need help resizing floating table

Question

Hello all!

Any suggestions for this issue here?

Thanks!

Here is the code I am currently working with:

\section{Empirical Results}

\renewcommand{\thetable}{\arabic{table}}
\begingroup
\begin{table}[h]
\caption {\label{tab:table1} Estimates of Pay-Performance Sensitivity: Coefficients of Ordinary Least Squares Regressions of CEO Salary on Shareholder Wealth (Standard Error in Parentheses} 
\begin{tabular}{llllll}
 Independent Variable 
 & CEO Salary & CEO Total Compensation                                      & \\    
 \hline
 Intercept   &  0.0555555   & 0.055556 \\       
 Change in Shareholder Wealth (\% Change) & 0.0408163 (.05) & .025 (.05) \\      
R-square & .005 & .005 \\ 
F-statistic for $\beta$ & 0.0246913\footnotemark[1] & 0.024691\footnotemark[1] & \\
Sample Size  &  0.0199999   & 0.020000 \\                                                   
\end{tabular}
\begin{tabbing}
$\footnotemark[1]$Significant at the 1\% Level.
\end{tabbing}
\end{table}
\endgroup

When I try to implement @bernard's code, I run into some issues with the code:

\newcommand{\rvec}{\mathrm {\mathbf {r}}} 
\usepackage{graphicx}
\usepackage{subfigure}
\usepackage{amsmath}
\usepackage{xcolor}
\usepackage{color, soul}
\usepackage[symbol]{footmisc}
\usepackage[utf8]{inputenc}
\usepackage{makecell, threeparttable, booktabs}
\renewcommand{\theadfont}{\normalsize}
\usepackage{lipsum}

This^ results in an error, as well as the following code:

\documentclass[letterpaper, 10 pt, conference]{ieeeconf}  
\documentclass[twocolumn]{article}

Any suggestions? Thanks!

Edit for @samcarter, I will go with the column width table. Can I improve it from here or is this as good as it's going to get? It looks awkward in a few places:

Here's the code that was provided to me to create it:

\section{Empirical Results}
%TABLE:
\begin{table}[htbp]
    \caption{\label{tab:table1} Estimates of Pay-Performance Sensitivity: Coefficients of Ordinary Least Squares Regressions of CEO Salary on Shareholder Wealth (Standard Error in Parentheses} 
    \begin{tabularx}{\columnwidth}{@{}XXX@{}}
        \toprule
        Independent Variable & CEO Salary & CEO Total Compensation\\    
        \midrule
        Intercept   &  0.0555555   & 0.055556 \\       
        Change in Shareholder Wealth (\% Change) & 0.0408163 (.05) & .025 (.05) \\      
        R-square & .005 & .005 \\ 
        F-statistic for $\beta$ & 0.0246913\footnotemark[1] & 0.024691\footnotemark[1] \\
        Sample Size  & 0.0199999 & 0.020000\\
        \bottomrule 
    \end{tabularx}
    \begin{tabbing}
        $\footnotemark[1]$Significant at the 1\% Level.
    \end{tabbing}
\end{table}

Thanks again!

Edit 4/23:

Here is where I am now. (Getting an error at \end{tabularx}):

\documentclass[twocolumn]{article}                                                         

\usepackage{graphicx}

%for table
\newcommand{\rvec}{\mathrm {\mathbf {r}}} 
\usepackage{graphicx}
\usepackage{subfigure}
\usepackage{amsmath}
\usepackage{xcolor}
\usepackage{color, soul}
\usepackage[symbol]{footmisc}
\usepackage{booktabs}
\usepackage{caption}
\usepackage{tabularx}

%TABLE:
\begin{table}[htbp]
    \caption{\label{tab:table1} Estimates of Pay-Performance Sensitivity: Coefficients of Ordinary Least Squares Regressions of CEO Salary on Shareholder Wealth (Standard Error in Parentheses} 
    \begin{tabularx}{\columnwidth}{@{}LLL@{}}
        \toprule
        Independent Variable & CEO Salary & CEO Total Compensation\\    
        \midrule
        Intercept   &  0.0555555   & 0.055556 \\       
        Change in Shareholder Wealth (\% Change) & 0.0408163 (.05) & .025 (.05) \\      
        R-square & .005 & .005 \\ 
        F-statistic for $\beta$ & 0.0246913\footnotemark[1] & 0.024691\footnotemark[1] \\
        Sample Size  & 0.0199999 & 0.020000\\
        \bottomrule 
    \end{tabularx}
    \begin{tabbing}
        $\footnotemark[1]$Significant at the 1\% Level.
    \end{tabbing}
\end{table}

have a look at the tabularx package. If you'd post a minimal working example (MWE) helping you would be easier. — samcarter_is_at_topanswers.xyz, Apr 18 '18 at 15:54

score 3 · Accepted Answer · answered Apr 18 '18 at 16:32

Never use \resizebox with a table: it leads to inconsistent font sizes, and can make your tables unreadable. I suggest loading makecell, which allows for line breaks in standard cells, and threeparttable to manage table notes, rather than the tabbing environment. Here is posssible code, with some improvements the rules from booktabs:

\documentclass[twocolumn]{article}
\usepackage[utf8]{inputenc}
\usepackage{makecell, threeparttable, booktabs}
\renewcommand{\theadfont}{\normalsize}
\usepackage{lipsum}

\begin{document}

 \section{Empirical Results}

\lipsum[1]
\renewcommand{\thetable}{\arabic{table}}
\begin{table}[h]
\begin{threeparttable}
\caption {\label{tab:table1} Estimates of Pay-Performance Sensitivity: Coefficients of Ordinary Least Squares Regressions of CEO Salary on Shareholder Wealth (Standard Error in Parentheses}
\begin{tabular}{lll}
\toprule
\thead[l]{ Independent\\ Variable}
 & CEO Salary & \thead[l]{CEO Total\\ Compensation} \\
 \midrule
 Intercept & 0.0555555 & 0.055556 \\
 \thead[l]{Change in \\Shareholder Wealth\\ (\% Change)} &\makecell[l]{0.0408163\\ (.05)} & \makecell[l]{0.025\\ (.05)} \\
R-square & 0.005 & 0.005 \\
F-statistic for $\beta$ & 0.0246913\tnote{1} & 0.024691\tnote{1} \\
Sample Size & 0.0199999 & 0.020000 \\
\bottomrule
\end{tabular}
\smallskip\footnotesize
\begin{tablenotes}[flushleft]
\item[1]Significant at the 1\,\% Level.
\end{tablenotes}
\end{threeparttable}
\end{table}

\end{document}

Thanks for this! When I swap out this document class for the one I currently have, there is an error, and when I add the packages you posted, I also get an error. How should I implement these changes? — texmex, Apr 18 '18 at 16:45
Could you post (or link to) a small document showing the problem? — Bernard, Apr 18 '18 at 16:50
@user159324: The reason is quite simple; you loaded the ieeeconf class first, then the article class. You can declare two classes for the same document. Drop the latter. Some other remarks: for the definition of \vec, \lmathbf is enough (it is an upright font). Also, package subfigure is obsolete (not maintained for more than 12 years), use the subfigure environment from subfloat. Also, needless to load color if you load xcolor (with option [table] if you use colour in tables, as it loads colortbl and extends it). Last, if your input encoding is utf8, load soulutf8. — Bernard, Apr 18 '18 at 18:16

score 2 · Answer 2 · edited Apr 25 '18 at 06:15

2

Your table is wider than your column. To squeeze it into the column width, you could use a tabularx environment, which will automatically determine the available width for the columns and add line breaks as needed.

Furthermore I suggest to use the booktabs package for nicer tables.

\documentclass[twocolumn]{article}

\usepackage{tabularx}
\usepackage{booktabs}
\usepackage{caption}
\usepackage{array}

\newcolumntype{L}{>{\raggedright\arraybackslash}X}

\begin{document}

\section{Empirical Results}

\begin{table}[htbp]
    \caption{\label{tab:table1} Estimates of Pay-Performance Sensitivity: Coefficients of Ordinary Least Squares Regressions of CEO Salary on Shareholder Wealth (Standard Error in Parentheses} 
    \begin{tabularx}{\columnwidth}{@{}LLL@{}}
        \toprule
        Independent Variable & CEO Salary & CEO Total Compensation\\    
        \midrule
        Intercept   &  0.0555555   & 0.055556 \\       
        Change in Shareholder Wealth (\% Change) & 0.0408163 (.05) & .025 (.05) \\      
        R-square & .005 & .005 \\ 
        F-statistic for $\beta$ & 0.0246913\footnotemark[1] & 0.024691\footnotemark[1] \\
        Sample Size  & 0.0199999 & 0.020000\\
        \bottomrule 
    \end{tabularx}
    \begin{tabbing}
        $\footnotemark[1]$Significant at the 1\% Level.
    \end{tabbing}
\end{table}

\end{document}

Another possibility: With table*, the table can span both columns

\documentclass[twocolumn]{article}

\usepackage{booktabs}
\usepackage{caption}

\begin{document}

\section{Empirical Results}

\begin{table*}[htbp]
    \caption{\label{tab:table1} Estimates of Pay-Performance Sensitivity: Coefficients of Ordinary Least Squares Regressions of CEO Salary on Shareholder Wealth (Standard Error in Parentheses} 
    \centering
    \begin{tabular}{lll}
        \toprule
        Independent Variable & CEO Salary & CEO Total Compensation\\    
        \midrule
        Intercept   &  0.0555555   & 0.055556 \\       
        Change in Shareholder Wealth (\% Change) & 0.0408163 (.05) & .025 (.05) \\      
        R-square & .005 & .005 \\ 
        F-statistic for $\beta$ & 0.0246913\footnotemark[1] & 0.024691\footnotemark[1] \\
        Sample Size  & 0.0199999 & 0.020000\\
        \bottomrule 
    \end{tabular}
    \begin{tabbing}
        $\footnotemark[1]$Significant at the 1\% Level.
    \end{tabbing}
\end{table*}

\end{document}

edited Apr 25 '18 at 06:15

Mico

506,678

answered Apr 18 '18 at 16:18

samcarter_is_at_topanswers.xyz

158,329

Hello! Thanks for the response. I'm new to overleaf and latex. How do I implement those first lines of code? Do I go back to the beginning of the document and just copy and paste it in? Thanks again! – texmex Apr 18 '18 at 16:20
@user159324 The \usepackage stuff has to be somewhere before \begin{document}, the table obviously where you want it to be. – samcarter_is_at_topanswers.xyz Apr 18 '18 at 16:27
thank you for these example tables. Very helpful. I like the look of the latter one, but can this not go into the body of the paper? It won't copy and paste into the empirical results section. I also like the look of the one that is column width, but I get an error in this case: "LaTeX Error: Environment tabularx undefined." Any tips? Thanks! – texmex Apr 18 '18 at 23:02
@user159324 For using tabularx you need to load the package of the same name. As for your problems with the two column solution: please make a minimal working example (MWE) that reproduces your problem. – samcarter_is_at_topanswers.xyz Apr 18 '18 at 23:51
please see the original post for a follow-up question. thanks! – texmex Apr 19 '18 at 12:30
@user159324 Maybe left aligned would be better than justified columns – samcarter_is_at_topanswers.xyz Apr 19 '18 at 12:35
Do you mean simply left-justifying things manually? Or does this take an adjustment in the code? – texmex Apr 19 '18 at 14:06
@user159324 See my edit – samcarter_is_at_topanswers.xyz Apr 19 '18 at 14:11
after trying to implement your edit, I get an error at the line "\end{tabularx}" – texmex Apr 20 '18 at 15:37
@user159324 Then please make a MWE that demonstrates your problem. – samcarter_is_at_topanswers.xyz Apr 20 '18 at 15:38
I edited the original document. Let me know if you need any more information about the table. – texmex Apr 23 '18 at 18:00
@user159324 You did not copy all the packages and definitions from my preamble – samcarter_is_at_topanswers.xyz Apr 23 '18 at 19:02

Mico · Answer 3 · 2018-04-25T06:23:18.220

(This answer addresses only the most recent version of the OP's query, which is marked "Edit 4/23".)

Note to OP: If you have a substantially new question, you should post a new query. Otherwise, the answers that were posted previously (and which were hopefully helpful) become meaningless to future readers of your posting.

The reason your latest code doesn't compile -- other than the fact tbat it's missing \begin{document} and \end{document} statements -- is that you didn't define the L column type.

I can't help but remark that your table is confusing and hence not easy to undertand. For instance, the legend refers to one dependent variable (CEO salary), but the table clearly reports the results of two regressions, with two separate dependent variables. Worse, assuming the F-statistic numbers you posted are real, it looks like you've misinterpreted the meaning of statistical significance. If the F-statistic is really 0.0247, then there's not a 1% but actually a 99% chance that any (linear) association between the dependent variables and the regressor is due solely to chance. Put differently, it looks like there's an exceedingly low likelihood that you've uncovered a meaningful linear relationship between change in shareholder wealth and CEO compensation: Is this what you meant to report? By the way, one usually doesn't need to highlight insignificant statistical results with asterisks.

Incidentally, the F-statistic tests the significance of the entire regression, and not just the significance of an individual regressor and its coefficient \beta. By the way, how does one get fractional sample sizes of roughly 0.2? Sample sizes are generally expressed as integers...

Finally, the subfigure package has been deprecated for a decade or more. Don't use it! Use either the subfig or the subcaption package.

Here's my attempt to clean up some of the challenges pointed out above.

\documentclass[twocolumn]{article}                                                         

\usepackage{graphicx}

%for table
\newcommand{\rvec}{\mathrm {\mathbf {r}}} 
%% \usepackage{graphicx} % no need to load a package twice
%%%%%%\usepackage{subfigure} % don't load this deprecated package!
\usepackage{amsmath}
\usepackage{xcolor}
\usepackage[symbol]{footmisc}
\usepackage{booktabs,caption,tabularx}
\newcolumntype{L}{>{\raggedright\arraybackslash}X}
\begin{document}
%TABLE:
\begin{table}[htbp]
    \caption{Estimates of pay-performance sensitivity} 
    \label{tab:table1} 
    \raggedright
    OLS regressions of CEO salary and CEO total compensation on change in shareholder wealth. Standard errors in parentheses. 

    \medskip
    \begin{tabularx}{\columnwidth}{@{}LLL @{}}
    \toprule
    Independent variable & CEO salary & CEO total compensation\\    
    \midrule
    Intercept   &  0.0555555   & 0.055556 \\       
    Change in Shareholder Wealth (\%~Change) & 0.0408163 (.05) & .025 (.05) \\  
    \addlinespace    
    R-squared & .005 & .005 \\ 
    F-statistic & 0.0246913$^{*}$ & 0.024691$^{*}$ \\
    Sample Size  & 0.0199999 & 0.020000\\
    \bottomrule 
    \end{tabularx}

    \smallskip
    $^{*}$ Significant at the 1\% level.
\end{table}
\end{document}

these numbers were just filler. You are correct . . . that would have been nonsense. Thanks for the heads up though! — texmex, Apr 27 '18 at 20:59

score 1 · Answer 4 · answered Apr 25 '18 at 06:59

as supplement to excellent @mico answer (with a careful review of statistics), limited only to table design:

changes are in:

used \small font size in table
last two columns are of type l
added are vertical space (2pt) above and below cells contents (with \makegapedcells from the package makecell)
for the headers of columns is used macro \thead from makecell

\documentclass[twocolumn]{article}
\usepackage[symbol]{footmisc}
\usepackage{booktabs, makecell, tabularx}
\renewcommand\theadfont{\normalsize}
\renewcommand\theadgape{}
\setcellgapes{2pt}
\newcolumntype{L}{>{\raggedright\arraybackslash}X}
\usepackage{caption}

\begin{document}
%TABLE:
    \begin{table}[htbp]
\caption{Estimates of pay-performance sensitivity}
\label{tab:table1}
    \raggedright
OLS regressions of CEO salary and CEO total compensation on change in shareholder wealth. Standard errors in parentheses.

    \medskip
    \setlength\tabcolsep{4pt}
    \small
    \makegapedcells
\begin{tabularx}{\columnwidth}{@{}Lll @{}}
    \toprule
\thead[l]{Independent\\ variable} 
            &  \thead[l]{CEO\\salary} 
                                & \thead[l]{CEO total\\ compensation}\\
    \midrule
Intercept   &  0.0555555        & 0.055556          \\
Change in Shareholder Wealth (\%~Change)
            & 0.0408163 (.05)   & 0.025 (.05)       \\
    \addlinespace
R-squared   & 0.005             & 0.005             \\
F-statistic & 0.0246913$^{*}$   & 0.024691$^{*}$    \\
Sample Size & 0.0199999         & 0.020000          \\
    \bottomrule
\end{tabularx}

\smallskip
$^{*}$ Significant at the 1\% level.
    \end{table}
\end{document}

need help resizing floating table

4 Answers4

Linked