STATISTICS
Pretend that you are working a real job and writing a report for your supervisor, who is very busy and usually very nice but can also be a crotchety old war-horse. The point: don’t try to entertain her; communicate the experimental procedures and results completely and efficiently, so as not to waste her time. Assume that she understands statistics reasonably well (i.e. do not teach statistics in your report – use it to relate the experiment’s results).
DO YOUR OWN WORK. Copying another person’s EWA in whole or in part is plagiarism – a violation of the University honor code.
Late policy: if you cannot turn in your EWA by lab time on the due date, you may turn it in up to one week later with a 25% penalty.
General
The report must be typed, double spaced on one side of the paper, using 11- or 12-inch font in the body of the report. Most good EWAs are about 5 pages long. Avoid use of personal pronouns like “I”, “we”, “our”, “us”, “you”, etc. When writing about what was done for the lab exercise, use past tense; when writing about what is seen in graphs and tables, use present tense. Use section headings. Break each section into paragraphs as appropriate. Use proper grammar; avoid abbreviations. Do not use slang or informal phraseology. Use a spelling checker. Use page numbers. Make it “look professional”.
Title and Abstract
The cover page should show a carefully written title centered near the top, followed by a one-paragraph abstract of the report. At the bottom of the cover page put your name and the date. The title should be one or two lines, informative yet brief; for example, the title “Session 4 EWA” is not informative to a reader unfamiliar with the exercise.
The abstract should be a stand-alone summary of the entire report presented in terse, factual tone (i.e. “for the 6:00 news”): purpose of the experiment, brief summary of what was done, and highlights of the results, all in a half page or so. Make every word count. Be specific – give values of key descriptive statistics, test results, etc., rounded to two significant digits. No figures or tables should be in the abstract.
Paper-clip the blank grading sheet to the cover page; write your name and section number at the top of the grading sheet.
Materials and Methods
This section should begin on the first page following the cover page. In the first paragraph, give an introduction including a detailed description of the experiment’s purpose – why was this experiment done? Then, in the next paragraph or two, describe in past tense the basic measurement process and data collection methods in detail: the reader should be able to repeat the experiment completely based on what you have written here (though these should not be instructions on how to do the experiment – rather an explanation of what was actually done. DO NOT copy from the lab manual – use your own words). Then, in a final paragraph or two, say what statistical analyses were performed, with the same level of detail. Mention StatCrunch and give its website. Do not give any numerical or graphical results or interpretations in this section.
Results
Borrowed from Don Edwards
This section presents experimental results in well-crafted tables and graphs without interpretation. The idea is to let the reader look at the experimental results and draw some conclusions on their own before reading yours. The only writing in this section should be a “road-map paragraph” at the very beginning or the section explaining the contents of each table / figure: Table 1 shows the collected data…etc… Figure 1 gives…etc. Then present clean, carefully crafted tables and figures: see the blue box below for advice on these. Put the raw data in the Appendix.
Discussion
This final section gives your honest interpretation of the findings of the experiment, patterns that you see in the data (e.g. shapes of histograms, unusual observations, group comparisons) and your personal, honest explanations for these patterns (without saying “I” or “me”, of course). Give values of key statistics (rounded to two significant digits) wherever appropriate to support your discussion. If you did formal statistical tests or confidence intervals as part of your analyses, interpret these results carefully in this section. Avoid jargon – use language that would be understandable to as many people as possible. Point out any shortcomings or limitations for the experiment (for example, if we used the class as subjects, this is not really a random sample of college students). Give ideas for further experimentation that might address any shortcomings or unanswered questions arising from this experiment.
Appendix
Place the raw data here, in a well-labeled table.
About Figures and Tables Whenever possible, a figure or table should stand on its own as useful information, without the reader having to search the report. Crafting figures and tables is an art: how to neatly present a complete numerical or graphical “story” requires careful thought. Some specific points: 1. Avoid cryptic abbreviations in axis labels, column/row headings, etc. Write out brief but meaningful names for variables, categories, etc. followed by units in parentheses, for example: Blood Alcohol Content (g/dl) as opposed to BAC. On the other hand, do not go overboard in labeling – don’t make the figure / table too “busy”. If you can’t avoid abbreviations in labeling, explain your abbreviations in the legend (see item 3). 2. For tables, don’t just cut and paste from StatCrunch. Keep only the rows / columns that are of interest. Round all numeric values to 3 significant digits. Right-justify numeric values in their cells. Center column headings. 3. If it has only numeric information, it is a table; otherwise it’s a figure. Under each figure or table type a legend beginning with (e.g.) Figure 1: or Table 2: With a legend there is no need for a title above the figure / table. Use the legend to “explain” the information in the figure or table completely but briefly: make every word count. Legends should usually be 1-3 lines long. Legends often contain sample size information, source of data information, and explanations for abbreviations that had to used in the figure or table. 4. If using colors, use only two or three and make it look nice. 5. When you think you’re finished, pretend you are a reader who knows nothing about the data: looking at the figure or table, will any annoying questions come to mind? (e.g. “How many subjects???”). Every table or graph should include units of the variables and sample size information somewhere in the labeling – even if it seems repetitive
Pretend that you are working a real job and writing a report for your supervisor, who is very busy and usually very nice but can also be a crotchety old war-horse. The point: don’t try to entertain her; communicate the experimental procedures and results completely and efficiently, so as not to waste her time. Assume that she understands statistics reasonably well (i.e. do not teach statistics in your report – use it to relate the experiment’s results).
DO YOUR OWN WORK. Copying another person’s EWA in whole or in part is plagiarism – a violation of the University honor code.
Late policy: if you cannot turn in your EWA by lab time on the due date, you may turn it in up to one week later with a 25% penalty.
General
The report must be typed, double spaced on one side of the paper, using 11- or 12-inch font in the body of the report. Most good EWAs are about 5 pages long. Avoid use of personal pronouns like “I”, “we”, “our”, “us”, “you”, etc. When writing about what was done for the lab exercise, use past tense; when writing about what is seen in graphs and tables, use present tense. Use section headings. Break each section into paragraphs as appropriate. Use proper grammar; avoid abbreviations. Do not use slang or informal phraseology. Use a spelling checker. Use page numbers. Make it “look professional”.
Title and Abstract
The cover page should show a carefully written title centered near the top, followed by a one-paragraph abstract of the report. At the bottom of the cover page put your name and the date. The title should be one or two lines, informative yet brief; for example, the title “Session 4 EWA” is not informative to a reader unfamiliar with the exercise.
The abstract should be a stand-alone summary of the entire report presented in terse, factual tone (i.e. “for the 6:00 news”): purpose of the experiment, brief summary of what was done, and highlights of the results, all in a half page or so. Make every word count. Be specific – give values of key descriptive statistics, test results, etc., rounded to two significant digits. No figures or tables should be in the abstract.
Paper-clip the blank grading sheet to the cover page; write your name and section number at the top of the grading sheet.
Materials and Methods
This section should begin on the first page following the cover page. In the first paragraph, give an introduction including a detailed description of the experiment’s purpose – why was this experiment done? Then, in the next paragraph or two, describe in past tense the basic measurement process and data collection methods in detail: the reader should be able to repeat the experiment completely based on what you have written here (though these should not be instructions on how to do the experiment – rather an explanation of what was actually done. DO NOT copy from the lab manual – use your own words). Then, in a final paragraph or two, say what statistical analyses were performed, with the same level of detail. Mention StatCrunch and give its website. Do not give any numerical or graphical results or interpretations in this section.
Results
Borrowed from Don Edwards
This section presents experimental results in well-crafted tables and graphs without interpretation. The idea is to let the reader look at the experimental results and draw some conclusions on their own before reading yours. The only writing in this section should be a “road-map paragraph” at the very beginning or the section explaining the contents of each table / figure: Table 1 shows the collected data…etc… Figure 1 gives…etc. Then present clean, carefully crafted tables and figures: see the blue box below for advice on these. Put the raw data in the Appendix.
Discussion
This final section gives your honest interpretation of the findings of the experiment, patterns that you see in the data (e.g. shapes of histograms, unusual observations, group comparisons) and your personal, honest explanations for these patterns (without saying “I” or “me”, of course). Give values of key statistics (rounded to two significant digits) wherever appropriate to support your discussion. If you did formal statistical tests or confidence intervals as part of your analyses, interpret these results carefully in this section. Avoid jargon – use language that would be understandable to as many people as possible. Point out any shortcomings or limitations for the experiment (for example, if we used the class as subjects, this is not really a random sample of college students). Give ideas for further experimentation that might address any shortcomings or unanswered questions arising from this experiment.
Appendix
Place the raw data here, in a well-labeled table.
About Figures and Tables Whenever possible, a figure or table should stand on its own as useful information, without the reader having to search the report. Crafting figures and tables is an art: how to neatly present a complete numerical or graphical “story” requires careful thought. Some specific points: 1. Avoid cryptic abbreviations in axis labels, column/row headings, etc. Write out brief but meaningful names for variables, categories, etc. followed by units in parentheses, for example: Blood Alcohol Content (g/dl) as opposed to BAC. On the other hand, do not go overboard in labeling – don’t make the figure / table too “busy”. If you can’t avoid abbreviations in labeling, explain your abbreviations in the legend (see item 3). 2. For tables, don’t just cut and paste from StatCrunch. Keep only the rows / columns that are of interest. Round all numeric values to 3 significant digits. Right-justify numeric values in their cells. Center column headings. 3. If it has only numeric information, it is a table; otherwise it’s a figure. Under each figure or table type a legend beginning with (e.g.) Figure 1: or Table 2: With a legend there is no need for a title above the figure / table. Use the legend to “explain” the information in the figure or table completely but briefly: make every word count. Legends should usually be 1-3 lines long. Legends often contain sample size information, source of data information, and explanations for abbreviations that had to used in the figure or table. 4. If using colors, use only two or three and make it look nice. 5. When you think you’re finished, pretend you are a reader who knows nothing about the data: looking at the figure or table, will any annoying questions come to mind? (e.g. “How many subjects???”). Every table or graph should include units of the variables and sample size information somewhere in the labeling – even if it seems repetitive
Comments
Post a Comment