Bench Test

A good way to ‘bench test’ your survey instrument for your group project (make sure your questions, scales, and answer format are going to yield the appropriate data for analysis) is to populate a survey's data set using random numbers.

Here, I created a data set that organized data from various sources (survey, a calculated column based off of survey data, and web traffic data from server logs or Google Analytics), that you might see from surveying people using and giving feedback about a new website.  To see  if these questions and coding would work for the ultimate analysis, I  used various ways of generating random numbers in Excel, such as the Random Number Generator function in Analysis Pak (good for continuous measurements), or the RAND (good for customizing with formulas)  or RANDBETWEEN function (good for mimicking scaled response questions, binary responses) to generate a sample number set. SS DB5 template filled in.JPG 

[pic 1]

data set for DB6.xlsx   use this template here to organize your data, and your output.

• Using the same column names I used, generate your own random sample data that would be appropriate representations for each column.  Create a sample size of between 30-70 respondents.
• Create three additional columns representing something you would collect in addition to the ones displayed here, and generate appropriate sample data. Hint, look in Google Analytics for ideas! It can be another user segmentation question such as gender or year born, a respondent's perception question such as the 'liked best or least' questions, or, it can be a performance measure such as Seconds per Page, or number of # Page View. Try to do one of each. Take a screen shot, and describe these new columns, and why you think they are good 'new variables' to include in the survey?

With the appropriate Data Columns:

1. Identify what type of data is for each column (nominal, ordinal, interval, ratio). What other types of graphs or statistics can you apply to describe this data? 
2. Create a Histogram chart (Chart function) for each appropriate column of data, and describe the following:

1. Alter the bin range or number of bins to visually depict the best way to portray the data's shape. What is this shape telling you about your data set?

1. Use Descriptive Statistics function in xls. Analysis Toolpak to generate the appropriate column’s data statistical characteristics. Which columns are appropriate to use this on? Demonstrate those approaches by taking a screen shot  and include a brief description. What is this analysis telling you about your data?
2. Use the Correlation function in xls. Analysis Toolpak. Which columns does it make sense to use this type of analysis on? What does this analysis tell you about the correlation between the variables? Do you think there may be correlation with or without causation?
3. Use the Regression function in xls. Analysis Toolpak. What would you use as your Y (dependent variable) and your X (independent variable(s)? What does this analysis tell you about the association between the dependent variable and the independent variable(s) you chose?

Bonus Question:

Create a cross tabulation table (Pivot table in xls) using appropriate columns. Feel free to sub-group columns (use a range to consolidate responses into groupings) to make the table easier to understand. Create a 2 variable and 3 variable Cross-tabulation tables. Describe what your tables can say about your data.

Attach your generated data with your additional columns as an excel spreadsheet. Post your output in the message thread using screen shots of your output with your descriptions/analysis.