Comments - How to Lie with Visualizations: Statistics, Causation vs Correlation, and Intuition! - Data Science Central2020-08-09T15:24:04Zhttps://www.datasciencecentral.com/profiles/comment/feed?attachedTo=6448529%3ABlogPost%3A238265&xn_auth=noThank you, Duncan. I will try…tag:www.datasciencecentral.com,2017-12-17:6448529:Comment:6694432017-12-17T17:37:52.217ZDominic Lusinchihttps://www.datasciencecentral.com/profile/DominicLusinchi
<p>Thank you, Duncan. I will try that, although for statistical analyses I do not use Excel but either Stata, SPSS or R. But I will try that just to see what results I get. Best - Dominic</p>
<p>Thank you, Duncan. I will try that, although for statistical analyses I do not use Excel but either Stata, SPSS or R. But I will try that just to see what results I get. Best - Dominic</p> For Dominic ...
You are righ…tag:www.datasciencecentral.com,2017-12-17:6448529:Comment:6696182017-12-17T11:59:21.761Zduncanhttps://www.datasciencecentral.com/profile/duncan132
<p>For Dominic ...</p>
<p></p>
<p>You are right that Excel does repeat the 95% levels in the ToolPak Regression utility. In the dialogue box for regression you can change the confidence level from 95% to whatever you want. When you do That, you will get the 95% results and your own results instead of 95% twice.</p>
<p></p>
<p>Duncan</p>
<p>For Dominic ...</p>
<p></p>
<p>You are right that Excel does repeat the 95% levels in the ToolPak Regression utility. In the dialogue box for regression you can change the confidence level from 95% to whatever you want. When you do That, you will get the 95% results and your own results instead of 95% twice.</p>
<p></p>
<p>Duncan</p> http://hackski.com/wcc/99d622…tag:www.datasciencecentral.com,2015-01-28:6448529:Comment:2446282015-01-28T04:46:23.717ZSriram Sitharaman (Latentview)https://www.datasciencecentral.com/profile/SriramSitharamanLatentview
<a href="http://hackski.com/wcc/99d6221a1d" target="_blank">http://hackski.com/wcc/99d6221a1d</a>
<a href="http://hackski.com/wcc/99d6221a1d" target="_blank">http://hackski.com/wcc/99d6221a1d</a> The data transformations (not…tag:www.datasciencecentral.com,2015-01-12:6448529:Comment:2399792015-01-12T17:03:55.213ZSione Paluhttps://www.datasciencecentral.com/profile/SioneKPalu
<p>The data transformations (not all) can be said to be invariant:</p>
<p><strong><a href="http://en.wikipedia.org/wiki/Invariant_" target="_blank">http://en.wikipedia.org/wiki/Invariant_</a>(mathematics)</strong></p>
<p>The data transformations (not all) can be said to be invariant:</p>
<p><strong><a href="http://en.wikipedia.org/wiki/Invariant_" target="_blank">http://en.wikipedia.org/wiki/Invariant_</a>(mathematics)</strong></p> Ah interesting. Great catches…tag:www.datasciencecentral.com,2015-01-09:6448529:Comment:2393672015-01-09T18:58:59.437ZAlex Joneshttps://www.datasciencecentral.com/profile/AlexJones
<p>Ah interesting. Great catches!</p>
<p>Honestly, I never use Excel for analysis-- just too many challenges (scale, robustness, etc)! But that's really the overarching theme of this article-- that statistics (and even visualization) isn't just something you run in Excel or do without proper forethought. </p>
<p>So in the spirit of things, the flaws are by design and you're on point-- there are a lot of errors and we should be cautious of being bamboozled by "data-driven results" because not…</p>
<p>Ah interesting. Great catches!</p>
<p>Honestly, I never use Excel for analysis-- just too many challenges (scale, robustness, etc)! But that's really the overarching theme of this article-- that statistics (and even visualization) isn't just something you run in Excel or do without proper forethought. </p>
<p>So in the spirit of things, the flaws are by design and you're on point-- there are a lot of errors and we should be cautious of being bamboozled by "data-driven results" because not all of them are created equal! </p> And another thing - which has…tag:www.datasciencecentral.com,2015-01-09:6448529:Comment:2395372015-01-09T17:57:09.739ZDominic Lusinchihttps://www.datasciencecentral.com/profile/DominicLusinchi
<p>And another thing - which has nothing to do with your analysis. I'm sure you have noticed that Excel generates 95% confidence interval for the regression coefficients - twice!! This has been going on since I can remember using Excel (i.e. the 90s). I am using Excel 2010 now, and Microsoft has still not fixed this problem. Have they fixed it in the most recent version? I'm not a programmer, so I don't know how difficult it would be to fix this problem.</p>
<p>And another thing - which has nothing to do with your analysis. I'm sure you have noticed that Excel generates 95% confidence interval for the regression coefficients - twice!! This has been going on since I can remember using Excel (i.e. the 90s). I am using Excel 2010 now, and Microsoft has still not fixed this problem. Have they fixed it in the most recent version? I'm not a programmer, so I don't know how difficult it would be to fix this problem.</p> Alex: should you not use a di…tag:www.datasciencecentral.com,2015-01-08:6448529:Comment:2393172015-01-08T22:20:31.911ZDominic Lusinchihttps://www.datasciencecentral.com/profile/DominicLusinchi
<p>Alex: should you not use a different data analytic approach since OLS regression assumes that observations are independent? Is it safe to assume that stock prices are independent? (I'm not a financial analyst.)</p>
<p>Alex: should you not use a different data analytic approach since OLS regression assumes that observations are independent? Is it safe to assume that stock prices are independent? (I'm not a financial analyst.)</p>