Using

Microsoft Excel and SOFAstats: An Instruction Manual Jamela Towler and Nathan Kuik University of Washington

Table of Contents I. Introduction....3 II. Introducing the SOFAstats Interface..4 III. Importing Data into SOFAstats....5 A. Preparing Excel Spreadsheets.5 B. Importing the Spreadsheets into SOFAstats.6 IV. Data Analysis.9 A. Creating Tables and Crosstabs...9 B. Creating Graphs.........11 V. Exporting Graphs13 VI. Wilcoxon Signed Ranks Test.........14 VII. Additional Resources15

I. Introduction
Welcome to this brief tutorial on using SOFAstats. Stats can feel overwhelming and unrelated to the mission of an organization, but they are a necessary evil. Hopefully this instruction book can ameliorate some of your woes. We are not trying to make stats fun; we are trying to provide step-by-step instructions so tSB can compile their own stats. Before jumping into the computer instructions, we would like to provide a brief overview of SOFAstats. SOFAstats is an open source statistical analysis program. Open source means the developers allow the program to be downloaded and used free of charge, and they are continually fixing bugs and adding features. Here is a brief summarization of the positives and negatives of the program. Pros: Free to download and use Developers consistently Add features and fixes which are also free Works adequately with Excel if the spreadsheet is organized correctly Produces aesthetic graphs with little work Works on Macs and PCs Running statistical analyses and creating frequencies is more straightforward than other software Cons: Less polished than more expensive programs Complicated process for using graphs Difficult to modify data structure once imported Unique interface may be unfamiliar to new users Bottom-line: Since SOFAstats is free and relatively easy to use, it provides individuals and organizations a good opportunity to create powerful graphs and information, but requires some program knowledge and manipulation.

II. Introducing the SOFAstats Interface

SOFA statistics presents an easy to use, point and click interface that allows for multiple types of analysis and presentation of results. The first screen that appears once SOFAstats is opened will be referred to as the interface screen. The interface screen presents a variety of options for the user to select. Aside from multiple links a user can select for help - including Get help from community and Online help the options on the interface screen allow for users to create datasets, upload data, create tables and charts, run basic statistical tests and select from saved projects and preferences. The first button available to users, labeled Get Started, provides step-by-step examples and screen shots similar to the tutorial you are currently reading. The purpose of this tutorial is to inform users about the different options available to them on the interface screen, and our tutorial will provide users with easy to follow instructions on how to utilize each of the options presented.

III. Importing Data into SOFAstats

A. Preparing Excel Spreadsheets Before importing data into SOFAstats, it is necessary to fix some aspects of the Excel spreadsheet so it imports correctly. It is important that you delete any missing values within the Excel spreadsheet to begin a statistical analysis in SOFA. Here is how to do this: 1) In Excel, make sure youre in the Home menu at the top, CLICK the binoculars/Find & Select, and choose Replace: 2) A window will appear called Find and Replace:

3) First, Type -9 in the Find what field and leave the Replace with field blank:

4) CLICK Replace All 5) Next, repeat the same step (opening Find and Replace), but type #NULL! and leave the replace section blank:

6 6) Now, go to File, choose Save As, and save the file as Excel 97-2003 Workbook. This step is extremely important; SOFAstats cant import newer file types. Name the file something like "

7) Click Save

*Note: Make sure to use the spreadsheet found in the portfolio. Any type of analysis will not work in SOFA unless the responses of each participant are on the same row. To put it another way, the pre and posttest scores need to be on the same row. Each participant needs to be on their own row in Excel.

For more help and instructions, visit this link.

B. Importing the spreadsheet into SOFAstats

1) Open SOFAstats 2) CLICK Import Data on SOFA menu screen:

3) In Select file to import BROWSER, choose the Excel spreadsheet that you saved as an xls file. CLICK Import

If you dont like the long file name, you can make it whatever you want in the SOFA Table Name line. When the window pictured below pops up, click Has Header Row:

A series of windows like this will pop up. Click Numeric for all of them:

4) A message will appear that says Successfully imported data, CLICK ok 5) Now close the Select file to import screen by PRESSING close 6) To look at your dataset in SOFA the Enter/Edit Data tab a. A message will appear called Data in Project, CLICK open to look at the dataset. If it looks accurate, PRESS close

Now you are ready to do your data analysis in SOFA (Frequencies, cross tabs, charts, etc.)

For more help and instructions, click this link.

IV. Data Analysis

A. Creating Tables and Crosstabs in SOFA stats
A table in SOFA stats will show the frequency and percentage of responses to each variable. 1) CLICK Report Tables on the SOFA stats interface:

2) We then need to make sure the radio button labeled Frequencies is selected. 3) Add the variable you wish to examine in ROWS.

10 4) 5) 6) Lets start with Age (Q1_1). CLICK Add and select Age (Q1_1) as the variable you would like presented in the rows. Once you have selected Age, CLICK ok, and then an example of a table should appear in the output window. In order to show the correct results for Age, you need to CLICK the Show Results button on the right side of the window. Then a table presenting the frequencies of each age cohort will appear in the output window. 7) 8) 9) 10) 11) In order to create a crosstab, we first need to select the radio button labeled Crosstabs next to the Frequencies button. We now need to first select our variable to display on the ROWS once again. Remember, to do so, CLICK Add, select Age once again and CLICK ok. Now, lets select the variable to display in the COLUMNS. To do so, CLICK Add under the Columns: heading, and select the variable of interest. In this example, lets select Q3_1, or Gender and CLICK ok. Again, a sample of a crosstab should appear in the output window, but in order to see the correct results we need to CLICK the Show Results \ button. From here, a crosstab showing the frequency of each age cohort by both female and male, numerically presented as one and two, respectively.

11 12) Finally, we can Add a Title and Subtitle to our tables by typing them in the boxes available. We can also present percentages for each row by CLICKING the Config button and selecting Column %. We must remember that after each change we make we must again CLICK Show Results. After we have finalized the tables and crosstabs to our satisfaction, we can follow the same instructions we used for our Charts to export the tables, in Section V.

13)

For more help and instructions, click here.

B. Creating Graphs
1) Click on Charts

12 2) Make sure the Table drop down menu has the correct data set loaded:

3) In the values drop down menu select a variable for example Age (Q1_1). Leave Charts By box blank to only examine one variable:

4) Next, CLICK on the Chart type you would like to use to display the frequency for by selecting the radio button next to your preferred display type 5) Title your chart in the Title box 6) Select Report Data as frequency or percentage 6) Click Show Results

SOFA has many too many chart options to discuss in this document, and a small amount of clicking around can create some powerful graphs.

For more help and instructions, click here.

13

V. Exporting Graphs
a. CLICK on View at the bottom of output section of chart screen:

b. You will be directed to the SOFA report screen in a web browser c. The next step will depend on your web browser. The goal is to save the webpage to a PDF. This can only be done in Google Chrome. If the webpage opens in any other browser, copy the URL and paste it in Google Chrome. CLICK on the in the upper right-hand, and then Print.

d. In the print window, change the Destination to Save as PDF. e. After saving the PDF, open it with Adobe Reader, click Edit and select Take a Snapshot. Draw a rectangle around the graph you want, and this will copy the graph. f. Copy the graph in whatever document you need it. Note: This is a cumbersome way of getting the graph or chart to a place where it can be manipulated, but there is currently no way to simply export the graph. If the procedure doesnt work through Google Chrome, it is also possible to use Window 7s Snipping Tool. Moreover, it is possible to select parts of a print preview in Apples Safari web browser on a Mac, so downloading and using this is also an option.

Click here for the instructional video.

14

VI. Wilcoxon Signed Ranks Test

Bear with us for some brief technical jargon. The gold standard for program evaluation involves producing statistically significant results. This simply means finding difference between pretests and posttests responses that would have an extremely low probability of happening by chance. In other words, the change in response is so drastic or widespread we can estimate that the difference did not simply just happen. If the directions seem overly simplistic or condescending, this is not our intention. We want to ensure tSB understands the directions for conducting the tests. Now we will explain how to perform a Wilcoxon Signed Rank Test in SOFAstats. It is not necessary to understand what the test means or does, but it is important to understand enough to perform and interpret the test. This test should be run when comparing pre and posttest responses to the same question. It does not work to compare on variable (i.e. age, year in school, gender, etc.) with different variable (i.e. answer to a certain question). 1. To perform the test, click Statistics: 2. Select Wilcoxon Signed Ranks and then CONFIGURE TEST: Here is what the following screen will look like:

15 Now lets walk you through the steps necessary to complete the analysis 3. Make sure the correct table is selected in the Table dropdown bar. In this case the table is Pre_Post_Prophets_blank_for_missing:

4. Select the pretest question in the dropdown bar for Group A, and the posttest question in the dropdown bar for Group B. For an example we will compare Question 11 (Q11):

16 Note: only pretest and posttest question scores on the same question can be compared. It does not work to compare one pretest question result with a different posttest question result. Also, SOFA will only compare the answers for the individuals with an answer for both the pre and posttest. It would skew the results to include answers of only the pretest or the posttest. 5. Click Show Results. This will run the analysis:

17 6. Here are the results: Lets walk through the results. As mentioned, this test uses math to determine whether or not the results, or responses to survey questions, would happen by chance. So we need to look at a few things in the results page; the sections of the results are numbered that we will explain. 1 2 3 4 5 1. This is simply a summary of the table the data was used from. It tells where the results were saved (Output also saved to), and whether or not any data was filtered from the analysis.

18 2. This is perhaps the most important this in the result output. This number signifies whether or not the results were statistically significant (small chance of just happening). The smaller the number, the better. When the number is 0.05, it is considered to be statistically significant. A p value of 0.05 signifies that the results have a likelihood of occurring 5 out of 100 times, which is not likely. A number smaller than 0.05 signifies an even smaller probability of happening by chance. 3. As it says, this is the Wilcoxon Signed Ranks statistic; it is simply the mathematical output of the analysis. It is not as important as the p value. 4. The table summarizes the data involved in the analysis. Variable is the question used; N is the number of entries or responses included; median is the middle of the answers; min is the smallest answer; and max is the biggest answer. The table helps the user understand how the answers changed, if they did. For example in this analysis, the pretest had answers from 1 to 5, whereas the posttest had answers from 3 to 5. This is an improvement and should be framed that way, but it is not statistically significant. Note: It is extremely important to know what 1 and 5 stand for when running these tests. Here, 1 is strongly disagree and 5 is strongly agree, with the others in the middle, and this greatly affects how data are interpreted. 5. This is SOFAs explanation of the p value.

Lets run another analysis to give another example. This time we will use question 16 from the survey, which is, I can make a difference in my community. After selecting the pre and posttest responses in Group A and Group B respectively, here are the results: After completing this test, you should first look at the p value. The p value is 0.034. This means the results were statistically significant. If the results were presented in a journal, they might look like this: A Wilcoxon Signed Ranks Tests revealed a statistically significant change in youths perception of their ability to make a difference in their community with pre-tSB program average (Mdn = 4.0) and post-tSB average (Mdn = 4.0), n=22, p = .04. You will not be submitting this to a journal, and writing the results like this makes it hard to know what it means, so here is another way of interpreting the results:

19

The Wilcoxon Signed Rank result for Q16 provides support for the Service Boards positive influence on youths perception of community adults valuing of youth. And a third way: TSB students were more likely to feel like they could make a difference in their community after they participated in tSB.

For the video, click here.

20

VII. Additional Resources

YouTube Channel with instructional videos: http://bit.ly/ReNevY SOFAstats own page: http://bit.ly/MwPUSB More videos: http://bit.ly/QA4CVx