MouseHouseApp Blog

Presenting Your Data

Posted by Sophie Holtzmann

So! You’ve finished your months-long research project of precise measurements and recordings, and now what? Was your question answered? Did you ask the right question? Does your data even address the question? All of these questions rely on the ability to take data and effectively analyze and present it. But that still leaves the very large question of how to do that. In this post we hope to outline the rough do’s and don’ts of data presentation, so that you can maximize the output of your data and effectively communicate that to others so that you may garner more feedback and more conclusive results.

presentation-skills-of-PhD-students-practice-bullet-points

The first step is figuring out the actual data. To begin, let the data lead the discussion of conclusions. Do general statistical analyses look for interesting trends in data lines. “General statistical analyses” will vary greatly based on your field and area of study within that field. For example, if there’s a lab which works with mice, variables such as mouse weight, hair shine or hair loss or gain could show interesting trend lines that may not necessarily have been a part of the initial goal of the study. If working in lab where there are task based responses being measured, accuracy and response time may show interesting trends which once again may not be directly involved in the hypothesis. A general analysis of data will vary greatly by field, but these basic analyses can occasionally show interesting trends. Similarly, the lack of anything abnormal in these analyses can also be helpful, in that they show that there were no subsidiary effects from the manipulation. These trends can tell you insights about your experiment that you may not have previously considered. 

Once this general analysis of measurements has been made, you can begin asking the questions you want answered and manipulating the data to find a conclusion, whether it confirms or contradicts the hypothesized result. It’s important to determine the questions first, so that you analyze the data to answer your question, and not to get a specific result. Obviously the specific analyses you’ll use will be wholly dependent on the question you’re asking. In a very base sort of example, the statistical measures of mean, median and mode all have very specific purposes based on what you’d like to know, and many mistakenly use them interchangeably. Similarly, be sure that the measurements and statistical analyses you do are actually answering the question you ask. So if I administer a test, if I want to know how well the class did, I can take an average for questions answered correctly, but if I want to know how well students did on any given part of the exam, I should look for modes on the questions students answered correctly. This example is just to highlight that not all analysis are created equally and it will greatly hinder results if the methods of analysis are not thoroughly thought through. 

Now, assuming you have run effectual analyses on the data, you can move forward in attempting to present that information, whether to your lab, colleagues or in a publication. There are always multiple ways to present the same data, but once again they are specific to the question the visual aid is attempting to meet. In a base example, the same numbers can be presented in a bar graph or circle graph. However, a circle graph is only useful if the numbers are significant as parts of a whole, whereas with bar graphs, the numbers can be of importance on their own, rather than in relation to one another. A good measure of decided what visual representation is best is to determine what matters. For example, do the details matter or does the big picture matter? If details are important, perhaps a table is the right tool to display that, or if the big picture is what matters, graphs or trend lines may be more helpful. One thing to definitely avoid is using vague visual aids when the details are important. This can lead viewers to believe the data is being manipulated in a misleading way, regardless of whether it is or not. 

So as soon as you have the visual representations chosen, it is still important to accurately and effectively convey this information in a presentation. If doing a formal presentation, there are important logistic considerations to take into account. These are small, detail-oriented solutions, but the lack of these considerations could ruin a presentation. If using a presentation aid such as powerpoint, be sure to take care to make the slides clear and readable from far distances. For example, use bright colors and fonts in sizes upwards of 24 in Sans Serif, or other very clear text. As a general rule for presentations, SHOW data and TALK conclusions. This simple mantra keeps listeners engaged with you and the fewer words presented on screen raises the probability that they’re listening to you versus reading the screen. Order the slides and information such that there is a funneling effect for the listeners. By funneling effect, we mean starting with broader concepts and more base understandings to more specific and complicated nuances of the study or data. Make sure to take the audience with you, and explain each step in logic which you went through to come to the conclusions you came to. 

If you work in a collaborative manner with a larger general lab, you can also offer to make all data open and accessible (based on the sensitivity of the data of course), so that others can bring their perspectives the data and perhaps handle it in manners which may not have occurred to you. This way you can maximize what your data can do for you and your experiment. This transparency also gives your presentation more validity. 

So in conclusion, with the confusion of data, it can be very simple to give it order and find real solutions in your analysis. The presentation of the data can be the greatest tool or the greatest hindrance based on how effectively it is done. An important thing to remember is that even if the data disproves your hypothesis, this is still okay! A falsification is still a conclusion and can be important in aiding future research. With these simple tools, hopefully you can present your data effectively and accurately.

Topics: How to

Subscribe to Email Updates

Posts by Topic

Connect