Institute of Solid State Physics


513.127 / 513.128 Solid State Seminar

Every semester, the Institute of Solid State Physics organizes a seminar where recent scientific results are presented. Sometimes external speakers are invited to present at the seminar but it is primarily a means to communicate internally at the institute. Every scientific member (professors, postdocs, PhD students, and master students) is expected to give presentations. The professors take turns organizing the seminar and it is traditional for the organizing professor to give the first presentation of the semester. The schedule of talks can be found via the institute website. Master students must give a presentation to satisfy course requirement in the master curriculum. The seminar course can be either "passed with success" or "failed". To successfully pass, the presentation must score a number of points that is sufficient to reach Grade 3 (Befriedigend). The maximum number is 100 points.

Nicht genügend (5) = 50 < Genügend (4) = 60 < Befriedigend (3) = 70 < Gut (2) = 80 < Sehr gut (1).

It is possible to receive a maximum of 55 points for your own presentation and a maximum of 45 points for attending the presentations of others. You acquire 3 points for each talk you attend in the Solid State Seminar series, up to a maximum of 45. As proof of attendance, bring a list of all talks attended to Elisabeth Stern when requesting a grade.

Grading of your presentation

Max 5 points: Abstract. Please submit a clear and interesting summary of your talk 4 weeks in advance

Max 15 points: design of transparencies. Consider the following aspects:

  • Figures (especially headings and tick labels) need to be large enough that they can be clearly seen from everywhere in the lecture room; typically this will mean (with exceptions possible) that there should be only one figure per slide.
  • The text on the transparencies must be well legible also from a distance (use large and typically bold fonts; typically do not use more than 5 bullets points per page). Typically avoid complete sentences on transparencies.

Max. 15 points: Good introduction to the topic

  • Clear abstract uploaded one month before the presentation.
  • Clear description of how the described work relates to the research field.
  • Clear description of the relevance of one's investigations / the described topic.

Max. 10 points: Clear presentation of the obtained results or when discussing a 'literature topic' clear description of the relevant details.

Max. 10 points: Clear and extensive answers to questions during/after the talks.

For talks with a duration of < 30 min, 10 points will be deduced. Talks > 45 min are to be avoided and also here points can be deduced (one needs to be able to describe a topic within a given timeframe).

Talk Guidelines and Their Reasoning

As the title says, this section contains a set of rather detailed guidelines that help creating good, well structured talks. We thought that such a document would be useful for two reasons: First, very often rehearsed talks are criticized for the same issues, which gets old soon. Of course, new students haven't heard the critique before, and can't know - hence this document. Secondly, probably because of time reasons, the reason for these strong opinions on several design issues is never made clear. Also this is something this document hopefully fixes. Of course, this makes this document rather long. Therefore, we have also summarized the most important aspects in a short checklist.

1.) Slide Layout

It's a simple psychological effect, but important nonetheless: People tend to believe things more if it requires less effort. In other words, if you have crowded, complicated slides, people need to invest a lot of work into deciphering what you've done. This effort makes them upset and/or tired, and they will leave your presentation with negative emotions. Conversely, clear, easy-to-understand slides allow people to focus on the science and get excited more easily. In short: Whatever is on your slide, make it easy to see. With that in mind, here are a couple of rules for designing transparencies:

Bigger is better: When you're preparing your talk on your computer, you're likely sitting very close to a quite big screen. This does not scale into an actual presentation, where you often have a small canvas and people in the back rows sit pretty far away. Furthermore, many people are older than you and don't have as good eyes. As a result, even if something strikes you as "too big" on the screen, in presentation mode it's likely not. Always try to use the space on the slide by making captions, figures, labels as big as possible. Hard guideline: Never use a font size smaller than 22 (exception: References). Always use bold fonts - they are much easier to read.

Only one idea per slide: There is no limit to the number of transparencies that you can use. Hence, there is no reason to clump things together. For each slide, make sure it tells exactly one thing - never more, since that only distracts from what you're currently trying to tell. Typically, this means: (1) Use only a single graph per slide. (2) Keep the number of individual text items (all related to this single idea) below or equal 6. Humans excel in very quickly spotting up to 6-7 items at once, but are known to struggle at a larger number of items simultaneously offered.

Graphs: Keep in mind that graphs are the major tool to convey your content. Be aware of the extra effort necessary to customize your graphs to support your story in front of your target audience. Be assured, this effort pays off!

  • (1) Make sure to plot the quantity that you want people to focus on, rather than deriving that quantity from a graph. For example, if you want people to focus on the evolution of the difference of two quantities, plot that difference, rather than the quantities themselves.
  • (2) Make sure that the linewidth is big enough. Here, a good guideline is 3-5 pt thickness.
  • (3) Make sure the lines are logically connected, but let them change only in a single systematic (e.g. they might show different coverages or different systems, but never change both at the same time). Try to reproduce the logic via the color of the lines (e.g. by going from hot to cold), but avoid different styles (long dash, short dash, dash dot, etc.). Different styles tend to be difficult to discern, especially if the projection is not very sharp.

Be tidy:

  • (4) It is very hard to focus on multiple lines at the same, especially if the legend is placed somewhere in a corner. Therefore, try using as few lines as possible - 3 is a good number, but never more than 5.
  • (5) Do not try to circumvent the rule of "only one idea per slide" by putting more information into a single graph. Instead, make sure the graph focusses on what you need to show. Choose the plot range accordingly. If required, highlight specific regions if they're important, and you can't avoid showing the rest (this happens typically when you show spectra and need to focus on individual peaks).
  • (6) A comprehensive graph must show clearly legible axis labels. It may further contain (legend) labels and text to discriminate between elements shown in the graph. Try to use common fonts and as little variation in the font sizes as possible, avoid shadings and coloring effects within letters. These simple measures make each graph to appear less busy.

Colors and contrast: Little known facts, that really help to avoid hard-to-read presentations, are:

  • (1) that beamers and monitors use different color spaces (monitors use RGB, while beamers use CMY). Hence, what looks like good contrast on your monitor might be all but invisible in a projection. As a rule, avoid light green altogether, even if it looks good on your monitor (yellow is fine, though). Also, specifically older beamers have trouble with contrast - try to maximize the contrast, especially when using greyscale figures.
  • (2) that one should never combine a red feature on blue background and vice versa. Human eyes/brains cannot discern these two colors when put in the same spot. The reason for this is that the eye receptor for "red" has a second absorption peak in the blue. Any effort to "resolve" the colors distracts and may even cause headache.

Write down important aspects: Frequently, people will have a hard time listening to you the whole time. They might not be able to hear you because of the acoustics, they might be chatting in between, maybe they have a hard time to understand your English. Also, it tends to be pretty hard to know what is really important and what is additional information. It greatly helps following a talk if the important parts (findings, intermediate conclusions, etc.) are explicitly written on the slides.

No sentences: You want people to listen to you, not focus on reading. Put only keywords on the slides, avoid full sentences. Never, under any circumstance, place a facsimile of a paper with highlighted sentences on a slide.

Equations: Even for people good in math, reading equations takes more effort than reading text or looking at a graph. For people not in the field, new, unknown equations are often highly confusing. It is thus best to avoid them as much as possible. Whenever you want to use an equation, ask whether it could be replaced by something more illustrative (often an animation, but see below). If you find that you really need that equation, there are a few things that you should pay attention to. First, make sure that while it is showed, you explain it properly (i.e., every symbol in there). Remember that you know what everything means, but the audience typically will not. Ideally, a slide with an equation should be self-explanatory, i.e. every symbol should be explained on that slide. Also, it greatly helps to focus on the important aspects if you use color or something similar to highlight some aspects of the equation - whatever it is that you will focus on during the next few slides.

Animations: Turning graphs into sketches that appear step by step can greatly help to facilitate the understanding, in particular when you're trying to visualize complex processes. However, those animations can be quite time consuming. More importantly, if they are too slow, they break your flow, which is quite irritating for the audience. Avoid "flying in" / "fading in" or similar animations - also since they tend to distract. If you really want them, make sure they are set to "very fast" in PowerPoint, rather than "normal".

Moving Objects: Evolution made sure that moving objects always catch attention. As a consequence, people tend to stop paying attention to what you're presently saying when something moves on your slide. Avoid unnecessary moving objects (like rotating logos). Where necessary (e.g., to visualize vibrations), use motion sparingly, i.e. start and stop it specifically, rather than showing it continuously.

Page numbers / progress bars: You know how when watching a kettle, it never boils? The same is true for talks: When putting in page numbers (especially in the type "page x of y") or progress bars (which are used by default in the LaTex beamer template, or so it seems), the audience gets the feeling that the talk never stops. Avoiding page numbers is strongly recommended. That being said, page numbers are required to have during rehearsals, making it easier to point to selected slides for comments.

Acronyms: Remembering acronyms requires a bit of effort. Moreover, if you use acronyms a lot during your presentation, someone who has for some reason missed the definition might get completely lost in your talk. Therefore, try to avoid acronyms that the audience needs to remember. You have plenty of space on the slide. Spell them out, even if their meaning is obvious (to you!)

Important Things Go Up: Many venues have flat rooms, and when someone tall sits in front of you, they obscure a part of the slides. It is fair to assume that many people will be unable to see the bottom third of your presentation. Hence, everything that is really important should be found at the top of the slides, even if this feels very counterintuitive.

Consistent Capitalization: Please make sure to consistently capitalize your words. While it's not per se important whether you use title case or not, being consistent whether you do or not is.

2.) Content and Structure

Like the layout of slides, the general layout and structure of a talk should be easy to grasp and follow for the audience. Note that in many aspects, this is separate from the scientific content (which is often fairly difficult, and there is little you can do about that). Here are a few simple guidelines that should help. This time, you will note that they are quite redundant.

What are you talking about? To catch the attention of the audience, it is important that they know what they are going to learn. Therefore, it is important that the talk clearly revolves a certain, well-defined scientific aspect. I find it ideal to phrase that directly as a question - that the audience can't answer before your talk, but which is clear at the end of it. Avoid generic terms in your outline like the outline on the left.

    Bad Outline
  • Motivation
  • Experments
  • Results
  • Conclusions
    
    Good Outline
  • Motivation: Heat dissipation limits the density of qubits in quantum computers
  • Problem: Dielectric losses in insulators caused by defects with dipole moments
  • Action: Oxide deposition was optimized to reduce losses
  • Conclusion: Vacuum is a lossless dielectric, redesign to use vacuum where possible

Don't save the conclusions until the end. State the main results on the contents slide. Provide arguments to support the claims on the contents slide throughout the talk. Repeat the conclusions at the end.

What does the audience want to hear? There is a fundamental difference between what the presenter (i.e. your) wants to say, and what the audience wants to hear. Typically, the presenter wants to focus on what they did, i.e. where they put a lot of effort into. (Interestingly, this never changes and is true both for young students and seasoned senior researchers). Unfortunately, almost always the parts that required the most effort are also the ones that didn't work out. Conversely, the audience has little interest in learning about what doesn't work - they are more interested in the answer to the question you posed, rather than the path that led to it. Also from a psychological perspective, it is bad to concentrate about failures, even if it worked out eventually. The audience will remember the failures more than your successes, leaving an overall worse impression. Therefore, as frustrating as that may be to the presenter, it is best focus on the final, last approach that worked. In other words, talk primarily about the key findings (not the minor aspects that stole so much of your time).

Focus! Only focus on a single story. At a conference, people hear a lot of talks. It is easier to remember talks revolving about a particular issue, rather that something that had a lot of different aspects. Keep side-stories (i.e., additional effects that are not relevant for the scientific question you are tackling (see first point)) to a minimum.

Talk about the science: Part of the audience at your seminar talk and your master defense consists of professors who will give you a grade. It should be clear to them why you needed to study physics to perform your master project. Keep the science in focus as much as possible. This tends to be more of an issue with master theses that are performed in collaboration with a company. The company's goals may not align with you spending a lot of your time going into the details of the underlying physics. However, you will be graded on how you explain the physics aspects of your project.

Don't put in too many different aspects. This has very much to do with the previous two points, but it's important to say, nonetheless. A common rookie mistake is to put in too much content into a talk. Often, this originates from the feeling that the talk may give the impression you've done too little. Be assured that everyone who sits in the audience knows that for everything you show, there are also many things you don't, and that coming to the relevant insight required a lot of work.

General structure / outline. All talks follow the same general structure: They start with an introduction, then present the results, and finally summarize what has been said. In a short (<15min) talk, you usually do not need an outline slide - putting down the scientific questions that you're going to answer typically suffices, i.e. is outline slide enough. More sophisticated outline slides are only needed if you talk about more than one story or aspect. However, if you do use an outline slide, make sure it is specific. In other words, such a slide should tell what you're actually telling, rather than using generic placeholders (introduction, results, conclusion…).

Motivate each slide. Between each slide / each new point that you're trying to make, you should explain the audience why they're listening to it. A common mistake is to say that "we also looked into that thing" (yeah, you did, but why do I care?), but rather explain why this is also important to understand the scientific question you're talking about.

Explain graphs. While you're seeing the same type of graph over and over again, making it very obvious to you, the audience will almost always be unfamiliar with it. Listening to your conclusion about the graph while still figuring out what exactly you plotted requires a lot of attention, and typically does not work very well. It thus greatly helps if you take your time to explain graphs step by step: First, explain what's plotted on the axis. Then, explain what each line in the graph stands for. Only then, tell the audience what they see scientifically (i.e., learn) in the graph.

Don't apologize, be confident. If you're not believing in your result, or if you think they aren't good enough, why should the audience trust you? Seriously, this can be major problem, even for otherwise very good science. When you present your results, stand behind them!

3.) Public speaking

You should go through the slides with your adviser before the seminar. Practice your presentation with your working group. They will give some friendly advice on what could be improved.

Experienced speakers look into the audience as they speak. This is harder than it looks. Normally when you speak to someone, they nod if they understand, they look confused if they don't understand, they scowl if they disagree, and they look away if they are uninterested. We adjust our presentations depending on the feedback we get from our listeners. If the audience is larger than a few people, you get conflicting feedback. Speak to the people who are paying attention and try to ignore the feedback from the rest.

You should look into the audience regularly during your presentation but the first 10 seconds are the most critical. A speaker that immediately turns away from the audience to look at the slides will make the impression that he or she is inexperienced. This will make it harder to convince the audience that your conclusions are correct. Look at the audience at least until you advance to the second slide.

It is almost impossible to be too loud. Humans can detect a very wide range of sound amplitudes and your really have to scream in someone's ear for it to be too loud. Speak as though the people in the back have difficulty hearing. This will be louder than your normal voice.

4.) Courtesies

No matter how senior you are, the rehearsal are attended by 5-20 people. Please do not waste their time - it really adds up! Therefore:

Open your presentation early: Please be there 5-10 minutes before your rehearsal time and make sure your presentation is prepared. Watching someone struggling with the beamer and/or the software to open a presentation gets old very fast.

Look through the presentation on the beamer once. It's quite useful to run through the presentation on a beamer at least once, to spot the worst of mistakes (bad contrast, colors, too small fonts, too crowded).

Be aware of the timing. While a rehearsal is there to improve your talk, having a rehearsal going completely wrong is very frustrating (also for the audience). This typically happens if the talk is significantly too short or too long, or if you have never throught about the wording for the transition between the slides. Please take the time and talk though the talk at least once before the rehearsal or until you feel confident.

Don't apologize, be confident. There is no need to point out weak spots of your presentation before you give it. This will only lead to a bad impression before you even started. You'll be getting (hopefully constructive) feedback anyways.

TU Graz templates for presentations and theses.