All survey sponsors crave more participants. More participants mean more data, which signals interest in the topic and fertilizes robust analytics. But all good surveys must come to an end, so when do you decide to stop accepting new responses? For a one-shot survey, the curtain has to come down; if you plan to return to the survey next year, you will be more open to keeping the survey open. I have encountered these nine stop signs.
Every set of survey data has values missing. This person didn’t say how many people report to them; that person didn’t fill in base compensation. Worse, missing values cause many algorithms (notably multiple regression) to drop all the data for a person if even a single number missing. So, if your survey data aims to predict your law firm’s fees for reviewing securities law filings and you have completed 65 such matters, if 10 of them are missing total fees, you have an analyzable set of only 55 matters.
A “script” is the term I will use for the programming an analyst does to turn raw survey responses into groomed output, such as tables, plots, or complete reports. Scripts have huge advantages over spreadsheet manipulations.
Reproducible: The small chunk of R code shown below reads in the Excel file that was exported from my hosting software, renames several variables (column headers in Excel) to make them more tractable, and drops a few columns that won’t be needed for the next step of the analysis.
When you present survey data, you can choose among five primary modes: text, lists, tables, charts, or infographics. Each serves a different purpose and demands different skills from those who use them. Each mode has its strengths and weaknesses. Let’s consider them in order of their difficulty to implement.
Text has the obvious advantages of being easy to produce and readily understood. With word-processing, you can apply bold, underline, italics and size to emphasize different numbers, but mostly it is a strung-along verbal description of numbers.
We can cite several advantages to using online surveys for data collection:
Respondent convenience: Online surveys can be accessed from any device with an internet connection, which makes it easy for respondents to complete the survey at their convenience.
Cost-effective: Online surveys are less expensive to administer than traditional paper-based surveys, as they do not require printing or mailing costs. As compared to mailed surveys, you also know more about whether a survey reached its recruitment target when you send out emails.