The Art Of Collecting Your Own Data
It has been a good six weeks into university, and it is finally recess week! Over the course of these few weeks, I have created and sent out more surveys (6, 5 of which was for my quantitative reasoning (QR) module, UQF2101J, the Pursuit Of Happiness, in the National University of Singapore (NUS)) than I have ever done in my life, and through this I truly felt like I learnt the art of being able to collect my own data.
Collecting data initially seemed daunting to me, since I used to associate ‘data’ with computer analytics and numbers, which I was terrible at. So to collect my own data, much less find datasets online, intimidated me at first. But pages like google forms and survey monkey are easy tools to use to collect data, and they are extremely user friendly too! Such websites definitely made the experience easier for me. Personally, I prefer google forms, since it’s free to get access to all functions, but if you’re not too concerned with prices here’s a more detailed breakdown on the differences between the two:
Now onto why you should collect your own data. Firstly, you can collect data on whatever you want! You don’t need to scour the internet for hours on end for online datasets in hopes that you find one that exactly fits your hypothesis (but of course, if you do, that would be great!). This allows you to collect data on specific questions.
You can also build on existing data that you already have. For instance, perhaps you have more questions pertaining to a topic that the dataset you found doesn’t have. You could then collect your own data instead to expand on this! This was what I did with my group mates, Min Yi and Kok Lee, in my class, on perceptions of happiness after we went through the data from the IPSOS survey on the perils of perception. Which leads me on to my next point, creating surveys allow you to collect data on how people around you perceive a certain subject! Then you could compare this with how things are actually like and see if there are any noticeable, interesting differences.
You could also spin-off questions based on the survey questions that other organisations used in collecting data. In my recent ongoing mini project on perils of perception, we actually replaced the original survey questions with emojis instead, so as to see if emojis as a visual aid would change the answers of the respondents, which was really fun! You could also collect your own data if you feel like you do not agree with the findings of another dataset. In a recent collaboration with Erika, also in my QR class, we did not agree with the WHR’s methods of collecting data, so we created our own survey to see if we were right! I won’t spoil our conclusion for you, you can take a look at our articles in our publication here:
Collecting your own data means you can get data that is more catered and relevant to your social environment. For example, if I wanted to collect data on university students, it is just easy to send the survey to my friends!
However, the above could be one of the very downfalls of collecting your own data. This means that your dataset is often a product of convenience sampling, and is a statistic not a parameter. The results of the survey cannot be extrapolated to the entire population, which leads me on to a concern I had: the lack of responses to my survey. I was afraid that people would be unwilling to do my survey and I would just end up with, 10 responses, all of which came from my close friends who I would, for lack of better word, coerce to do. Fortunately, I did not have to worry about this problem, and I will elaborate on this in the rest of the article. However, despite this, the sample size of self-collected data is usually small, and not randomised. This is because you are technically only collecting data from people you know and hence they might be of the same demographic as you, making the data set not as diverse as one that was for example, done on a large, international scale like the World Happiness Report.
Now on to my personal experience collecting data. The creation of the forms was relatively easy. Google forms had many options, multiple choice, short answer etc, so designing the survey was the easier part. I was afraid of the lack of responses, but I learnt to be shameless, and just had to publicise my survey everywhere: I revived old group chats, spammed family group chats (so much so that my family members sent ‘Huh? Survey again ah? Different or same one?’), and even asked my friends to send the survey to their other friends. Also, when you create surveys with teammates, they will help to publicise too, so eventually my surveys managed to garner around 50 (first ever survey I sent out for my individual mini project, which you can check out here):
to 120 responses each, which I thought was not bad considering I thought I would only get 10 responses. Very, very conveniently, google forms allows you to export your results to google sheets, producing your dataset which you can then move to excel to clean the data! Cleaning the data was the toughest part for me. I learnt from Erika that it would be easier to replace worded responses with numbers, so that it would be easier to represent the data in graphs. This was also quite tough, since it was hard to find a way to represent the data that would be easy to understand to the layman, and is intuitive too. Sometimes, the least complicated looking dataset would represent the data collected in the most efficient and understandable manner! Google forms have pie charts and bar charts automatically created for your responses depending on the type of question you asked in your survey, whereas in survey monkey you can choose the way you want to represent your data if I remember correctly.
Some tips for those who want to navigate the Google Forms interface easily!
Also, non-QR related, the other survey I conducted was to find which logo my interest group liked the most, since there were various colour computations, whether or not they wanted an icon in the logo, placement of the icon, etc. So the survey actually helped me to come up with the best computation for the logo design that my interest group wanted best! So it helped my logo team and I easily decide on which logo we should go for that included the opinions of everyone. (This was also the most recent survey I had conducted, so I was glad that all the previous survey creations made doing this one alone much easier and quicker!)
So that’s all I have on self-collected data! Thank you for reading up till here and I hope everyone is doing well! See y’all in my next post!
