Manually Create Graphs (using coef and standard error)

[Cat Santos]

Dear All,

I hope you are doing very well.

I need some help as I would like to manually create a coefplot out of my data, considering that my data is constituted by the variables: Time, Coefficient, Std. err., z, P>z (pvalue), 95% conf. interval -- low, and 95% conf. interval -- high.

My Y axis should be the coefficient, and my X axis the time variable. I also would like to add mlabels to each coefficient point with significance stars (* P ≤ 0.1; ** P ≤ 0.05; *** P ≤ 0.01).

Is this possible to do?


Here is what my dataset looks like

time  coefficient  std_err    z_stat  pvalue   ci95_low    .ci95_high

1      -.0245       .0208597    -1.17   0.240    -.0653842    .0163842
2        .007       .0007071     9.90   0.000     .0056141    .0083859
3        .014       .0106125     1.32   0.187    -.0068001    .0348001
4      .00575      .039004     0.15   0.883    -.0706964    .0821964
5      -.0145     .0183504    -0.79   0.429    -.0504661    .0214661
6       .0045     .0053794     0.84   0.403    -.0060433    .0150433
7   -.0341667   .0307992    -1.11   0.267    -.0945319    .0261986
8       .0095     .0084595     1.12   0.261    -.0070802    .0260802
9    -.013575     .0126924    -1.07   0.285    -.0384517    .0113017
10     -.00974     .0101287    -0.96   0.336    -.0295918    .0101118
11      .01745     .0135724     1.29   0.199    -.0091515    .0440515
12     -.00085     .0132536    -0.06   0.949    -.0268266    .0251266

Thank you so much in advance.

Cat

[George Ford]

lgraph would work.

[George Ford]

you could scatter the coefficients and rarea the CI.

[Maarten Buis]

clear
input time  coefficient  std_err    z_stat  pvalue   ci95_low    ci95_high
1      -.0245       .0208597    -1.17   0.240    -.0653842    .0163842
2        .007       .0007071     9.90   0.000     .0056141    .0083859
3        .014       .0106125     1.32   0.187    -.0068001    .0348001
4      .00575      .039004     0.15   0.883    -.0706964    .0821964
5      -.0145     .0183504    -0.79   0.429    -.0504661    .0214661
6       .0045     .0053794     0.84   0.403    -.0060433    .0150433
7   -.0341667   .0307992    -1.11   0.267    -.0945319    .0261986
8       .0095     .0084595     1.12   0.261    -.0070802    .0260802
9    -.013575     .0126924    -1.07   0.285    -.0384517    .0113017
10     -.00974     .0101287    -0.96   0.336    -.0295918    .0101118
11      .01745     .0135724     1.29   0.199    -.0091515    .0440515
12     -.00085     .0132536    -0.06   0.949    -.0268266    .0251266
end

gen str3 star = "*"   if pvalue <= 0.1  & pvalue > 0.05
replace  star = "**"  if pvalue <= 0.05 & pvalue > 0.01
replace  star = "***" if pvalue <= 0.01

twoway scatter coefficient time, mlabel(star)

[George Ford]

g str3 star = (cond(pvalue<.01, "***", cond(pvalue<.05, "**", cond(pvalue<.10, "*", ""))))

[George Ford]

twoway connected coefficient time , yline(0) || rarea ci95_low ci95_high time, color(gs10%50)

[Cat Santos]

Dear George Ford and Maarten Buis, thank you so much for your help! Appreciate it!