Have you ever wondered what effect light levels have on people? Are we happier or more alert when light levels are high, and do low levels of light make us drowsy or more irritable? Investigating these sorts of questions not only leads to potentially fascinating results, the process of undertaking this sort of experiment is a great learning journey too.
In this project we have built a digital light monitor that is also used to administer and record the results of a questionnaire. We will take you through all the steps from building the tool to using Excel (or any similarly powerful spreadsheet program) to analyse the data.
All you need to do is generate one or more hypotheses: what are the possible effects of light levels on people? For testing we used the hypothesis that people tend to be happier the brighter the light levels are...
1. Building the survey instrument2. Setting up the questionnaire3. Using the survey application4. Administering the questionnaire5. Extracting and understanding the data6. Checking for correlations7. The code
1: Hardware required:
You need the following components to build the survey instrument:
2: Flashing the survey code onto the device
Connect IP02 or IP03 to the CS11 as shown:
1. With the CS11 attached to the IP02/3, connect the IP02/3 to a computer USB port.
2. Double-click the reset button on the front of the CS11
3. The CS11 will show up as a drive on your computer.
4. Drag and drop the software file onto the CS11.
Your CS11 is coded and ready to use
1: Response scales:
You can include up to 10 questions in any questionnaire.
Each question has to use one of the following response scales (when people answer the question they have to choose an available option):
2: Formulate your questions:
When you formulate your questions have the response scales in mind: the answers in the scale you choose needs to make sense.
3: Create a text file and save it onto SD card
1. Open a txt editor on your computer
2. Create a new file and save it as “survey.txt”. The name has to be identical to this.
3. Type in your questions in the format shown below.
4. When finished save survey.txt to the microSD card you are using
5. Insert the microSD card into the slot on the CS11
1: Power it up:
Ensure that the microSD card is mounted into the CS11 then power it up
When the program first powers up the XinaBox logo appears for a few seconds
The application then checks 3 things:
1. Is the SL06 gesture detector attached?
2. Is a microSD card inserted and correctly formatted?
3. Is there a file called survey.txt on the microSD card?
If there is a problem with any of these a helpful message is displayed and the application will not progress.
If all of the system checks pass then the screen shown will appear.
You can now begin the survey.
2: Starting the survey: measuring lux
When you swipe right above the instrument will take a light level reading…
- This is done automatically - you have 15 seconds to put the instrument near to the person you are interviewing to measure the light levels.
- After 15 seconds a light reading is taken – lux is measured.
- Ensure that you use those 15 seconds to place the instrument near to the person you are interviewing. The light sensor on the instrument needs to capture the same quality of light that the subject is experiencing.
- Control your experiment by applying the same approach to measuring lux for all of your subjects (e.g. ask them to hold it in their hands in front of them facing up).
1: Ask the question exactly as it is written:
If all has gone well you should now see a screen that shows you the question you need to ask and the response scale to use
Read out the question exactly as it appears then read out the response scale.
Ask the person you are interviewing to pick the one answer that best applies.
2: Enter the response
Once your interviewee has answered the question you need to input their answer. Wave your hand over the SL06 gesture monitor to select a response option
- The selected response option is shown at the top. Swipe left or right until the appropriate option is shown.
- Swipe up to select this option.
3: Confirm the response
To ensure that incorrect data is not recorded you will need to confirm the option selected.
Swipe up to confirm or swipe left to change.
When the selected answer has been confirmed the message “RECORDED” is shown on-screen for a few seconds then the next question is presented.
If all the questions have been answer the message “THANK YOU” is shown and the system resets, ready to start asking the next respondent.
1: Extracting the data
All the answers provided are recorded onto a csv file on the microSD card.
You will need to copy this csv file from the SD card onto your computer:
- Remove the microSD card from the CS11
- Use a microSD card reader to place the memory card into your computer
- Copy and paste the file DATA.csv from the microSD card onto
2: Making sense of the data
When you open the DATA.csv file it will look something like this (but without the colour coding):
Look at the first yellow block of cells highlighted above:
- There were 3 questions in the survey – the first yellow block shows all 3 answers from the same person.
- The data that is of most value is in the LUX and ANSWER(1-5) columns – this is what we will analyse.
- The Timestamp, question number, Scale Type and Answer columns are to provide context for the data we will analyse
In the image above you can see responses from 4 different individuals (see the colour coded bands).
1: Why are we weighting?
We asked some questions and each questions was chosen based on hypotheses, e.g. people tend to feel happier in brighter environments.
Each question is answered using a scale, and each scale is allocated a score: the more positive the response the higher the score:
- The scores are called weights: values we allocated to each answer.
- Note that a higher score in the Answer (1-5) column = a more positive response.
We can also say that the higher the LUX value the brighter the light is.
If there is a correlation between LUX and Answer (1-5) for a particular question, then this could be evidence to support the hypothesis. Lets take a look:
2: Segregating each question
The first thing we need to do is isolate data for a specific question:
- Open the data file in Excel and Highlight the top row with the column headers
- Click on the Data tab then select the Filter option.
- Small arrows appear next to the column headings.
- Click on the arrow next to the column heading Question #
- Choose one AND ONLY one question from the list.
3: Calculating the correlation coefficient
With only 1 question selected, copy and paste the data to a separate tab.
Use a cell in a Column to the side of the data and type in the correlation formula:
Compare the columns LUX and Answer (1-5)
VERY ROUGH guide to results. Where CC = correlation coefficient:
- A CC of > 0.5 or < -0.5 is very interesting: that is quite high / strong. A positive score this high is evidence for the hypothesis. A high negative suggests the opposite.
- A CC or > 0.25 or < 0.25 should not be completely ignored, but it does not support the hypothesis. Lower scores than these suggest our hypothesis is false.
- If CC = 0 or 1 or -1 then something has gone wrong – check everything