• Numbers less than 10 are words (In our study, seven men did not complete the treatment)
• Numbers 10 or more are numbers (There were 345 children in the study)
• In a sentence, numbers below 10 that are listed with numbers 10 and above should be expressed in numbers ((In our study, 20 girls and 4 boys were anaemic).
• Use words to express any number that begins a sentence, title or deading. Try to avoid starting a sentence with a number (Twenty four per cent of the patients were malnourished)
• Numbers that represent mathematical or statistical functions should be written as numbers (We multiplied the scores with 3).
• Use a zero before decimal points when numbers are less than 1. (The P value was 0.012)
• Do not use space between a number and the percent sign (In total, 33% of the participants were obese)
• Use one space between a number and its unit (The mean weight of women was 62.5 kg)
• The P value is with capital P. (The P value was 0.012)
• P values between 0.001 and 0.05 should be reported with three decimal points (More girls than boys were malnourshed (t=3.0, df 45, P = 0.004).
• P values reported on a computer output as 0.000 should be reported as P < 0.001.

• Data editing is when the researcher reviews the recorded data for accuracy, consistency and completeness
• During Data summarization, the researcher summarizes the data. This consists of descriptive analysis such as tables that classify observations according to key factors
• The last step is sometimes referred to as estimation stage. The researcher estimates the measures of occurrence or of effects, using confidence intervals. This final step involves interpreting the results. It also needs some consideration of factors that were not measured.
  

(Reference: Rothman KJ, Greenland S and Lash TL: 2008 Modern Epidemiology. Third Edition. Lippincott).

1. Use number codes for categorical variables
2. Choose suitable variable names, and label them to avoid confusion
3. Check for double records, and for unlikely data values
4. Make regular back-ups of the data set
   

1. Do frequency analysis of all categorical variables
2. Use box plots, histograms, and other tests to see if the continuous variables have a normal distribution
3. Identify and deal with missing values
4. Then redo the frequency and distribution checks
   

1. Title of the proposed paper
2. Author list, and identify who is responsible for data analysis and documentation
3. Specific research questions to be answered
4. Identify the outcome and explanatory variables to be used
5. Statistical methods
6. Information about the database location and file names
7. Identify journals, and if suitable, scientific meetings, where you plan to present your results.
   

• Rule 1: Read many papers, and learn from both the good and the bad work of others.
• Rule 2: The more objective you can be about your work, the better that work will ultimately become.
• Rule 3: Good editors and reviewers will be objective about your work.
• Rule 4: If you do not write well in the English language, take lessons early; it will be invaluable later.
• Rule 5: Learn to live with rejection.
• Rule 6: The ingredients of good science are obvious—novelty of research topic, comprehensive coverage of the relevant literature, good data, good analysis including strong statistical support, and a thought-provoking discussion.
• Rule 7: Start writing the paper the day you have the idea of what questions to pursue.
• Rule 8: Become a reviewer early in your career.
• Rule 9: Decide early on where to try to publish your paper.
• Rule 10: Quality is everything.

1. Recognize and declare editorial interests
2. Be aware of interests beyond the commercial: Most researchers are aware of the damaging nature of commercial competing interests. However, be also aware that your research can be influenced by the wish to promote an idea, or a research programme, rather than a commercial product.
3. Consider whether there is a ghost in the machine: Ghostwriting occurs when individuals who have made a large contribution to the research project, or to writing of the article, are not named as authors. All authors should be named, and their names, and their roles and competing interests made clear in the publication.
4. Remember the protocol
5. Consider whether the data are important even if the results are not exciting

• Cohort studies track people forward in time from exposure to outcome.
• Case-control studies work in reverse, tracing back from outcome to exposure.
• Cross-sectional studies measures both exposure and outcome at one time point.
  

• Register the trial before you start the clinical or field trial. Information on how to register a trial is found here.
• Use the CONSORT Guidelines and the Flow diagram
  

1. It should state the main findings of the study in one or two sentences
2. The discussion should consider the methods, and address possible shortcomings
3. It should compare the results with other published findings
4. It should discuss the implications of the results. Sometimes you need to answer the questions «What does this mean for practice of policy» and «Are there any needs for further studies».
   

• Report percentages to only one decimal point if the sample size is larger than 100
• Report percentages with no decimal point if the sample size is less than 100
• Do not use percentages if the sample size is less that 20 (In the sample of 14 children, 3 were malnourished)
• For ranges use "to" or comma, but not "-". This to avoid confusion with the minus sign. remember to use the same number of decimal places as the summary statistics. (The mean height was 176 cm (95% CI 172 to 180) or The mean height was 176 cm (95% CI 172, 180).