Does this use the data provided, please
1 SPSS LAB 2 A. Part 1: summary statistics...................................................................................................................................................................... 2 B. Part 2: Use SPSS to do hypothesis tests ................................................................................................................................................ 7 Now you will open an existing dataset, the one we used for our class. They investigated the credit card usage along with charity selection. The file contains just a few items (=questions) from the full study. Download and open the dataset in SPSS ► download and save the dataset “Lab Credit_card_survey_data short.sav” from CARMEN ► Start SPSS (in the menu “Spreadsheets and Statistics” on lab computers, or just type ‘SPSS’ after pressing the windows key) ► Select file – open (fig. right) A new window called ‘Open File’ opens ► Browse to the location where you saved the dataset ► Select the file “Credit_card_survey_data.sav” and click ‘Open’ 1. EXPLORING AN EXISTING DATASET The first time you open a new dataset you always ‘explore’ it by examining the variables in ‘Variable view’. Familiarize yourself with the new dataset ► Select ‘Variable View’ (see bottom left in the ‘SPSS Data Editor’ window) ► Familiarize yourself with the variables: (1) read the variable names in the column ‘Name’; (2) read the descriptions in the column ‘Label’; (3) investigate the labels of the variables in the column ‘Values’ (fig. below) The column ‘Measure’ indicates the scale level of the variables in the dataset. The scale level determines how you can summarize the variable. ► For each of the variables, verify that the correct scale level is set in the column ‘Measure’ ► Scale level is not only determined by the question (e.g. ‘What is your age’ = 24, but also by the possible answers – any value vs a list of possible ranges) ► Now check the value of ‘Measure’ for the variables on row. Correct them if needed (at least one of these is wrong, and possibly all!) using what is given in the labels and value coding. ANSWER QUESTIONS 1 ON YOUR ANSWER SHEET CONTENTS A. Part 1: summary statistics 2 2. Using Descriptive statistics in SPSS for discrete variables (Frequencies, Pie Charts, Crosstabs, Bar Charts) Now that you know how to perform basic data manipulation, let’s work through some basic descriptive statistics to summarize this data. Let’s start with discrete variables by using frequencies, pie charts, crosstabs and bar charts. 1) Frequencies ► Open the ‘Analyze’ menu, then select ‘Descriptive Statistics’, and choose ‘Frequencies.’ ► In the ‘Frequencies’ window select “Charity Selected to Support[pickcause]”and “Political Identity [politics]” variables. And press arrow button. ► Click ‘OK’ and the frequency table for “Charity Selected to Support[pickcause]”and “Political Identity [politics]” will appear in the SPSS Output window. A new window ‘Output1 – SPSS viewer’ opens. SPSS prints in this window all your output. It now printed a table with the title ‘Frequency Table’ showing the Frequency, Percent, Valid Percent, and Cumulative Percent of the variables. You can copy-paste this into the answer sheet or just extract the relevant pieces of information. On the answer sheet you only need to provide mean and standard deviation for ‘scale’ level variables and frequencies for nominal/ordinal scaled variables. NOTE: The second frequency table is provides the frequency, the percentages, and the valid percentages. The valid percentages remove the 35 respondents that did not answer “Political Identity [policics]” question. The percentage you want to use (Percent or Valid Percent) depends on whether or not the ‘Missing System” respondents are identifiable due to a non-answered question as well as the specific question you are trying to answer. For example, we cannot say much about people that simply chose not to answer a question for unknown reasons, but in the above case, if all 35 people are just those that have not provide answers to this question we may want to exclude them in the percentages we use to answer managerial questions! ANSWER QUESTIONS 2 ON YOUR ANSWER SHEET 2) Pie Charts: Next you will learn to create a simple chart for a nominal/ordinal variable. ► Open the ‘Graphs’ menu and select ‘Legacy Dialogs’, then select ‘Pie…’ and drag or double click the first chart type “Summaries for groups of cases”. ► Once the chart appears in the chart preview click and drag “Charity Selected to Support [pickcause]” onto ‘Slice-by’ box under the chart. ► Click ‘OK’ ► The chart preview should now show the pie chart for the variable. 3 ANSWER QUESTIONS 3 ON YOUR ANSWER SHEET 3) Crosstabs: Next you will learn to create a Crosstab for a nominal/ordinal variable. ► Open the ‘Analyze’ menu and select ‘Descriptive Statistics’, then select ‘Crosstabs’ and drag “Charity Selected to Support” onto “Row(s)” box, and “Political Identity” to ‘Column(s)’ box. ► Select “Cells”, and check “Observed” under “Counts”, click ‘Continue’. ► Click ‘OK’ Answer QUESTION 4 ON THE ANSWER SHEET For 4 b) Please Select “Cells”, and uncheck “Observed” under “Counts”, then check “Row” under ‘Percentages’. Click “Continuous”. For 4 c) Please Select “Cells”, uncheck “Row” and check “Column” under ‘Percentages’. Click “Continuous”. 4) Bar Charts: Next you will learn to create a Crosstab for a nominal/ordinal variable. ► Open the ‘Graphs’ menu and select ‘Legacy Dialogs’, then select ‘Bar…’ and drag or double click the last chart type “Stacked” and “Summaries for groups of cases”. ► Once the chart appears in the chart preview click and drag “Charity Selected to Support [pickcause]” onto ‘Category Axis’ box under the chart. Drag “Political Identity [politics]” onto ‘Define Stacks by’ box. 4 ► Click ‘OK’. ► Note again that this chart can be edited extensively. You can double click the graph on output window, then select “Show Data Labels” on Chart Editor. Choose “Percent” and move it to the “Displayed” box. Since we have different sample, so the label on the graph will not be the same as yours ► Click “Close” and click “X”. You will have the stacked bar plot with count and percentage labels. ANSWER QUESTION 5 ON THE ANSWER SHEET 5 3. Using Descriptive statistics in SPSS for continuous variables (Descriptive, Histogram & summary statistics, box plots) Now let’s investigate interval/ratio variables using the ‘Analyze’ menu. ► Open the ‘Analyze’ menu, then select ‘Descriptive Statistics’, then choose ‘Descriptives.’ ► In the ‘Descriptives’ window select all 7 “Attitude…” variables. ► To change the descriptive statistics that are found, click ‘Options’ ► In the ‘Descriptives: Options’ window select all the ‘Dispersion’ statistics and click ‘Continue’ ► Click ‘OK’ in the ‘Descriptives’ window to have the descriptive statistics for the 7 “Attitude…” variables appear in the SPSS output window Answer Question #6 on the Worksheet The next thing we will do with interval/ratio variables in this Lab is to create a histogram to understand the distribution of a variable. ► Open the ‘Graphs’ menu and select ‘Legacy Dialogs’, then select ‘Histogram’ ► Once the chart appears in the chart preview click and drag “Monthly Amount Charged to…” onto the Variable of the chart. ► Click ‘OK’ 6 The histogram for the monthly amount charged to credit card can now be viewed in the SPSS Output window. ► Note that this histogram can be edited for color, shape, labels, and much more by right-clicking the chart, selecting ‘Edit Content’ and choosing ‘In Separate Window’. This will bring up a chart editor in which you can make changes to the histogram as you see fit. Note that this is very valuable as you prepare your data for your project, as construction of clearly labeled and ‘nice-looking’ charts is very important in highlighting your results for decision-makers! Answer Question #7 on the Worksheet The last thing we will do with interval/ratio variables in this Lab is to create a boxplot to understand the distribution of a variable. ► Open the ‘Graphs’ menu and select ‘Legacy Dialogs’, then select ‘Boxplot’; use “Simple” and “Summaries of separate variables” ► Once the chart appears in the chart preview click and drag all 7 “Attitude…” variables onto the boxes represent. ► Click ‘OK’ Answer Question #8 on the Worksheet NOW LET’S TURN TO BIVARIATE STATISTICAL TECHNIQUES THAT YOU CAN USE TO INVESTIGATE RELATIONS BETWEEN TWO VARIABLES (TWO QUESTIONS IN A QUESTIONNAIRE). THE STATISTICAL TECHNIQUE THAT YOU USE DEPENDS ON THE SCALE LEVEL OF THE TWO VARIABLES (QUESTIONS). 1. Investigating the difference in the means of two groups You should have noticed that you have information on respondents’ opinions on several charity organizations. These issues are measured in the questionnaire using Likert scales (Very Bad – Very Good). Recall that Likert scales are assumed to be interval scales. Hence, you may compute means and standard deviations to summarize and analyze such data. The political identity variable ‘politics’ is a nominal variable (there are 4 categories in this variable). We can use this data to investigate the following research question: PART 2. SPSS LAB II: Hypothes test 7 Research question B: do Republicans have different attitudes than Democrats toward WHO? ► Determine the underlying variables that you are investigating here, and their scale levels Compute the mean and standard deviations for Republicans and Democrats for the Likert statement ‘who’ to investigate research question B. ► Select from the menu ‘Analyze > Compare Means > Means’ ► A new window opens called ‘Means’ (figure on right) ► Select ‘Attitude WHO [who]’ into the ‘Dependent List’ ► Select ‘Political Identity [politics]’ into the `Independent list’ ► Press the button ‘Options’ A new window opens called ‘Means: Options’ ► Make sure that the mean and variance are computed. These should appear under ‘Cell Statistics’ (see fig. on right). Note that there are several other statistics that you can have SPSS calculate! ► Select ‘Continue’ The window ‘Means: Options’ closes and you return to the window ‘Means’ ► Select ‘OK’ The results are printed in the output window. You will find the means and standard deviations for this Likert item for the different parties in the table ‘Report’. Note: This is still descriptive statistics only; we have not conducted tests yet! PLEASE ANSWER QUESTION 9 ON YOUR ANSWER SHEET We now have the means for attitudes toward WHO across four groups (Republican, Democrat, Independent and Other). This gives us an idea of what is happening in the sample, but management decisions are made for a population. To draw conclusions about the population, we can run hypothesis tests investigating differences across these two populations (Republican, Democrat). Investigate research question B (the relation between ‘who’ and ‘politics’) ► Formulate the null and alternative hypotheses for research questions B ► Choose the statistical test that applies to these hypotheses (what are your comparing across these two populations) PLEASE ANSWER QUESTION 10a ON YOUR ANSWER SHEET ► Go to ‘Analyze > Compare Means > Independent Sample T-test.’ ► Select the two variables as indicated in the figure below. ► Choose ‘Define Groups’ and set the two internet usage groups using the corresponding values (1 and 2) as indicated below. ► Click OK and evaluate the output. What is the P-value for the independent t-test? (Let’s focus on the two