the principal goal of surveys is to inform and simplify the world around us by giving us key pieces of information somewhat ironically that goal of simplification involves many complicated steps join me and we'll walk through ten steps in conducting surveys keep in mind that these steps are broad brushstrokes they are easy to describe but often hard to execute step 1 know your population in their most basic form surveys are a tool to represent a broader group you have to decide that group and your reason for studying them early on are you interested in entire country a region or a rare population of certain individuals step 2 determine a mode will your survey be face to face online mail or telephone no survey mode is perfect for all scenarios you have to choose the mode carefully in light of your needs and resources step 3 create a questionnaire surveys need questions no-brainer huh in practice writing good questions requires painstaking consideration step 4 test your questionnaire so you think you wrote off some questions huh well when I sing in the shower I think I could tour with Conway Twitty hello old are all survey questions should be pre tested either by focus groups cognitive interviewing split ballot experiments field pre testing or some other methods step 5 get a good sample sampling is the magic of surveys your sample consists of the people you are trying to interview a great questionnaire will yield a junk survey with a bad sample step 6 collect the data this is where we find more tricky bits collecting data ultimately involves making phone calls knocking on doors posting HTML or mailing paper data collection must be rigorous documented and consistent step 7 organize and clean the data suppose you did a telephone survey with a caddy system just got a data set that looks like a messy spreadsheet with questions and strange orders scattered across multiple columns you also probably have many pieces of information that need to be rescaled & renumbered precisely and accurately this is also a good time to calculate response rates a quality measure consisting of the number of completed interviews divided by the valid sample step 8 wait the data now even if you did everything perfectly up to this step your final sample of respondents probably won't perfectly match the population post stratification waiting will now help you to correct many known imbalances step 9 analyze the data ain't nobody got time for that fancy spreadsheet survey analysis is fascinating addictive and extremely useful you want to present question order wording and basic descriptive statistics by creating a top line report then you might want to explore subpopulations with crosstabs you might even poke at causal relationships using regression graphics are cool too step 10 present the results if you conducted a good survey you just spent a lot of money to share valuable information with important people you now need to communicate your results to your audience in a way that is accessible and interesting finally give someone a high five because you've just gone through the process of a survey now let's do a quick review of those ten steps step 1 know your population step 2 determine a mode step 3 create a questionnaire step 4 test your questionnaire step 5 get a good sample step 6 collect the data step 7 organize and clean the data step 8 wait the data set 9 analyze the data and step 10 present the results 10 steps survey excellence you