.
One-Way ANOVA à Use this test when you want to compare the means for a certain variable between three or more groups assuming a normal distribution, approximately equal variance, and independence of the observations in your sample.

Univariate à This model is “general” in the sense that one may implement both regression and ANOVA models. One may also have fixed factors, random factors, and covariates as predictors.
In the advanced package of SPSS, you can also run multivariate generalized linear models as well as repeated-measures models. If you need to do these kinds of analyses, learn how to use STATA (free on campus, but more difficult to use) or be prepared to sell a kidney to buy the $800 yearly license for the SPSS advanced add-on module (on top of the $600 dollars you’ve already spent on SPSS Base).

Bivariate à When one usually thinks of correlation, this is it. You can choose between Pearson, Kendall’s tau-b, and Spearman correlation coefficients. SPSS will give back a matrix in the output window with either one-tailed or two-tailed tests of significance.
Partial à Use this correlation when correlating two variables while controlling for a third or more variables. You can use partial correlations to explore whether your variables are overlapping or discredit the existence of a particular intervening variable.
Distances à Do you want to find out the Euclidean distance, squared Euclidean distance, Chebychev, block or Minkowski distances? Though not. Read on.

Linear à Perform simple (one predictor variable) and multiple linear (multiple predictor variables) regressions using this test.
Curve Estimation à This isn’t a “final” test in and of itself. It explores the relationship between one dependent and one independent variable in terms of linear, logarithmic, inverse, quadratic, cubic, power, compound, S-curve, logistic, growth, or exponential based models. You might use this to determine the correct test with which to analyze the data
Ordinal (Logistic) Regression à Use this type of regression when your dependent variable is a binary outcome (or at least non-continuous). The regression will then measure the probability of a particular outcome.

Factor (Analysis) à This analysis is used mainly to develop questionnaires to ensure that the researcher is measuring the construct they are attempting to explain or predict. Save this tidbit of information for grad school in quantitative psychology.