Bi 253 Lab Exercise 1. Data Analysis and Reporting 1 Introduction The purpose of this lab exercise is to show the techniques for reducing and presenting data and how to write a lab report. For most of the exercises this term you will write a short report in the style of a research article. This will involve finding at least one relevant article from the scientific literature and learning how to reference it. The first step is to collect some data, but for this introductory lab you will have some fake raw data to work with. These will be analyzed and you will make some figures that go into the lab report. The analysis and figure preparation will use Excel and the text will be written in MS Word. When you produce lab reports subsequently in the course you may use software of your choice. There are several good freeware clones of MS Office that will work adequately; in particular Libreoffice is a good choice that will run on Mac OS-X, Windows (although perhaps not Windows 8) and Linux. Some of you may have laptops or tablets in lab with you, and then you can work somewhat independently. For those who do not there is the lab station computer. If more than one person in your group is sharing the lab station computer, then you will make a single lab report which you will all use instead of an individual report. In the future, however, all parts of your lab report, with the exception of the raw data, must be produced by you alone. 2 Reducing Data One simple but important thing to keep in mind is unit conversion. Sometimes you will collect data that must be converted to appropriate S.I. units. For example, you might collect weight data from your student partners in pounds (lbs) but for scientific purposes, these values must be converted to kilograms (kg). Sometimes there are multiple standards in use. Blood pressure in the United States medical community is only reported as millimeters of mercury (mmHG) which is a terribly old standard not used elsewhere. The correct S.I. unit for pressure is Pascals (Pa). 1 To do the conversion you can (1) find a web calculator, (2) use any number of smart phone apps or (3) if you have a mac or linux computer use the “units” program that comes bundled with the operating system. Finally, if you have a number of values to convert MS Excel is most convenient, and this is what we will do later in this lab exercise. For your first simple exercise, convert the following measurements to the correct S.I. units: 1. 60 lbs to kilograms 2. 40 miles per gallon to liters per kilometer 3. 87 degrees Fahrenheit to degrees centigrade 4. 140 mmHg to Pascals The next data reduction consideration has to do with reporting numbers. Consider a behavior experiment in which you observe pillbugs escaping a hot, bright light source. Pillbugs like cool dark places, and so will move quickly away from the light. Suppose you have determined that the pillbug retreats at 2.5 cm per minute. Other students collect similar data which you pool to get an average. You might have something like the following. pb1 pb2 pb3 pb4 pb5 2.5 3.1 2.9 1.6 1.8 The average is, of course, (2.5+3.1+2.9+1.6+1.8) 5 = 2.38cms−1 If your TA sees this result in a lab report, there could be a bit of an issue. How so, looks perfectly alright? Well, to make the point with a bit of exaggeration, suppose your pillbug speed was reported as 2.38475920 cm s−1 . Did you really measure the speed that accurately? How certain are you that the speed was not 2.38475921 cm s−1 ? Here are some key terms relative to measurement: • Resolution refers to how well you can read your instrument. If you have a ruler divided into 1 mm tics, then you can resolve distance to 1 2 mm, or arguably to 0.5mm if you see that a point is half way between two tics. If you are reading a digital meter, then the resolution is the number of decimal places you get. A digital thermometer might report 26.5◦ C which means that it has 0.1◦ C, or 1/10th of a degree resolution. • Precision has to do with repeatability and essentially the quality of the instrument you are using. Suppose you use a caliper to measure the length of your pillbug; see Figure 1. If the caliper is a cheap plastic tool with lots of slop in the mechanism, then each time you make a measurement you may get a different reading. You will be able to resolve the scale similarly each time, say to 0.1 mm, but there may not be much consistency. On the other hand, if you purchase a $300.00 caliper that is a precision instrument, then each time you use it the readings will be the same (when measuring the same object). • Accuracy. Even if you have a precision instrument that has high resolution, you still don’t know if it is correct. For critical measurements many lab instruments have to be themselves measured and calibrated by a “standards laboratory”. In such a laboratory are kept, for example, objects of an exact known length. OK, so how do they know their objects are exactly right? Well, ultimately all measurements in the U.S. are related back to the National Institute for Standards and Technology (NIST). Very often you may find an instrument with a sticker on it that will say “NIST traceable” or “NIST certified” with a date. The certification only lasts for some time, then you have to recheck your instrument. If you used a ruler of unknown accuracy and low precision to measure your pillbug distance, and perhaps using a similar quality instrument to give you time, then how should you report your velocity? Maybe 2.4 cm s−1 is more appropriate than 2.38, because you really can’t know the real value to 0.01 cm s−1 . Very often the trouble begins when you scale values by multiplication or division, because your calculator will report many decimal places – it is just doing math and does not know about your science experiment! By the way, one of the worst offenders for lab instruments are temperature meters. They often will give you a readout to 1/100th of a degree, but if you look up data on the sensor used in the device, often it is only “good” to 0.5◦ C. Said another way, the resolution is inappropriate for the precision. 3 Figure 1: Measuring your pillbug 3 Data Reduction and Analysis with Excel R In this section you will work with some data as it might be collected during a real experiment. Also, your TA may have you fill in some of your own data in a table – see the very end of this exercise. In Table 1 are some values for human weight, height and and age. Some fake student data weight (lbs) height (inches) age (yrs) 250 69 25 185 75 31 175 68 43 200 71 19 Table 1: Data from some students who might be in your lab group Many, or perhaps most of you may be familiar with Excel, but it is important that everyone have a handle on how to use Excel for reducing and plotting data. So if you are an expert and someone in your group is less knowledgeable, please give them a hand. The following exercise may be pretty lame if you know your way around the program. First we will enter into the spreadsheet the data from table 1: Enter the data into the cells as shown. If an entry is not fitting entirely 4 within a cell, go to the “Format” dropdown menu and choose “AutoFit Column Width”. Your screen should look like Figure 2 Figure 2: Entering data into Excel Now that we have entered in the data, we want to do a calculation; in this case we will calculate the basal metabolic rate (BMR) for our hypothetical individuals. The formula for BMR is different for males and females, so for simplicity we will assume that the data values are for males. The BMR formula takes in S.I. units (kg, cm, years), however we have data in the form of: (lbs, inches, years), so we will have to convert the units to what we need. We could do each one by hand as in the exercise above, but the whole point of spreadsheet programs like Excel is to do multiple calculations like this for us. First, lets make a new column for the converted weight numbers. This column should go between the current weight column and the height column. To do this: • Highlight the current height column (one way to do this is to click on the top cell, hold shift and then click on the bottom cell). • Right click on one of the highlighted cells. • Choose the “insert option” 5 • From the popup menu, choose the “Shift Cells Right” option. • Click “OK”. Now your screen should look like Figure 3. Figure 3: Set up for a simple calculation In the top cell of this new column, type something like “Mass(kg)”. Now we need to convert lbs. to kg. Looking at your set of conversion factors, it looks like 1 pound = 0.454kg. Thus, we will have to multiply each persons weight in pounds by 0.454 to get their weight in kg. Figure 4 • Select the first cell under “Mass(kg)” by clicking on it. 6 • Type “=B3*0.454”. Notice that everything you type after “=” comes up in the equation box above the cells. • Hit enter. Figure 4: One converted mass value You should now have the weight of the first person in kg. in that cell, as per Figure 4. To do the rest of the column, you dont have to type in an equation for each number, since its simply the same equation with sequential cell numbers. We just need to copy the formula in the row where you entered it into other rows, i.e., do a “row copy”. Here are the steps, and the result is shown in Figure 5 • Right click the cell we just calculated. • Choose the “Copy” option. • Now select the entire column, just as before. • Right click on one of the selected cells. • Choose the “Paste” option. You should now have an entire column of weights in kg. Now, repeat these steps with the appropriate changes (naming columns, conversion factors) to convert the “Height” column from inches to cm. The next thing we will do is perform a calculation. By reviewing the scientific literature (there will be more to say about that shortly) we learn that you can determine the basal metabolic rate (BMR) from height, weight and age values. How to do this is given by the following equation: 7 Figure 5: The formula will be applied to all four rows Pmen = 13.75m 5.0h 6.76a Kcal + − + 66.47 1kg 1cm 1year day and Pwomen = 9.56m 1.85h 4.68a Kcal + − + 655.10 1kg 1cm 1year day Where m = mass (kg), h = height (cm), and a = age (years). The spreadsheet entry for the equation looks like this (Figure 6) which is entered with these steps: • Select the first cell in the BMR column by left clicking on it. • Type (without quotes) into the cell this equation: “=13.75*C3+5*E36.76*F3+66.47”. • Hit enter. • You should now have the BMR for that person in the first cell of that column. • To do the rest of the males in that column, copy the cell we just calculated, select the rest of the cells in that column (as before), right click and choose paste. 8 Figure 6: Entering the BMR equation You should now have BMR data for the entire male column of data. Its a good practice to glance over the final data output and make sure it makes sense. Typically, BMR output from this equation will be from 1000-2500 kcals a day. Use your own judgement to decide if the output looks wrong (obviously, if youre getting 6 million kcals a day as your output, something went wrong, but if its saying 700 kcals, its possible that you just have a very small person of a certain age). Now we will expand our data set by collecting data from each other (but only with your permission!). Of course we don’t have BMR values, but certainly we can get height, weight, age and gender. There are some meter sticks taped to the wall for obtaining height, and you probably know your weight. If not, perhaps your driver’s license is a good place to check... or you can just guess. Make a tables with column headings for height, for age, and for weight. There should be one table for males and one for females. Once you have your data, use Excel to determine these results: • From the female data re-run the calculations for BMR, as you did above for the fake male data. You can also do your new male data. • Calculate some simple statistics: the mean and standard deviation. The calculations should be separate for the male and female data. Your TA will help you find the mean and standard deviation with Excel. Standard deviation is a measure of how much variability you have in your data. It is a common question in biology to ask if two populations are different. For example, are voles from California larger than voles from Oregon. Suppose you go out to the field and get a lot of weights from California voles and a lot of weights of Oregon voles by spending the summer trapping the little buggers. Then you calculate the mean weight for each population of values. Say you find Calvoles have a mean weight of 22 9 grams and Orvoles are 23 grams. Are they different? Well, 22 is not 23 so they must be. Hmm, not so fast! Suppose the variation in weight is so great that half of the heavier Calvoles weigh more than half of the lighter Orvoles. You would not be very confident in saying that they are really different. A simple statistical test would tell you exactly how confident, and this test is based on the standard deviation of each set of weight values. Although unlikely, it is possible that the Calvole values come in like 21.5, 22.1, 22.0, 21.9 and so forth, while the Orvole values look like 22.9, 23.1, 23.5, 23.0. In other words, you never see Calvoles as heavy as Orvoles. Now the variation is much less, and the standard deviation measure would therefore also be less, and you might be pretty safe in saying the populations of voles do not weigh the same. Graphically, the way you illustrate the amount of variation with a bar plot is to add a bar to the bar! You put an error bar at the top of the bar showing the mean value. The length of the error bar is equal to the standard deviation. If two bars you are comparing are different in height, because the mean values are different, then by looking at the error bars on them you can get an idea if the mean difference really says something. If the error bars do not overlap each other, there is a very good chance that statistically the means are really different. By the way, if you perform a statistical test to show that two populations have different means (or some other statistical test), then you can make a statement like “population A is significantly different from population B”. Unfortunately, it is common when writing to use the word significantly to emphasize that there is a large or prominent or obvious difference, but when making a scientific report you cannot use the word significantly unless you have actually done a statistical test. It is a reserved word! 4 Data Presentation R Graphs for scientific presentation are more precise than the standard Excel spreadsheet graph. Axes markings and labels must be done correctly, and the data properly presented. Much of this is “style” and one could argue that style does not matter if the results are there, but this is not the case. If the reader has to puzzle over your data plot it is probably due to poor presentation. In this class we would like you to develop a sense of pride in presenting your results, and that means making figures with care. One important convention for x-y or scatterplots is to put the dependent 10 variable on the Y (vertical) axis and the independent variable on the X (horizontal) axis. The independent variable is the one the controls the dependent variable. For example, later you will make a plot of weight vs. age. We know that as people get older then tend to put on extra pounds, so weight depends on age, but not the other way around. Thus such a plot would have the variable age on the X axis and the variable weight on the Y axis. Here are some graph presentation techniques to keep in mind: • Axes should have tic marks, and the last tic mark should be at the end of the axis. That is, there should not be a long portion of axis extending beyond the last tic. • The labels for the tics should be large numbers, not the microscopic ones often produced by default. • Tics should be at reasonable intervals. For example, tics at 3.45 4.27 5.19 should be replaced with tics at 3 4 5 or perhaps 1 5 10, depending on the axis range. • Axes should be labeled descriptively, and indicating the units. For example, “Blood Pressure (mm Hg) ”, or better yet “Blood Pressure (Pa) ” Bar charts can have spaces between the bars that are adjusted on the basis of style and labeling, but for histograms, there are some fine points for best results: • The histogram bars should not have random spaces between them, as is common with bar graphs or bar charts. • Ideally, the each bar should be at the midpoint for the interval, with the bar width exactly the interval span. Like this: Suppose you have histogram intervals of 4 units. So bin 1 is 0 to 3.99, bin 2 is 4.00 to 7.99, bin 3 is 8.00 to 11.99 and so forth. The center of histogram bar 1 would be at 2.00 and the width of the bar such that it extends left down to zero and right to 3.99. You should not put the same data in a table and in a graph. Pick the best format and just show your data once. Tables should include headings and units for the values. 11 The following instructions take you through the process of making a scatter plot, which is data shown in two dimensions, x and y. What data will we use? Now that you have calculated some BMR data, you might be wondering what physiological factors influence BMR. According to recent research [1], the best known predictor of BMR is fat-free mass (the mass of your body minus any fat). Suppose that you wanted to see if this research is right. You go out and grab 15 random people, measure their fat-free mass (never mind how), then you measure their BMR. You now want to find out if there is any correlation between the two variables. Starting in Excel, enter your data into two columns (the first one labeled as Fat-Free Mass (or FFM) and the second one labeled BMR as shown in figure 7. Figure 7: Entering mass and BMR data. Now, highlight both sets of numbers, go to the “insert” tab, then choose “Scatter”. This is shown in figure 8 and gives the result shown in figure 9. If you wanted to change the style of any part of the graph, right click on that part of the graph and choose the “Format” option. For example, if you wanted to change the style of the data points, right click on one of them and choose the “Format Data Series” option. If you feel that the range displayed is not optimal, right click on the axis that you feel is wrong and 12 Figure 8: Select the scatter plot graph option. choose “Format Axis.” From the “scale” tab, you can manually enter in maximum and minimum values for that axis. Figure 9: The initial data plot. Now you will want to add labels to your axes, as well as a title for your graph. The easiest way to do this is to go to the “Layouts” tab under “Chart tools” and choose the first option. You should now have a graph with a title and axes labels, all you have to do is click on the individual labels and/or titles and type your own labels. 13 Scientific graphs nearly always have tic marks on the axes, and Figures 10 and 11 show you how this should go. Figure 10: First step in tweaking the axes. The next thing we will do is add a trendline to the scatterplot. This is often done to give a good visual representation of the slope of the data values. When there is a lot of variation in the data, a non-zero slope may be hard to appreciate. The trend line is calculated using a statistical technique called linear least-squares regression and often the trend line is called a regression line. How the calculation works is a bit more than we want to get into for this lab, but Excel makes it easy to get the result. See Figure 12 for how to do this with your data. Your final graph should now look something like Figure 13. Note the comment on the figure that R2 = 0.9152. What does this mean? When you calculate a regression line you would like to know the extent to which your population of points actually follow the line. That is to say, you want to know if there is a correlation between the independent and dependent variables. If you had a random cloud of points, then just about any X value could have an associated Y value over the whole range of Y values. On the other hand, if there is a strong correlation, as the X values go up (or down) the Y values follow systematically. So a large X value can only have a large Y value associated with it. The regression calculation offers up this result: The correlation coefficient, sometimes just called the “R value”. An R value can range from 0.0 (no correlation at all) to 1.0 which is a perfect correlation. Well, actually it goes to -1 too, for the case where one variable 14 Figure 11: Second axes adjustment. goes down predictably as the other goes up. Now the “R2 value” is related: Not surprisingly it is the R value squared, and gives the “goodness of fit” between the trend line and the data points. For our data the R value is about 0.96, which is very good. That is to say the data are highly correlated. When this is the case you would expect the trend line to fit well, and thus we see the correspondingly high R value. Finally, it is important to note that correlation is not the same thing as causation. Dont forget that your graphs should always have units on them, as well as a descriptive title. Figure 12: Adding a trend line to your scatter plot. 15 Figure 13: The final plot So we have made a proper scatterplot, how about a bar chart? Recall you have two populations of height, weight and age data values, one for males and one for females. You used Excel to determine the mean and standard deviation. Now try to produce a bar chart with one bar representing mean height for males and another bar representing mean height of females. The height bars should include standard error bars. You can do the same for weight. This figure can be figure 1 in your lab report due next week. Your TA will help you with this if you are not arriving at a satisfactory result. 16 5 Organizing your lab report Now you should have your data reduced and figures prepared. You know what results you have obtained. The next step is to start building your lab report. A scientific report is organized this way: • Introduction. Here you give background information. For example, with the metabolic rate lab you might talk about how much food animals consume, a range of oxygen consumption for different kinds of animals, the effect of age or exercise or whatever on metabolic rate. You could talk about different methods for measuring metabolic rate. • Methods. Now you describe how you went about collecting your data. • Results. In the results section you present your data descriptively, and with tables and figures. Each table or figure must be numbered, in the order presented. Usually the same data is not shown in both a table and a figure. • Discussion. Now you talk about what you found in the context of what others have found. You might think this is just like the introduction, but there is an important difference. In the introduction you might make a statement like “Humans have a relatively high basal metabolic rate, compared with other mammals (Poindexter 1932)”. This is introducing the reader to what is known about human metabolism. When you are writing the discussion, you would make a statement more like this: “Due to our leaky apparatus it was not possible to see a clearly defined BMR, as has been shown using the Hossenpfeffer method (Poindexter 1932). In other words, you compare your results with expectation. You don’t have to explicitly reiterate every citation from the introduction. • References. These are the sources of information that you discuss in relation to the experiment performed and the data you obtained. For the next part to the exercise you should open MS Word and make a “skeleton” for your lab report. Make headings for each of the sections of the report (Introduction, Methods, etc). So you have your headings, but what do you say? How do you introduce the experiment? It is best to know what is going on, yes? How about we 17 know find a scientific article or two that relate to the experiment. Since we got data on basal metabolic rate in humans, this is what we should hunt for. Parenthetically, you should realized that in a research lab the literature search is the first step. You would never undertake an experiment without knowing what else has already been done. The PSU library has several very comprehensive scientific databases you can use. Biosis and Medline are two. In many cases, to get access to the full text versions of articles found with databases you must do your search from a university computer. The reason for this is that PSU pays for access to the journal contents. It is very expensive! In most cases, Google Scholar is a perfectly adequate as a way to find articles, albeit with a bias to the more recent ones. Doing a Google Scholar search for “basal metabolism human” yielded the results shown in Figure 14. Figure 14: Looking for relevant articles. Note the links to the right of the article citations where you can get the actual article. Sometimes you will see “find it at PSU” which means the PSU library has a copy or electronic access to one. If you do the same search from your home or laptop computer then the links may not appear, or when you click on them you will be directed to the publisher’s web site. Sometimes the articles are free, but often the publisher will want you to pay a large fee, perhaps $35.00 for an article. There is no need to do this! Just use a PSU computer. As a last recourse you can request an article through Interlibrary 18 Loan (ILL) and there is a link on the PSU library web site for this. It can take a few days, though. So the last article on the list looks interesting. When you click on the link you get the front page of the article and lots of ancillary information. A little hunting reveals a link under the heading “Access” to “Full Text (PDF)” which is what you want. Soon you will have the exact article in your hands. The first page of the article looks like Figure 15. Well, we will have to read the article and get information related to the metabolic data you “obtained” in lab. 6 Scientific Referencing When you write your lab reports you should adhere to the standards for scientific article referencing. The first thing to keep in mind is that a reference list, which you will provide in your report, is different from a bibliography. In a bibliography you may include many relevant works even if they are not explicitly referenced in the text of your article or report. When you write a scientific article or lab report, every point that must be backed up should have a reference associated with it, and every item in your reference list must be referred to from the text. What points need to be backed up? This is a bit of an art and there are no strict rules. For statements that everyone would immediately accept as true and obvious there needs to be no reference. For example, if you had a statement in your report that “Humans maintain a basal metabolic rate” no reference would be needed. It is unlikely anyone would object to this idea. On the other hand, if you made this statement: “Based on the insulating properties of skin and subcutaneous fat, humans can tolerate temperatures as low as 22 ◦ C without seeking shelter.” then you would need to back up the claim. The reader might want to know who, in fact, says so! The in-text referencing style varies somewhat, but the two most common methods look like the following examples. Here is the sentence, above, as you might have in your report now properly referenced: “Based on the insulating properties skin and subcutaneous fat, humans can tolerate temperatures as low as 22 ◦ C without seeking shelter (Warm & Fuzzy, 1976)”. If there are three or more authors, the convention is to list just the first author. Suppose the authors are John Warm, Evelyn Fuzzy and Norman Sneezing. Now your sentence would look like this: “Based on the insulating properties skin and 19 subcutaneous fat, humans can tolerate temperatures as low as 22 ◦ C without seeking shelter (Warm et al 1976)”. An alternative is to use numerical referencing. In this case each of the citations in your reference list has a number, and you use that in the text. Now the sentence would appear like this: Based on the insulating properties skin and subcutaneous fat, humans can tolerate temperatures as low as 22 ◦ C without seeking shelter5 . The number corresponds to the entry in your reference list. If the text reference is 5, then the Warm et al citation would be number 5 in your reference list. Lets look at the first article we liked from the search, the fourth one on the search page about basal metabolism and body size (Figure 15). Scanning the front page we can glean all the bits we need for our citation: • Authors are Craig White and Roger Seymour. • The journal is PNAS (proceedings of the National Academy of Sciences). • The article was published in volume 100, issue 7 and is on pages 4046 to 4049. • The title is “Mammalian basal metabolic rate is proportional to body mass 2/3 ”. Here is another article found by scrolling down further on the list of hits: An old classic review paper. Like, really old, but just the thing to reference basic stuff about metabolism. Figure 16 shows what it looks like. As a next exercise, extract the appropriate information from the article page and enter this and the previous citation into the reference section of your lab report document If you now look at the references at the end of this document, you will see how the reference list is properly formatted. Your task for the lab report now is to find more relevant articles, and add them to your list with correct formatting. With the remaining time in lab, you should actually write up a lab report! You can read the articles you find and start to talk about them. Of course, scientific research articles are not easy to read and you are not an expert in the 20 field of human metabolism. On the other hand, a good deal of information can be gleaned from the introduction and discussion and without getting bogged down in the details of the methodology. To move on to the next step, you should export your MS Excel figures and tables and incorporate them into the MS Word document. At this point you will at least have a report with the correct sections, a figure and a table, and a few citations in your reference list. A decent start! Are you at a complete loss to find material to put in the discussion? How about doing a Google search for BMR? If you do this many hits will come up. One of the first things you will notice is that there is a similar concept called “RMR”. Hmm, maybe I could talk about the difference and why there are two similar quantities measured. Start reading one or more of the reference articles or your text book to get more bits to discuss. References [1] Mifflin MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA, Koh YO. (1990) A new predictive equation for resting energy expenditure in healthy individuals. Am J Clin Nutr 51( 2): 241-247 [2] Kleiber, M. (1947) Body size and metabolic rate. Physiol. Rev. 27:511-541 [3] White, C. R. and Seymour, R. S. (2003) Mammalian basal metabolic rate is proportional to body mass 2/3 . PNAS 100:4046-4049 If you do everything you learned in this first lab exercise, will that guarantee a perfect score on all subsequent reports? The lesson above show the minimal steps to prepare a lab report but it is the quality of your report that will give you points above a low score. Here are the factors that go into grading a lab report: • Overall you will get credit for the clarity of your presentation. If the TA has to struggle to understand what you are trying to say, or if you thoughts are not well organized then it will be difficult to get full credit. • The proper reduction of data is very important. Was it best to get mean data? Is a plot or a table or a bar graph the most appropriate to present your results? 21 • The quality of graphs and tables is very important. Axes should be properly labeled, numbers well organized so that it is easy to read. Should your graph have just symbols? Symbols and lines through the points? • You should have at least one and preferably more references from the scientific literature. You will lose points if not. • In your discussion do you point out what is different about your data compared to what you would expect? For example, if you get blood pressure readings from fellow students, do they match literature values? If not, why not? • Are your references organized and presented properly? Points will disappear if not. • Did you bring a bribe for your TA last week? (Just kidding - don’t do this!!) The teaching assistants in this course will endeavor to have a very consistent approach to grading. That is not to say that it will be perfect, but they will do their best. First, do be aware that attendance in the lab is mandatory. This is rule number one. If you don’t show up, but get the data from another student and then write a wonderful lab report, it will still not count. Furthermore, you must write up your data by yourself. If your TA determines that you have swiped a report from another student (even if you and the other student agree to it) and perhaps made a few tweaks here and there so it appears to be non-identical, the report will be given a zero and you may find yourself explaining to the appropriate academic dean why you should be allowed to continue at the university. Academic dishonesty is a big deal. 7 Collecting Your Own Data Your TA will have you collect your own data to practice doing an analysis. Below is a table for this purpose. There are meter sticks in the lab for obtaining height. Once you have filled out the tables, you should be able to use Excel to calculate the mean and standard deviation for male/female height. A bar chart of these data could be figure 1 of your report. 22 Men height (cm) age (yrs) weight (lbs) Table 2: Data from your lab group Women height (cm) age (yrs) weight (lbs) Table 3: Data from your lab group 8 Questions to answer in your lab report 1. Use the class data to run a t-test on some of our class data. An example of a null hypothesis could be There is no difference between the height of males and females in our lab. Then calculate the mean, standard deviation, and variance of whichever aspect you chose. We want to know if these two groups are significantly different. This should be a two-tailed t-test and we will assume equal variance. What is the p-value? How do you interpret this result? For example, does the pvalue support the null hypothesis? (This is made easier by reading the supplementary paper Basic Statistical Analyses.) 2. What section of your lab report would you write your p-value? What about your interpretation? 3. Within a lab report what kind of information needs to be backed up with a citation? 23 9 Grading of your lab report If you do everything you learned in this first lab exercise, will that guarantee a perfect score on all subsequent reports? The lesson above show the minimal steps to prepare a lab report but it is the quality of your report that will give you points above a low score. Here are the factors that go into grading a lab report: • Overall you will get credit for the clarity of your presentation. If the TA has to struggle to understand what you are trying to say, or if you thoughts are not well organized then it will be difficult to get full credit. • The proper reduction of data is very important. Was it best to get mean data? Is a plot or a table or a bar graph the most appropriate to present your results? • The quality of graphs and tables is very important. Axes should be properly labeled, numbers well organized so that it is easy to read. Should your graph have just symbols? Symbols and lines through the points? • You should have at least one and preferably more references from the scientific literature. You will lose points if not. • In your discussion do you point out what is different about your data compared to what you would expect? For example, if you get blood pressure readings from fellow students, do they match literature values? If not, why not? • Are your references organized and presented properly? Points will disappear if not. • Did you bring a bribe for your TA last week? (Just kidding - don’t do this!!) 24 Figure 15: A reference25 from a first web search. Figure 16: A snapshot of part of the front page of a really old classic article. 26