Giving science talks

Some advice on slide design and presentation so your audience doesn’t cringe or fall asleep.

Choosing your presentation software/device

  • Omnigraffle Professional: This software is the most fun to use.  “Fun” is relative, I suppose, but this program is addictively productive.  I use it almost daily for making graphics and notes, but the “Pro” version allows slideshows.  You can add mouse-over notes for each element on a slide.  But doesn’t do animations, play embedded movies, or have built-in equation editor. I also doesn’t have good slide sorter interface, so it’s not convenient if you have dozens and dozens of slides.
  • Apple Keynote: Best feature, in my opinion, is that elements are instantly draggable (unlike PowerPoint).  Also integrates easily with iPhoto, iMovie, iThings, etc.  If you are Mac based, this is the program to get, even if you own Powerpoint.
  • Microsoft PowerPoint Presentations always look like they have been designed by first graders.  Actually, some presentations I’ve seen first graders make are better.  I call it Microsoft PowerNap.
  • Google Doc Presentation: Works great, but not at all when the internet is down…but you can (should) always export a copy to physical drive beforehand, just in case.
  • Overhead projector: If you have images and objects to show, this is the way to go.  The advantage to overhead projector is that you can physically write on paper while audience watches…that can be good.  If you use plastic sheets, you can erase the information later if you’d used dry-erase markers.

Presentation content

The following recommendations are for a 20-minute research presentation.

Title slide: Title, your name, institution, and e-mail address if you think people would be interested in contacting you. This slide will be displayed as you are getting introduced. Don’t read the text to your audience, unless you know that it contains people who clearly cannot read. Make sure you have an interesting photograph on this slide, too. [2 min (if your host is brief), 1 slide]

Introduction: Show some pretty visuals (photographs, SEMs, drawings, paintings, Quicktime movies, sound files) of your subject and get your audience excited about the “issue” or question; put your issue in the context of several already-published articles from the primary literature; summarize (very briefly) your past research, if any, on the topic; provide clear statement of hypothesis, and give “road map” of what you will talk about in the rest of the talk. Avoid slides with a lot of text. [2 min, 4 slides]

Materials and methods: Show more pretty visuals of your organism and justify (if you haven’t already done so in the Introduction) why your experimental organism is perfect for addressing the issue mentioned above; show experimental equipment and methods (ideally including a photograph of a person doing something); show experimental design (with sample sizes); mention statistical analyses that were used and how they will address hypothesis. Avoid slides with a lot of text. [Approximately 2 minutes, 4 slides].

Results: First mention whether experiment worked (e.g., “90% of the birds survived the brainectomy treatments”) and show some photographs that show interesting or comforting qualitative results (a surviving bird looking content); remind audience of specific hypothesis; present chart of data and explain whether hypothesis is supported; explore data (e.g., “I noticed something unexpected”); end with brief summary slide of main results. [7 min, 4 slides]

Here are some thoughts about graph design:

Use line plots (below) to show means (=averages). Much better than bar graphs because you can easily see the variation around the mean; also removes the temptation to fill bars with different colors for no reason! Error bars can be standard errors, standard deviations, etc., so be sure to specify which in figure legend. Also specify the sample sizes for each treatment. Comparison of means can be done with unpaired Student’s t-test, analysis of variance (ANOVA), etc.

Use box plots (below) to displays medians (when one or more groups have non-normally distributed data). The “box” contains 50% of the data points, and the middle line of the box is the median. The tips of the projecting bars show minimum and maximum values. There are ways ways to make these, so it is critical to explain graph elements in the figure legend. List sample sizes, too. Comparisons of medians can be done with Wilcoxon rank sum tests, Wilcoxon signed rank tests, and Kruskal-Wallis tests, among many others.

Use scatterplots (below left) and regression plots (below right) to show relationships between continuous variables. Regression assumes causality (X causes Y, e.g.).

Use bar graphs (below) to show count (=discrete, discontinuous) data. Statistically compare counts with goodness of fit tests and contingency chi-square tests, for example.

Conclusions: Discuss why your results are sound and interesting (convince audience, too); describe relevance of your findings to other published work; state relevance to real organisms in the real world; talk about future directions. If you have more than 1 slide that say, “Conclusions”, rename the earlier ones so that the audience doesn’t actually think you are concluding. [4 min, 3 slides]

Acknowledgements: I hate this slide, and you should, too. Ending a great talk with a laundry list of thank yous will put your audience to sleep, and will put a huge void in between your amazing conclusions and the questions from the audience. If you need to thank people, say it on an early slide (title slide, perhaps), and then mention specific help you received during Materials and methods section. [0 min, 0 slides]

Questions: Leave your conclusion slide up, so that your audience can refer to it. Audience will invariably want you to go back to Results slide X…so you should have a handy list of slide numbers so you can quickly jump to that slide. On Powerpoint, type the #, then return. [3 minutes]

Note: If you are giving a talk in a 60-minute time slot, practice your talk until it is approximately 40 minutes long (and with approximately 40 slides). What happens to other 20 minutes? You will probably be introduced 4 minutes late, your host will spend up to 5 minutes introducing you (more, if you are infamous), interruptions (questions) during your talk with bleed away 5 minutes, and then you will have 5 minutes at the end of your talk for questions, yielding a seminar that is exactly 59 minutes…which gets your audience back to their lives with a minute to spare. Although your host and perhaps several members of the audience may be willing to stay indefinitely to hear you answer questions, try to remember that the majority of your audience will become impatient if you drag the seminar beyond the one hour mark.

[Parting slide]: Optional slide showing “for more information” such as your e-mail address, laboratory web page, and suggestions for further reading (journal articles, books) for those interested.

General layout and design considerations

  • Construct all slides in “landscape” orientation. This assures that the tops and bottoms of slides fit onto the projection screen and that view from the back of the room is not obscured by the heads of people sitting up front. If your image is inherently vertical, shrink it so that it fits horizontally (do not, as some have done, have separate presentations for wide and tall slides).
  • Even with horizontally-positioned slides, try to keep the important information in the top 90% of the slide area. Often the bottoms of slides cannot be seen by people in the back rows because people’s heads are in the way. This is not the case in all lecture halls, but if you are unsure of the room, or unsure of how tall the audience members will be, plan cautiously.
  • Unless you have a lot of free time, it is usually advisable to construct slides in your presentation with black text (and black illustrations) on a white background. This strategy makes the task of printing handouts infinitely easier and far less wasteful of toner ink in laser printers and photocopy machines (i.e., dark backgrounds use ink, a white background does not). And illustrations and charts are invariably first produced with black text and lines, and conversion of all black elements and text to white involves substantial tweaking and fussing. Finally, when your background is white, you are ensured of at least some light in the lecture room, which is important in a lecture hall when you want the audience to be able to see your face or hand movements or want the audience to stay awake (Microsoft PowerNap is visual chloroform to many people these days, so brightness is an issue!).
  • When the lecture room is completely dark, however, it is of course true that white (or light) text and drawings on a dark background improve clarity of the slide. A blue background is particularly popular, but particularly annoying for people younger than 53. Also, if you choose a colored background, be aware that certain text colors may not be discernible to people with some types of color deficiencies.
  • Whatever background scheme you choose, just be sure not to mix the two within a single presentation or within a single slide (i.e., “don’t cross the beams, Egon”).
  • If you have only a photograph on a slide, set the background of that slide to black.
  • Adjust font size so that that approximately 10 words fit horizontally (24 point is usually a good size), and line spacing so that only 10 lines would fit per slide. Naturally, if a slide contains scanned text as part of a scanned illustration, be sure that the font size is comparable to 24 point … or be prepared for the audience to squint and whine.
  • Do not use more than 2 or 3 typefaces in a presentation.
  • Your audience will read 100% of the text on a slide, so delete any text that is not essential.
  • If your research area is fond of abbreviations, don’t assume that the audience will remember for more than 2 slides what DDCP-2A’ or URAQT signify. Remind the audience verbally or put a “definition bar” at the bottom of later slides so that even people who slept through the earlier slides will have the ability to follow your presentation if and when they wake up. In fact, avoid using abbreviations if at all possible–save them for the manuscript, which is more likely to be read by people in or near your research area.
  • Use italics instead of underlining.
  • Avoid using strings of all capital letters in slide titles (and elsewhere). Strings of all capitals are very difficult to read quickly, and also obscure information inherent in allele names (e.g., Adh) and other nouns that possess mixtures of capitalized and uncapitalized letters (e.g., rDNA, species names). For example, compare “THE FATE OF ADH MUTANTS IN DROSOPHILA MELANOGASTER POPULATION CAGES” and “The fate of Adh mutants in Drosophila melonogaster population cages.” In addition, “all caps” is usually interpreted as the print equivalent of YELLING!  If you really like using all caps, you are probably old. Or you have a broken keyboard.  Things happen.
  • Similarly, not put slide or figure titles in “title” case. Compare “Effect of Cheap Malt Beverages on Adh Expression in Wild Type and Mutant E. Coli” with “Effect of cheap malt beverages on Adh expression in wild type and mutant E. coli.” In the latter version, a reader can “get” the meaning of italics and capitalization with less brain power.
  • Do not include repeated banners, logos, or backgrounds on your slides.  They are often pretentious, always distracting. Really. If you have some innate desire to use a logo, up your meds or at least restrict the logos to your opening and ending slides.
  • Do not use transition fades, bouncing text, or swhooshing noises. Some in the audience will politely chuckle, but most people will be silently cringing. Using these gimmicks is simply an admission by the speaker that the topic is uninteresting, or that the presenter does not think of himself/herself as a strong public speaker. If you’re still in first grade, though, go crazy.
  • Do use animations of images when movement is actually part of your research. Perhaps you are studying sprint speed of red and green lizards, and have good photographs of each: animating these images to move across the screen at different speeds would be an amazingly effective way to express results to a sleepy audience, though you might want to retain the bar graphs, too (or you could have a chart timed to appear after the lizards are done running).
  • For true animation, you will first need to compose in a program that is capable of producing “animated GIFs”, and then find a program that can convert the file to a Quicktime format. Powerpoint may eventually have the capability to show GIFs, but I’m not sure when this feature will be introduced. You can also approximate the effect within Powerpoint by simply placing a series of images in consecutive slides, and then advancing through these slides at a good clip (you can even specify that you would like Powerpoint to advance a specified range of slides at, say, 10 frames/second).
  • Time-lapses using a series of scanned photographs can also be constructed with GIF-building programs. Again, however, this file would need to be converted to Quicktime before it could be used in a Powerpoint presentation.
  • Do import movies–audiences these days assume that if you are using Powerpoint, that you must have a great movie hidden amongst your dry text and chart slides.  When using more general search engines, include “Quicktime” and “movie” along with the words specifying your quest. There are thousands of Quicktime files out there, so be optimistic.
  • Do use sounds if they are integral to your research. Bird song, bone shattering, etc.  would be great additions to a talk, and would eliminate the often problematic use of external tape players. You can set options for each sound, and have the sound-activation “icon” show next to photographs (e.g., “listen to the sound of the femur failing, again, and now compare that to massive skull trauma”).
  • Avoid mixing green and red on a figure–members of your audience may be red/green colorblind, the most common type of colorblindness. The obvious exception is when you are giving talks around Christmas.  Color deficiencies of some sort affects 32 million Americans (out of a total population of 300 million) –it is distinctly not rare, especially among white male audiences, which is sometimes your entire audience in some disciplines. Read up on colorblindness. If you are using Photoshop to tweak figures, the program has color deficiency tools built in..
  • Use figures instead of tables whenever humanly possible.
  • Graph titles are not appropriate for laboratory write-ups and manuscripts, but they are fine for slides:
  • Small icons near lines draw viewer’s interest

    If you can ever add miniature illustrations to your graphs, do it. Visual additions help attract and inform viewers much more quickly than text alone. Tables benefit from this trick as well.

  • Most graphing applications automatically give your graph an extremely annoying key that you should delete; directly label the different elements, instead. Interpeting keys is sometimes very difficult, and you should do anything in your power to make your graphs easy on the brain.
  • Use arrows to direct audience’s attention to particular parts of charts, especially for complicated figures. For figures created in charting programs and then exported to Powerpoint, make sure that the output fonts and line widths are legible once the image is scaled within Powerpoint. If you keep shrinking your graphs, your audience will squint for a few slides and then give up out of annoyance. You don’t want to make your audience mad.  Somebody might be packing.
  • If you have a complicated chart, it’s a good idea to add statistics directly onto the chart. For example, if you have conducted an ANOVA and some post-hoc comparison of interest, you might use lines with arrows to indicate which means are significantly different from each other. Next to the line you can say, “means sig. diff.” You can give further details orally.
  • Horizontal text on y-axis makes reading easier

    Orient the y-axis label to be horizontal (e.g., the “Growth per week” in rat slide below) whenever you have the space. Vertically-oriented labels are substantially harder to read, and often require the audience to tilt their heads (which tempts them to sleep, I suspect). People with neck braces and football players are always especially appreciative of horizontal text.

  • Never display two-dimensional data in three dimensions. Doing so marks you as a novice.
  • If you steal images off the internet, don’t steal bad ones. Make sure that you have selected the correct size for a computer display (at least 1024 pixels, usually). For a great tutorial, see this PDF: http://pandora.tcs.tulane.edu/art/Images_for_Powerpoint.pdf.
  • If you use an image or graph that you didn’t produce, cite the author by name or with a URL.  On. Every.  Single. Slide.  No exceptions.  And for the love of God, if you steal images without permission, never, ever post your Powerpoint file on Blackboard, Moodle, or the Internet…it will eventually be found by the author and s/he will write you a nasty take-down request, probably CCing your boss.
  • If you have a photograph that you would like to scan and include, be aware that the maximum resolution you would need (for an original image sized 11.25″ wide and 7.5″ high, the dimensions of a Powerpoint slide) is 72 dots per inch (dpi).
  • If you scan a small image (e.g., 1″ square) adjust the scan settings so that your final scan file is approximately 1.5-2.0 MB (e.g., your scan will be at some resolution far greater than 72 dpi). In Photoshop, adjust your image size under Adjust:Image (with “Resample image” turned off) so that the image is 72 dpi; turn “Resample image” back on; then limit the file to be no larger than 11.25″ wide or 7.5″ tall (the limits of a Powerpoint screen). In doing this your saved image (JPG or PICT format) is optimized for Powerpoint and should rarely be larger than 300k.
  • Because projection resolution is currently so poor, try to keep scanned illustrations as large as possible, especially if you wish the audience to read the finer print. A very common error in Powerpoint presentations is to shrink scanned illustrations (e.g., the citric acid cycle) so that a really large title can be included at the top of the slide. Just omit the title and expand the drawing to the limits of the slide dimensions. You’ll be glad you did.
  • If you overlay text onto a scanned image, apply a shadow to the letters so that the words are legible in both white and dark portions of the figure.

Delivery

  • A big part of giving a presentation is connecting your computer to the display projector successfully. Don’t wait for somebody else to do it for you. What this means is that you (you) need to buy or borrow a connector dongle and know how to use it. This also means that you (you) will need to power on the projector.
  • Your graphics and text might look fantastic on your laptop but terrible on the screen when presented. You should therefore test your presentation in the room where you will give the presentation. And do this at 24 hours in advance in case you need to redo everything.
  • Do not rely too heavily on your notes. It is irritating to the audience, who value eye contact. Aside from eye contact, people who deliver talks too closely to pre-written text tend to deliver really boring talks. If you have a tendency to write boring talks, try not writing out every darn word…you might find that you speak better than you write.
  • Do not keep “checking” to see whether a slide is still there. It almost invariably is. If there is a power outage or your computer falls off the table, then go ahead an check.
  • Do not read your slides to your audience. If you DO read the text out loud, you should preface it by saying, “Because some or all of you cannot read, I’m going to read this for you.”
  • When verbally referring to a specific portion of a slide, use a pointer or your finger to orient the audience.
  • When using a pointer or your finger, it is best to touch the image on the screen rather than situating the pointer in the projection beam to throw a shadow.
  • A stick or a finger (yours) is almost always better than a laser pointer (the projected dot is generally too small on cheap laser pointers). Laser pointers are, however, excellent cat toys ($7.99 at Targét).
  • If you must use a laser pointer, do not blind people by directing the beam in their eyes.
  • Do not chew gum, fiddle with your jewelry, or wear a hat — even if these things comfort you or are critical components of a carefully constructed persona.
  • Do not put your hands in your pockets. If you are likely to forget, fill your pockets with pushpins beforehand. If you are a hemophiliac, put something soft but unpleasant in your pockets (so you don’t bleed out from pushpin wounds).  Or wear a pocketless skirt (if you’re an XX or XX wannabe).
  • Do not draw more attention to bad slides by apologizing for them. If you do feel compelled to apologize for a slide, fine. But don’t preface every darn slide with an apology…that will make audience want to kill you.
  • Resist puns, obvious jokes, and overly rehearsed humor. Spontaneous, genuine humor is so much better.
  • Minimize your use of the crutches, “OK,”"like,”"um,”"er,”"sort of,”"ya know,” and “kind of.” Especially “like.” It’s sort of like, you know, when you use lots of filler words, it’s, like, people totally don’t even listen to you anymore, and, well, kinda sort of think you’re dim. Bored audience members have been known to actually record the number of “likes” in talks.
  • Etc. is pronounced, “Et cetera,” rather than,”Eck cetera.” It’s true, even if somebody you respect insists that you should say, “eck.” Moving forward, you should be suspicious of anything that person says.
  • Adjust your speed or ask whether there are any questions when you notice confused looks.
  • If people in the audience start closing their eyes, it is a sign that you are boring them. Speak up and become more dynamic.
  • When responding to questioners with faint voices, repeat the question loudly for the benefit of all. It’s a strange but true fact that older people, especially those with bad hearing, often sit in the back rows, so make this a habit for all questions.
  • Attempt a response to all questions even when you think there is an audience member who might be able to field it better than you.
  • If you want a blank screen to appear (for instance, to get the audience’s full attention), press the “b” on the keyboard (at least for Powerpoint). In a pinch, just ask somebody in the back of the room to stand in front of the projector beam.
  • If the moderator does not end the question period in a timely fashion, say, “Perhaps I could entertain further questions outside?” People will cheer.


Moderating

  • During the talk, if a person starts playing “air piano,” find a way to stop them so that nearby members of the audience can focus on the speaker. Hand them a note that says, “Hey, Mozart, you are distracting everyone.” Or, before the talk starts, ask people with large hats and annoying habits to please sit in the back, where they are less likely to be distracting to the rest of the audience.
  • Similarly, if somebody starts compulsively clicking their pen several times per second, go to the offender and give them a writing instrument that doesn’t click. Remind the person to take their medication before coming to talks.
  • If the speaker starts to go beyond the agreed-upon stop time, stand up and situate yourself near the front of the lecture room so that the speaker gets the hint.
  • If the speaker refuses to stop, pull out the taser. Usually, just showing the device is sufficient.

Useful literature

  • Briscoe, M.H. 1996. Preparing Scientific Illustrations: A Guide to Better Posters, Presentations, and Publications, 2nd ed. Springer-Verlag, New York.
  • Day, R.A. 1994. How To Write and Publish a Scientific Paper, 4th ed. Oryx Press, Phoenix.
  • Matthews, J.R., J.M. Bowen, and R.W. Matthews. 1996. Successful Science Writing: A Step-by-Step Guide for the Biological and Medical Sciences. Cambridge University Press, Cambridge.
  • Pechenik, J.A. 2006. A Short Guide to Writing about Biology, 4th edition. HarperCollins College Publishers, New York.
  • Tufte, E.R. 1983. The Visual Display of Quantitative Information. Graphics Press, Connecticut.

Useful links

PowerPoint in high education…Rebecca Schuman
“The Cognitive Style of PowerPoint”
(Edward Tufte)
“Death by PowerPoint” (flickr group)
“The use and abuse of PowerPoint in teaching…” (Allan Jones)
“Best Powerpoint slide. Ever.” (Gary Turner)
“The best PowerPoint presentation ever” (Doug Zongker)

Using this site

Anyone is allowed to link to this page, but please don’t download the darn thing and keep it on your site.  If you need a citation:

Purrington, C.B. Giving science talks. Retrieved <today’s date>, 2011, from http://colinpurrington.com/tips/academic/sciencetalks.

Comments are closed.