5/26/2017

Journal Guidelines for EHP from their official website

  • EHP will consider papers from dissertations that have been published in their entirety by a university in partial fulfillment of a degree.
  • For each manuscript, authors must submit a competing financial interests declaration form. On top of that, authors need to FURTHER indicated in the acknowledgement section.
  • For research articles, the word count is 7000, excluding references tables figures and acknowledgements and supplementary information. No way I am going to go over that.
  • Line number is mandatory!
  • Acknowledgements should be on the first page not last
  • They want a shorter title beside the long title, need to think about it.
  • Abstract needs to be equal or less than 250 words, structured with labels: background, objective, methods, results, conclusions
  • Main text has labels: introduction, methods, results, discussions, and conclusions. All subsection title needs to be fewer than 8 words.
  • References, Tables, and Figure Legends needs to be following the main text. Figures and supplementary material will be provided as separate files.
  • All references need to be author year. List alphabetically.
  • Do not use foot notes.
  • Confidence interval must be in the same column as the point estimation.

5/24/2017

Today I learned that

  • Conditional Logistic Regression model with time-varying independent variables can be implemented through Cox proportional hazard function ‘coxph’ in combination with a survival object Surv. Always have wondered about this.

5/23/2017

Today I did

  • Verified all numbers in tables, including Paper 2 Exhibit 3 Table 2, and Exhibit 5 Table 3.
  • Temperature Study Paper 2 Exhibit 4 Figure 2 has been updated to version 3. This plot on Overleaf needs to be updated again.

Latex

  • Multiple bibliography is done by the bibunits package, in combination with \begin and \end{bibunit} and \putbib[main]

5/19/2017

Today I did

  • In the secured citrix system, now the latest version of morbidity costs is saved as MorbidityCosts.RData under HDD001Temperature/Paper2.

Latex

  • Multibib package can be used to include multiple citations. Particularly will be useful for dissertation.

 

5/18/2017

Statistical Method

  • Regression Discontinuity and Interrupted Time Series analysis has long confused me. I found a presentation of University of Washington that discusses the differences. It is almost the only place that I found this information.  It seems like the biggest difference is that ITS relies on population data and RD on individual data. Overall RD is more robust than ITS.

Latex

  • Conditional equations in Latex. Answers from here.
  • \[
        f(x)= 
    \begin{cases}
        \frac{x^2-x}{x},& \text{if } x\geq 1\\
        0,              & \text{otherwise}
    \end{cases}
    \]

    However, This code generate equation without numbering. To change it, we remove \[ and \] and replace with \begin{equation} and \end{equation}. Another thing to modify is when we write functions this why, the format of \sum becomes its in-line format – the beginning and end of the range of independent variable appears to the right of the capital sigma. To change that, use \sum\limits_{i=1}^{m_y} instead of \sum_{i=1}^{m_y}.

5/17/2017

LaTeX

  • In Latex, when table and some texts are on the same page, table will show up before text. To reverse this order, use \begin{table}[htb]. No idea what it means. This site says it means “causes LaTeX to try to fit the float “here”, or at the “top” of the current page (or the next page), or at the “bottom” of the current page (or the next page).” Doesn’t really make sen the statement.
  • In Latex, for figures or tables to not show up in the list of figures or tables, use the package {caption} then change the caption of that figure/ table from \caption{} to \caption*{}
  • In Latex, tables are sometimes too tight. In this case, after \\ or \hline, add \\[-1em]. When there is a margin that needs to be filled, instead of \\[-1em], use &\\[-1em]

General Epidemiology

  • Just a reminder: in a regression model, exp(beta) is OR in logistic regression and RR in poisson regression.