Psychology 240 - Comp Lab 2: Descriptives, Z-Scores, Power

Psychology 240:  Computer Lab Handout #2

Descriptives, Z-scores, Power

Learning Objectives:  1) Run descriptive statistics

                        2) Transform data to Z-scores

                        3) Run descriptives on Z-scores

                        4) Run power analyses

We will be using the same data set as the one from Computer Lab 2.  This data set was gathered from 24 hypothetical adults.  The researchers were interested in seeing whether stressful life events (SLEs) predict workplace absences.  The participants in this study reported the gender of their managers, the number of stressful life events they had experienced in the last year, and the number of sick days that had taken in the last year.

Run descriptive statistics and transforming values to z-scores

1. Examine the frequencies of SLEs and sick days (analyze → descriptive statistics → frequencies).  (2 points)

1. What is the most common number of sick days?

0

1. What is the most common number of stressful life events?

0

1. Examine the means, SDs, and ranges for SLEs and sick days (analyze → descriptive statistics → descriptives).  Check “save standardized values” to create new, standardized versions of these variables.

Note:  Look at the Data View page.  What do you notice?  These two new columns are our standardized variables (z-scores) – (5 points)

1. What is a typical number of sick days? (i.e., what’s the mean?)

6.83

1. How far do people tend to be from that average (i.e., what’s the SD?)

4.90

1. What’s a typical number of SLEs?

5.54

1. How far do people tend to be from that average?

3.98

1. What are the ranges in the values for SLEs and sick days?

SLE Range is 15 and Sick is 19

1. Examine a histogram for sick days (graphs → legacy dialogs → histogram, then add sick days to the variable box).   Then examine a histogram for SLEs (graphs → legacy dialogs → histogram, then add SLEs to the variable box).  If you click on “display normal curve,” it will superimpose a normal curve over the data.

What do you notice about the histograms?  Do the histograms look normally distributed or skewed? (1 point)

In both histograms the variables are close to normal but not exactly normal.  

Both histograms look relatively normally distributed with SLEs being slightly more so.

1. Change scale of x-axis to make more informative (Click on Options→“bin element”).  Click on the Custom radio button under the x-axis heading.  Change the number of intervals to 20. Click Apply.  Click Close.

What do you notice about the histograms now?  Do the histograms look normally distributed or skewed? (1 point)

The bins got smaller because there are more bins presents.  There is more detail presents with the more bins.

The sick day’s histograms is slightly skewed but still pretty normally distributed.  The histogram for SLE’s show a normally distributed bell curve aside from those with 0 SLEs.

1. Examine a scatterplot of the relationship between sick days and SLEs (graphs → legacy dialogs → scatter/dot → simple scatter, then add sick days to the y-axis box and stressful life events to the x-axis box).  (1 point)

Does it look like there’s a relationship?  

Yes it seems that as SLEs increase sick days increase.  

Add a regression line to your scatterplot by double clicking on any dot and then clicking on the “add fit line at total icon,” located fifth from the left on the bottom row of icons.

        Is there a linear relationship?

There is a moderate positive correlation that is enforced by adding the fit line.

1. Examine correlations among all 3 variables (analyze → correlate → bivariate).

What do you notice?  Are the variables significantly correlated with each other?

(3 points)

        From the diagonal line if you look above or below the results are mirrored.  The variables are significantly positively correlated with each other.  Between stressful life events and manager the correlation is .50, between sick days and manager the correlation is .52, and between sick days and SLEs the correlation is .59.