Tables and Figures Module XXIII-B Slide 1 No Comments Slide 2 After collecting and analyzing the data from our research study, we are faced with the need to turn data into information, that is, we want to interpret our findings back into the context of the original content area, in such a way that it answers the purpose or question of the research project. Slide 3 We might summarize what we need to do by saying: Show data, Show data clearly, and show data accurately. One of the ways of doing this is to make appropriate tables and graphs. Slide 4 We will first consider tables. The purpose of a table is to summarize a large amount of information in a small area. This will facilitate comparisons over periods of time or between groups. It will also mean that the reader can easily view the description of our subjects or the results of our study and we only need to highlight the most important pieces of information in the text. We can make tables for either quantitative studies using numerical information or for qualitative studies using text. Slide 5 A research paper usually includes three types of tables. First, we need summary tables for demographic information about our subject. Second, we need descriptive tables for our study or outcome variables, and Finally, we need tables which include the results of our inferential statistical analysis. Many times the results of our statistical analysis will be included in the descriptive tables for our outcome variables. Slide 6 Each table should have a title which clearly says what the table contains. There should be information about the population, variable, sample size, and time and place of the study, if appropriate. Table titles appear above the table. Many tables have very poor titles, which give minimal or no information that relates the table to the research study. A good table has appropriate row and column headings which give the framework for the table. The body of the table should have appropriate spacing (not too little or too much between pieces of information). The number of division lines in a table should be kept to a minimum. Usually, you can do a better job of directing the reader's eye using spacing than using lines. Finally, if your data contains decimals, line up the columns on the decimal point. Slide 7 Tables should be numbered in the order of appearance in the text, and if you put a table into the text, it should be referenced somewhere in the paper. This reference could be as part of the sentence or could appear in parentheses within or at the end of the appropriate sentence. Slide 8 Different levels of measurement require different types of tables. Qualitative data (i.e. nominal or ordinal data) is usually summarized using frequency tables, while Quantitative data is usually summarized using means and standard deviations. If the distribution of the data is skewed, quantitative data may be summarized using medians and ranges or minimum and maximum. Slide 9 This table is a demographic table for a qualitative study. When the sample size is small, information is usually given for each of the participants. In this table the column and row headings are appropriate, the spacing is good, and the table is very easy to read. The table, however, has a very poor title. It tells you nothing about the study. Slide 10 This is a results table for a qualitative study. It summarizes and organizes the themes from two phases of the study. The title is better than the previous title, but there should be a line denoting the bottom of the table, and possibly a line after the first row of bold information. Slide 11 This table is for a quantitative study. It has a title that tells you that it is a table of means and standard deviations, says a little about the study/study variables and informs you that there were only 4 subjects in the sample, which is a very small sample size. This table combines both descriptive information about the study variable at the two points in time they were measured, but for the readers convenience gives the means and standard deviations of the differences. This is a combined descriptive and inferential table which includes the p-values for the differences and includes a footnote telling the reader that the analysis was done using a paired t-test. Lines and spacing have been used appropriately in this table. Slide 12 This table for quantitative data is trying to put too much data into one table. There are three different comparisons between groups: Control vs Study; Dominant vs Non-Dominant and With and Without Tape. These comparisons are made for four different muscle groups. The table presents both descriptive and inferential data for each of these 12 comparisons in the same table. While the table is readable the way it is, it might be better to make separate tables for the three comparison groups, since in the text these three comparisons will be discussed in separate paragraphs or sections of the paper. Slide 13 This table for quantitative results is set up appropriately in terms of title, spacing, and amount of information. It even gives the sample size, however, there are a number of abbreviations in the table, which the reader may or may not understand, and there is no indication of the statistical test that was used to determine the p-value for the inferential statistics. Slide 14 Output from statistical programs frequently give information with many more decimal digits than should be reported in research papers. Means and standard deviations should be reported to one more digit than the original data, that is the data that was entered into the computer. If the original data is whole numbers, means and standard deviations should be reported to one decimal digit, if the data is to one decimal digit, means and standard deviations should be reported to two decimal digits. Percents should be reported to one decimal place. Slide 15 Test statistics values, the Zs, ts, Fs and chi-squares, should be reported to two decimal digits. Correlation coefficients should also be reported to two decimal digits. Slide 16 The number of digits reported for p-values depends on whether the results are statistically significant or not. A significant p-value of .028 should be rounded to .03, a p-value of .002 is correct as written, and a p-value of .000, which means 0 to three decimal places is usually reported as p<.001. Non-significant results are reported using a p-value with 2 decimal digits. Since most data analysis is done using computer, p-values should be reported as actual values, not just p<.05 or p>.05 in indicating significance. Slide 17 This table has many good qualities, but the authors did not use the recommended number of decimal place for their correlation coefficients. The third decimal place that they have included does not add any useful information for the reader. Many journals discourage the use of leading zeros for data which the reader knows are decimals. In a table with many correlations reported, leaving off the leading 0 and giving the correlation to two decimal places will make the table much easier to read. Slide 18 We now turn our attention to figures. A figure is any chart, graph, photograph, drawing or depiction. Often a figure will have a short heading at the top of the figure, but titles for figures are called legends and appear at the bottom of the figure. Titles for figures should include enough information so that the reader can understand and interpret what is presented. This includes an explanation of any abbreviations or symbols that the reader may not know. As with tables, every element of a figure does not need to be discussed in the text. The researchers need to point out the information from the figure that is most important to understanding and interpreting the results. Simple, easy to understand figures help the reader, but many figures contain too much information or the quality of the reproduction of the figure is such that it is difficult to interpret. Slide 19 This slide gives a listing of statistical graphs and other types of figures that you might see in a research article. Slide 20 When deciding what type of figure to produce for displaying our data there are several things that need to be considered. First, what is the purpose of this figure with respect to sharing information with the reader. Is it appropriate for the type of data that is being displayed? And is it appropriate for the intended reader. Choice of size, spacing and color are important. Most of the time the figure that is submitted to a journal will be full page size, but this is not the size of the figure that will appear in the article. A good practice is to Xerox your figure with reduced sized several times to determine whether the detail that you need to show will be visible at the reduced size. It is best to use fonts that are similar in size, so that when the largest front is reduced and readable, other fonts in the figure will also be readable. Slide 21 Graphing programs usually will produce a graph with elements of the graph in color. You are able to change color combinations so that they fit your taste and needs. Graphs will appear in color in a Power Point presentation or on a poster, but graphs in a research article are most frequently in black and white. Thus it is important to choose colors that will give shades of gray and black that can be distinguished from each other, or some type of fill should be chosen (not too gaudy) to be able to distinguish the elements of the graph. Spacing and amount of information contained in a figure are also very important when you use a figure to share complex information with the reader. Slide 22 Some final pointers for displaying data are: One, avoid pictograms as a method of showing frequency. Pictograms are two dimensional pictures and the results will be viewed as areas, not height or length, giving a false impression of the relative importance of the categories. Slide 23 Two, be sure to use equal intervals when making bar graphs for qualitative data or histograms for quantitative data. Equal intervals will give comparable areas for the bars, while unequal intervals will overemphasize the categories with the widest interval length. Slide 24 Three, beware of chopping off the y or vertical axis. Removing a piece of the axis, especially for bar graphs and histograms, will emphasize differences between the heights of the bars, which may not be important if you were to see the entire bar. Slide 25 Four, too much data on a graph makes it very difficult for the reader to make any judgments or comparisons about the data. There is a point where more information is actually less understandable information. Slide 26 Finally, five, two dimensional graphs are usually better than three dimensional graphs for making comparisons. Slide 27 We will now look at examples of some of the types of graphs that can be used to display data. Our first example is a box and whiskers plot. This is an excellent type of graph to use when you want to compare distributions for quantitative data for two or more groups of subjects. In fact, you can make a comparison of two or more groups of subjects at more than one time point or under more than one condition on the same box and whiskers plot. The square in the middle of the box is the mean, the line is the median, and the box covers 50% of the data. The whiskers go out to the maximum and minimum values if there are not outliers. If there are outliers, the whiskers go out to the smallest data value within what is called the fence and the outliers are indicated on the graph. A box and whiskers plots allows the reader to make comparisons for both location and variability and easily shows if any of the groups has outliers. It would be helpful if somewhere on the graph we were told the sample sizes for the groups of subjects. Slide 28 This graph shows a comparison of a quantitative variable for experimental and control groups for three different time points. The height of the bars is the mean value, but there is no indication of whether the bars are standard deviation or standard error. Also there is no indication of sample sizes for the groups. Otherwise, the title and the labeling of the graph are good. Slide 29 Line graphs are often used when displaying results from analysis of factorial designs. A factorial design has a quantitative dependent variable and two independent variables called factors. In the graph, one of the factors is place on the X-axis and there are separate lines for the levels of the second factor. The mean is usually graphed for each of the factor pairs. Sometimes error bars are added to these mean value points. A different symbol is associated with each line, so the graph is easy to understand and interpret. Slide 30 Data were collected in the clinical lab for sodium levels for a group of male and female surgery patients. The frequency data can be displayed in two different ways. In the first graph the variable on the x-axis is sodium level recorded in intervals of values. There are different colored bars for males and females. This graph, gender distribution by sodium level, makes it easy to compare males and females within any given sodium level interval. Slide 31 The second graph has gender on the x-axis with bars of separate colors indicating different sodium levels. This graph, sodium level by gender, makes it easy to compare sodium levels with each gender grouping. Which graph is appropriate for the research paper depends on the questions that are being asked by the researchers. Slide 32 The data for sodium levels by gender could also be graphed using frequency polygons. When making frequency polygons be sure that the polygons are tied down to the X-axis at the next higher and next lower midpoints on the axis. Slide 33 Relationship data for bivariate data is usually displayed in a scatter plot, where each subject's data accounts for one point in the graph. If there are clearly dependent and independent variables for the relationship, the independent variable should be plotted on the X-axis and the dependent variable on the Y-axis. Module XXIIIB Graphs and Tables Grenith J. Zimmerman, Ph.D. Associate Dean for Research Copyright 2012 1 2 Show data Show data clearly Show data accurately 3 1 Module XXIIIB Graphs and Tables Summarize a large amount of information in small area Facilitate comparisons Supplement text Quantitative or qualitative Exact numbers Text-demographic information, etc. 4 1. Summary of tables for demographic information of participants. 2. Descriptive tables of study variables. 3. Results of inferential statistical analysis. 5 Title: Brief but clear Population Variables Sample size Time and/or place Ti d/ l Units of measurement Row and Column headings Body of the Table Spacing Number of dividing N b f di idi lines Lining up decimal points 6 2 Module XXIIIB Graphs and Tables Number tables in the order they appear in the text If you put a table in to the text, you need to reference it in the paper They had similar characteristics as shown in Table 1. ..... as indicated in the stress scale self reports (See Table 1). 7 Qualitative Data: Nominal or Ordinal Frequency Tables Quantitative Data: Interval or Ratio Means and Standard Deviation Quantitative Data: Skewed Median, Range, (Min, Max) 8 9 3 Module XXIIIB Graphs and Tables 10 11 12 4 Module XXIIIB Graphs and Tables 13 Means and Standard Deviations (One more digit than the original data) Original data 25, 31, 44, 27, 48, 21 Output: mean = 32.6754; SD = 10 8901 10.8901 Table: mean = 32.7; SD = 10.9 Percents 1 decimal place 0.5945945 59.5% 14 F, t, Z, Chi-square - use 2 decimal places Output: Z = 1.8596 Table: Z = 1.86 Correlation Coefficients - Use 2 decimal digits Output: r = .23945 Table: r = .24 15 5 Module XXIIIB Graphs and Tables Significant Results Output: p = 0.028; p = 0.002; p = 0.000 Table: p = .03; p = .002; p < .001 Non-significant Results 2 decimal digits Output: p = 0.2349 Table: p = .23 16 How many decimals should be used? 17 Any chart, graph, photograph, drawing or depiction Augment, not duplicate text Easy to read and understand 18 6 Module XXIIIB Graphs and Tables Scatter plots Line Graphs Bar Graphs Pie Charts Histograms Hi t Frequency Polygons Box Plots Drawings Photographs Decision Trees Computational Charts d Nomograms and N Organizational Charts 19 Does it achieve its specific purpose? Is it appropriate for the type of data? Is it appropriate for the intended reader? Is the choice of size, spacing and color size appropriate? Will it give a good quality of reproduction? 20 21 7 Module XXIIIB Graphs and Tables Avoid pictograms, because they give results in terms of area, not height. 22 Be sure that bar graphs and histograms have equal intervals. 23 Beware of chopping off the vertical axis. Should start at 0 24 8 Module XXIIIB Graphs and Tables Not too much data in one graph. 25 Two dimensional graphs display data better than three dimensional graphs. 26 100 90 80 70 Days 60 50 40 30 20 10 0 Educational Comparison Figure 1. Comparison of days from the initial injury to return to work on regular duty by group. 27 9 Module XXIIIB Graphs and Tables 8 Experimental Group Control Group 7 6 Degrees 5 4 3 2 1 0 Pre-treatment Mid-treatment Post-treatment Measurement Time Figure 1. Comparison of degrees of shoulder ROM in the transverse plane between the experimental and control groups over time. 28 15 watts 30 watts 50 watts 70 watts 0.24 0.22 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 30 rpm 50 rpm 70 rpm Speed Figure 1. Effects of work and speed on mechanical efficiency. 29 Males Females 40 35 30 Percent Mechanical Effic ciency (%) 0.20 25 20 15 10 5 0 134-135 136-137 138-139 140-141 142-143 144-145 146-147 mm/L Gender Distribution by Sodium Level for Surgery Patients 30 10 Module XXIIIB Graphs and Tables 31 Males Females 45 40 35 Percent 30 25 20 15 10 5 0 130 135 140 145 150 Sodium Level (mm/L) Sodium Levels by Gender 32 33 11 Graphing Module V-E Slide 1 No Comments Slide 2 When making presentations, whether oral or poster, it is usually much easier to convey information by using a graph than by using a table. The type of graph that is used to display research data will depend on the level of measurement of the data. We begin by looking at graphs for quantitative data. Nominal (data in unordered categories) and Ordinal (data in ordered categories) can be graphed using the same types of graphs. Remember that the order of the categories for ordinal data will determine order for the graphs, just as it did for making frequency tables. Slide 3 For data tallied into categories, one type of graph that is appropriate is the BAR GRAPH. A bar graph has the names of the categories on the X-axis and frequency or percent frequency on the Y-axis. The bars can be any width the researcher chooses, but all bars should be the same width. Slide 4 Graphs are called Figures, when they are put into research papers. Figures should have titles containing the same type of information as titles for Frequency tables. There is one major difference. Tables have their titles above the table, while figures have their titles below the figure. The figure given in this slide is a Bar Graph for the Allied Health student's program data. You will notice that the title is the same as the title we used for the frequency table. In this bar graph, the programs are in alphabetical order on the XAxis, going from Dietetics to Speech. On the Y-Axis, we have percent frequency. Slide 5 If the categories are ordered in terms of frequency from the largest to the smallest frequency, the Bar Graph is called a Pareto Chart. The Pareto Chart in this slide contains exactly the same information as the previous Bar Graph, except for the difference in the ordering of the bars. Slide 6 When our frequency data is divided into subgroups, such as by gender or by age group, we can graph the data for the subgroups in the same Bar Graph. In this slide, we have graphed the frequency data for the programs by gender. The dark gray bars represent males, the light gray bars, represent females. Notice that gender has been added to the title of the Figure. In this bar graph it is easiest to compare the genders within each program. Slide 7 If we wanted to compare the program distribution within each gender subgroup, it would be easier to use this bar graph. Both Bar Graphs are based on the same data, and which one you chose to use will depend on which of the two comparisons is most important to you. Slide 8 Another way of displaying nominal data is with a Pie Chart. To make a pie chart, you use the fact that a circle has 360 degrees. Change your percent frequency into relative frequency by dividing by 100, than, multiply 360 degrees by this relative frequency to determine how much of the pie corresponds to this category. For example the percent frequency for dietetics is 16.1%, so dietetics would make up 360 x .161 = 58 degrees of the circle. If you are drawing a pie chart by hand, you will need your protractor to be able to determine this sector of the circle. This side gives two examples of pie charts for the Allied Health student program data. Slide 9 We next consider graphs for Quantitative data. We will illustrate these graphs using the weight data for the 53 research subjects. In the previous module, we talked about class limits, which are based on the rounded data and give intervals that do not overlap, and class boundaries which cover the entire interval of possible values and overlap at the end points. Class limits are used for making frequency tables for quantitative data. Class boundaries are used for making graphs. The graph with class boundaries on the X-Axis and frequency or percent frequency on the Y-Axis is called a Histogram. Histograms are made using bars, but they are very different than Bar Graphs. First, the histogram is graphed on the number line, so this determines the order of the bars. Second, the class boundaries for the histogram cover the entire interval of data possibilities. If there is an interval with 0 frequency, there will be no bar above that interval. There will be no gaps between bars, unless there are intervals with 0 frequency. Again, the title goes below the figure and it is very similar to the title for the frequency table of data, which lead to the graph. Slide 10 Notice the graph on the left is graphed using frequency and the graph on the right uses percent frequency, but the two graphs are identical in shape. Slide 11 Because the bars in a bar graph can be moved, we can put data from two or more subgroups on the same bar graph, however, since histograms cover the entire X-Axis interval, and the bars cannot be moved or separated, we cannot put data from subgroups in the same histogram. If we wanted to compare males and females, we would need to make two histograms, one for males and a separate histogram for females. If we want to put quantitative data for subgroups on the same graph, we will need to make a different type of graph. Slide12 One type of graph that allows us to put data from multiple subgroups on the same graph is a Frequency Polygon. Frequency polygons are graphed on Class Midpoints. Class midpoints are found by adding the lower class boundary to the upper class boundary and dividing by 2. Frequency polygons are tied down to the X-Axis at what would be the next higher and next lower midpoints. Slide 13 This slide shows a frequency polygon for the distribution of weights, that contains the same information as the histograms. Our class boundaries for the first weight interval are 99.5 and 119.5, the midpoint of that interval would be (99.5+119.5)/2 = 219/2 = 109.5. The length of the interval can be found by subtracting the lower class boundary from the upper class boundary. The length of this interval is 119.5 - 99.5 = 20. We can find all of the other class midpoints by adding the length of the interval repeatedly to the midpoint of the first interval. So, our midpoints will be 109.5, 129.5, 149.5 and so on. Our frequency polygon will be tied down at 89.5 = 109.5 - 20, and 289.5 which is 269.5 + 20. Slide 14 This slide illustrates how frequency polygons for males and females can be graphed in the same frequency polygon. Slide 15 On the right, we have a box and whiskers plot for the weights of the 53 research subjects. Slide 16 The Box and Whiskers Plot is another type of graph that can be used for graphing quantitative data. A box and whiskers plot is based on the median and the interquartile range IQR. The interquartile range is the difference between the third and first quartile, the data values such that 75% and 25% of the data lies below them. Slide 17 This all seems very complicated, so let's discuss the steps that we have taken to construct the box and whiskers plot for the weight data. We will need to calculate the following values from our data set. The median (which is 165), the third quartile (75th percentile, which is 185), the first quartile (25th percentile, which is 127.5). We also need to note the maximum value which is 260 and the minimum value which is 100. Plot the median, and first and third quartiles on the graph and make the box using the quartiles, with a line through the box at the median. This is the box part of the box and whiskers plot. The height of the box is the interquartile range IQR = Q3 - Q1 = 185-127.5 = 57.5. Since 50% of the data is within this range of values, the height of the box gives us a measure of variability Slide 18 Next for the whiskers. Calculate the Interquartile range Q3 - Q1, which for the weight data is 57.5. Using this we can determine the fences for the data. The value we need to add to the top of the box and subtract from the bottom of the box is 1.5 IQR, which for the weight data set is 1.5 x 57.5 = 86.25. Slide 19 The whiskers go out to the minimum and maximum values of the data set unless these values are outside of the ends of the box by more than 1.5 IQR, in which care they will be outliers. If there are outliers, the whiskers go out to the last data values within the fence. Note that our maximum value of 260 is within the upper fence and our minimum value of 100 is within the lower fence. This means that there are no outliers, and the whiskers will go out from the box to the values 260 and 100. Slide 20 The interquartile range is the difference between the third and first quartile, the data values such that 75% and 25% of the data lies below them. These are the values for the top and bottom of the box. . The line in the middle of the box is the median. The whiskers go out to the minimum and maximum values of the data set unless these values are outside of the ends of the box by more than 1.5 IQR, in which care they will be outliers. Slide 21 This slide illustrates how helpful box and whiskers plots are for making comparisons between subgroups. Notice that the median value for the comparison group is larger than the median value for the Educational group. Also, notice that the box which contains 50% of the data is larger for the Comparison group, indicating greater variability. Since the median value is not in the middle of the box, the distributions are not symmetric, in fact the tail is to larger values. This means that the distributions are skewed to the right. There are no outliers. The dot in the box is the mean. Slide 22 In statistics, when we are interested in looking at relationship between two quantitative variables measured on the same group of subjects, the graph that summarizes our data is called a Scatter Plot. In a scatter plot, one of the variables is placed on the X-axis and the other on the Y-axis. If we want to use our data for prediction, the variable we want to predict is put on the Y-axis and is called the Dependent variable. The variable on the X-axis is the Independent variable. If there is a linear relationship, the scatter of points should line within a long skinny ellipse. A random scatter of points indicates not linear relationship. Graphing Module VE Graphing Module VE Grenith J. Zimmerman, Ph.D. Associate Dean for Research Copyright 2012 School of Allied Health Professions 1 Conveying Information Using Graphs The type of graph depends on the level of measurement of the data. oNominal (unordered categories) oOrdinal (ordered categories) oQuantitative 2 Bar Graphs 3 1 Graphing Module VE Distribution of Program for 31 Allied Health Students % Program (n) Dietetics 16.1 (5) EMC 12.9 (4) OT 16.1 (5) PT 22.6 (7) Rad Tech 19.4 (6) Speech 12.9 (4) 4 Pareto Chart 5 Bar Graph by Subgroups 6 2 Graphing Module VE 7 Pie Charts 8 Histograms 9 3 Graphing Module VE 10 Bar Graph vs. Histogram Bar Graph oQualitative data oData for more than one subgroup on the same graph Histogram oQuantitative data oOnly one group per graph 11 Class Midpoints Class midpoints can be found by adding the lower class boundary to the upper class boundary and dividing the sum by two. Midpoints differ by the length of the interval. Frequency polygons are graphed on class midpoints. They are tied down to the xaxis at what would be the next higher and next lower midpoints. 12 4 Graphing Module VE Frequency Polygon oBoundaries 1st interval 99.5, 119.5 oMidpoint (99.5+119.5) =109.5 2 oLength of interval 119.5-99.5 =20 13 14 Weight data 15 5 Graphing Module VE Box and Whisker Plot oBased on the median and the interquartile range oIQR =Q3-Q1 oQ3=3rd quartile oQ1=1st quartile 16 The Box Median = 165 Q3 = 185 Q1 =127.5 Box Minimum: 100 Maximum: 260 17 The Fences IQR = Q3-Q1 =185-127.5 =57.5 The Fences (end of the box +/1.5 IQR) Upper fence 185 +1.5(IQR) =185+86.25 =271.25 Lower fence 127.5-1.5(IQR) =127.5-86.25 =41.25 18 6 Graphing Module VE 19 Summary o The box height is the interquartile range IQR = (Q3 - Q1) o The line in the box is the Median o The Whiskers go out to the minimum and maximum values of the data set unless these values are outside of the end of the box +/- 1.5 IQR, in which case they will be labeled as outliers 20 Box and Whiskers Plot Used for Group Comparison 100 90 80 70 Days 60 50 40 30 20 10 0 Educational Comparison Figure 1. Comparison of days from the initial injury to return to work on regular duty by group. Figure 11. 21 7 Graphing Module VE Scatter Plot Looking for Relationship 22 8 Here are more notes to review for assignment. They opens up a video. However you probably already know this basic info. 1) SPSS Graphs, Bars and Histograms 2) SPSS Graphs - Box Plots 3) SPSS Graphs - Scatter Plots SPSS Graphs, Bars and Histograms PLEASE DOWNLOAD THIS VIDEO. All you need to do is click on the link (not the video box), a prompting will pop up and ask you if you want to save or open. You should open it and watch the video. You will probably need to view it in full screen in order to see the details of the SPSS program. SPSS Graphs Bars and Histograms Click to view SPSS Graphs - Box Plots PLEASE DOWNLOAD THIS VIDEO. All you need to do is click on the link (not the video box), a prompting will pop up and ask you if you want to save or open. You should open it and watch the video. You will probably need to view it in full screen in order to see the details of the SPSS program. SPSS Graphs Box Plot Click to view SPSS Graphs - Scatter Plots PLEASE DOWNLOAD THIS VIDEO. All you need to do is click on the link (not the video box), a prompting will pop up and ask you if you want to save or open. You should open it and watch the video. You will probably need to view it in full screen in order to see the details of the SPSS program. SPSS Graphs Scatter Plot Click to view